PART ONE
The Credit
Scoring Site

A bleak account
PART TWO
creditaccuracy.com
Dirty Data
Google
Web     creditscoring.com     creditaccuracy.com



creditscoring.com
in the media

Clark Howard
USA Today
Newsweek
Chicago Tribune
The Christian Science Monitor
HowStuffWorks
Federal Reserve
Credit Repair
The Detroit News
The Columbus Dispatch
The Augusta Chronicle
Bankrate.com
Bankrate.com
Realty Times
Nolo
About.com
MoneyCentral Radio
Realty Times
Money Maze Radio


The Big Meeting Transcript
A confrontation

PUBLIC FORUM: THE CONSUMER AND CREDIT SCORING

Highlights:
  • creditscoring.com: To not take too much of the nation's time with whining and long-winded stories, asked five succinct questions and even provided them in writing (indeed-- even on this web site prior to the forum). The idea was to get the questions in before the noon break: "... as promised we would like to give the audience a chance to raise some questions and issues so that we can be sure to cover them this afternoon." The questions did not survive lunch. Wasted plane ticket.
  • FTC: "I can't think of a subject that more consumers have expressed to us their desire to understand exactly what it is."
  • Fair, Isaac: The Wholesome, Brave, Hard-Working Pioneers (introduced as those who "pioneered" -- see PR)
    • Answers the big question from creditscoring.com: What horrible thing will happen if you release the consumer's score to him?
    • Fair, Isaac's presentation: an attempt to either bore you to death, or, if you do pay attention, confuse the issue. Pseudo technical babble about something completely secret-- so there is no way to challenge their statements or methodology, anyway. Here, in a national presentation, they run out of time and skip though their own material like they are ad-libbing a freshman book report. But, just shut up and take what they feed you, America-- it's PROPRIETARY (as they say, often).
    • On Non-Water Cooler Thinking-- Cold, Hard, Serious, Number Crunching, Numbers Don't Lie, Just the Facts, "It's better to count than to guess" statistics:
      • "We saw all of those numbers, so where do they come from? Well, we do not sit around the water cooler at Fair, Isaac and kind of pull them out of our ears, which again may come as a surprise to people... All of the numbers that are in those scoring models are the result of very painstaking, detailed statistical analysis of real data."
      • NON-WATER COOLER MENTALITY: "And so in fact what we've said is, we don't count any inquiries in the 30 days before this particular score is computed. So you can have a hundred inquiries in the month before the score is computed and they won't count at all. And you can have any number of mortgage or any number of auto loans in any --I'm sorry. It's a 14 day period at any time before the score is computer, and they'll only count as one inquiry each."
        So, with all this statistical, analytical, multiple regressionistic, yadda, yadda, yadda calculating going on, why didn't their algorithms spit out some unusual number of days-- like 12, 17, or 26?
    • "For most kinds of credit, 700 or maybe a little bit up in the 700's. Anything above that is considered golden for most kinds of credit. Anything below about 550 is considered awful."
  • The national credit bureaus:
    • Answer the big question from creditscoring.com: What horrible thing will happen if you release the consumer's score to him?: "It's an issue I'm not going to be able to address today."
    • A BIG BREAK FOR THE CONSUMER; SCORE ONE FOR THE FTC:
      "MS. TWOHIG: Thanks, Ray. I have one question about something Peter Mahoney said this morning. When he was arguing that the black box was a myth, which, I think, a lot of folks in the room since disagreed with, he said that you can contact each of the three repositories and they will give you what he called a feedback list. I'm not sure if that's the same as a factor list. And I was just wondering if you know anything about that?
      MR. CRESCENZO: The list of factors that are part of --that could be construed as a full list?
      MS. TWOHIG: I assume that's what he was referring to. Yeah. That's what he was referring to.
      MR. CRESCENZO: Well, I have a full list right here.
      MS. TWOHIG: So that is the case, that those factors that are used by the --
      MR. CRESCENZO: The whole list is available.
      MS. TWOHIG: Okay. I just wanted to clarify that. And if anybody called and asked for that, would you mail it out to them?
      MR. CRESCENZO: I would.
      (Laughter.) MS. TWOHIG: Call Ray.
      MR. CRESCENZO: We have been also --one of our other staff members has been working with the scoring development folks to try to expand the comment factor codes --the explanation codes --so that there is an educational message besides just the statement of lack of payment. So it's, you know, more of the evolution process that continues.
      MR. MEDINE: Ray, do you want to make that part of the record of this, so that people can have access to it and have a better understanding of how the process works?
      MR. CRESCENZO: The list?
      MR. MEDINE: Yes.
      MR. CRESCENZO: Yeah, sure."

      Now, go to his web site and try to find the factors "that could be construed" as a full list. Please send email if you find them. Send it if you don't, too.

      And, try the national credit bureaus' web sites:
      Trans Union
      Equifax
      Experian (be sure to "Ask Max")

      Try
      Freddie Mac
      Fannie Mae

      Try Fair, Isaac:
      www.fairisaac.com/servlet/SiteDriver/Content/929
      http://www.fairisaac.com/servlet/SiteDriver/Content/397#4

      Links from this page to the list from these credible sources will be provided when they get around to publishing it on the Internet. We can rank them in chronological order.

      But, don't hold your breath.

      While you're looking, try to find links to the FTC's (or any other) transcript of the Credit Scoring Forum. It may be hard, since "... 3 days prior to the meeting, you could check the following group calendar of events and find no record of this critical public forum: Fair Isaac, Associated Credit Bureau, Equifax, Experian, Trans Union, Fannie May, Fannie Mac."

  • MR. GOLDBERG: "Joe Goldberg from the Pennsylvania Attorney General's office and the Bureau of Consumer Protection. I have a very serious concern with what you said, and I also have a very serious concern with the credit bureaus. Nobody wants to take responsibility for the information that you are transferring. What Fair, Isaac is doing and what the other systems do, is making a statement about the character-- the financial character-- of a consumer, and you are conveying that to a third party. It's not purely internal. And there is a reliance on that statement for better or for worse. What you do need to do is, if you're going to take that step, I think you have the responsibility to tell the consumer the basis for the statement about the character, and the effect that changes in their behavior will have on that statement. I disagree that you don't tell them that. You are making a statement and conveying that to a third party. You have an obligation to allow this consumer to act. And by just giving a list of four factors, if that consumer adds to the consumer's detriment, in my opinion that's an unfair or a deceptive trade practice."
    (Applause.)(no response from the opposition)
  • Fannie Mae:
    • Myth of the Down Payment: "Because no longer is a low down payment loan by itself the highest risk alone in the bucket. And in fact, it may be --there may be plenty of opportunity to expand into, you know, a greater population of low down payment loans by using credit scores as a compensating factor."
    • Omigod, has it come to Approval By A Computer?:
      • "However, just as with credit scores, the use of Desktop Underwriter is meant to compliment the underwriter's judgment. It does not replace the underwriter's judgment."
      • Then on the same page: "An approved recommendation, which means that the loan meets Fannie Mae's credit underwriting guidelines and does not need further review by an underwriter."

        Eyewitness account: read from a prepared script.

  • Freddie Mac - "High Energy" and Humor (or Comedy):

    "Credit Scoring in the Mortgage Industry: Myths and Facts"
    1. "MYTH: Only One Type of Credit Score is Used in Mortgage Lending"
    2. "MYTH: Credit scores make lending decisions"
    3. "MYTH: Credit scoring models are a 'black box' and we don't know how they work"
      "You can contact each of the three repositories, and they will give you a list of the feedback messages --and there's a long laundry list of them --that each of the three repositories use when they send a credit score."
    4. "MYTH: Credit Scores are not based upon a representative population sample"
      • "... we have confirmed that generic credit scores --the stuff that's in FICO scores that we've recommended --are based upon a representative sample based on some of the stuff that Pete showed you, which is that there is a 6.7 percent on minority area representation versus an 8.3 percent of the overall population."
      • On an Experian page: "Developed from a representative national sample of roughly 1 million Experian consumer credit histories, the Experian/Fair, Isaac Model:... "
    5. "MYTH: Credit scoring has not been independently validated"
      "That is not an accurate statement. The Federal Reserve Board --actually Bob Avery is here, and you can ask him about this. They obtained some credit scores. They tested it against low income populations, as well as a broad sample of populations. Their findings were published in the July 1996 Federal Reserve Bulletin. I'm still looking for my autographed copy."

      Looking for any copy.
      Freddie Mac quotes the Federal Reserve.

      Checking with the Federal Reserve on that:

      • MS. WELSH: Hi. Kristy Welsh, K-R-I-S-T-Y, W-E-L-S-H, and I'm with creditinforcenter.com. And I had a question about what Mr. Cook said, that the Federal Reserve Board regulates or somehow evaluates the FICO scoring. How is this done? How is it tested? Are we going through a statistical analysis? Do we take case A, B, C and B and run them through the scoring? Hey, we know for sure that it matches what it should match. And if so --I mean, you said that this was done. Is this published anywhere? Can we get a copy of it?
        MR. COOK: Well, I may have mis-spoke. I don't think I said that we evaluate credit scoring systems per se. I will say that HUD has currently embarked on a study of the scoring systems of the GSE's --Fannie and Freddie. No. I do --I can tell you that in 1996 --and, Bob can correct me if I'm wrong, if he's still here --the Fed did undertake an analysis of credit scoring systems in the sense of trying to determine whether or not there were solid correlations between low scores and poor performance. And generally they found that there were. That credit scoring systems do tend to be predictive of behavior. But that leaves a lot of room, I think, for issues today.
        MS. WELSH: Right. So they haven't been evaluated for bias or discrimination or any of these kinds of things?
        MR. COOK: No. This was not an exhaustive study of each factor in a given scoring system and a relationship. I guess one of the questions that I didn't quite follow up on with Pete, you stated that the statistics are available for your clients to understand and appreciate the various risks associated with each of the factors, so that if they're called upon to justify the inclusion of a particular factor, it's predictiveness can be quantified for them based on the statistics that you've done. But let me ask you this, sort of a further question. Recognizing --sort of taking off on Debby's approach --that we're talking perhaps more than just the legal requirements here. Recognizing that probably you don't have a legal obligation to do the rest of the analysis, to do the third stage of a --what did we call it --disparate impact analysis, that while there may be a justification for using something like finance companies --a number of finance companies --is that the least --does it produce the lowest impact, or is there another set of factors that you could have used that would produce a lesser impact?
      • Federal Reserve: "Credit scoring models are proprietary and consequently public scrutiny is precluded. Suppliers such as Fair, Isaac & Co. provide substantial data to support the validity of their scores but do not publicly release information about how the formulas work."
    6. "MYTH: Credit scoring is unfair to low and moderate-income borrowers"
    7. "MYTH: Credit scoring is unfair to minority borrowers"
    8. "MYTH: Credit scores are always predictive for every single person"
    9. "MYTH: People without credit scores cannot obtain a mortgage"
    10. "MYTH: Credit scoring reduces the availability of low-cost mortgage credit"

  • Kirk Willison, Senior Vice President and Director of Government Affairs for Countrywide Home Loans:
    • "... And after listening to a lot of discussion today, I've become convinced that Congress was wrong, in that Congress should not have said that consumers really aren't entitled to their scores... "
    • SCORE ANOTHER ONE FOR THE FTC. The worst appears to be true-- it is contractual:
      "MS. TWOHIG: Kirk, I just want to make sure I understood one thing. Is it the case, though, that you don't give consumers their actual score when you explain to them the reasons why they might be denied?
      MR. WILLISON: Again, the agreement that we have with Fair, Isaac, is that we are not to release that score.
      MS. TWOHIG: Okay.
      MR. WILLISON: Our hands are tied."
  • Marcia Griffin, founder and President of Home Free USA:
    • "They need to know their credit score, and they need to know it early... "
    • "... the first secret is that the process has changed. And the second secret is to get your credit score. Get your credit report immediately, not from the credit reporting agencies, because when you write, they don't give you your credit score. They need to really get their credit report with a credit score, so that they can understand where they are... "

Public Forum :
The Consumer and Credit Scoring

July 22, 1999

AGENDA


9:00 a.m.

Introductory Remarks

Jodie Bernstein, Director, Bureau of Consumer Protection, Federal Trade Commission

9:10 a.m.

Credit Scoring: An Overview

This overview presentation is intended, in conjunction with the panel that follows, to give context to our later discussions. [PowerPoint Presentation available for download]


Peter McCorkell, Senior Vice President and General Counsel, Fair, Isaac and Company, Inc.

10:10 a.m.

The Use of Credit Scoring In The Mortgage Industry

Panelists will discuss, in addition to credit scoring, mortgage scoring and automated underwriting.
 
Panelists:
Pamela Johnson, Vice President, Single Family Mortgage Business, Fannie Mae
 
Peter Mahoney, Associate General Counsel, Freddie Mac
 
Carroll Justice, Executive Vice President, FT Mortgage Companies

11:00 a.m. - BREAK

11:15 a.m.

Consumers' Experiences With Credit Scoring

This session will identify the issues and concerns that consumers have regarding the use of credit scoring. We will first hear from the panelists, who will begin the issue identification process, followed by an open discussion during which everyone in attendance is encouraged to participate.
 
Moderator:
Peggy Twohig, Assistant Director for Financial Practices, Federal Trade Commission
 
Panelists:
Simi Batra, Consumer
 
Marcia Griffin, President, Home Free USA
 
Caryl Iseman, Owner/Broker, Action Mortgage Group

12:15 p.m. - LUNCH BREAK
A list of area restaurants is available on the table outside of Room 432.

1:30 p.m.

Is Credit Scoring Fair?

This panel will address, among other issues, whether the use of scoring adequately assesses risk for all populations of consumers, and how the use of overrides impacts the use of scoring. Following brief presentations by the panelists, we will have an open discussion of the points made by the panelists as well as any issues raised in the morning session that are appropriately addressed here.
 
Moderator:
David Medine, Associate Director for Financial Practices, Federal Trade Commission
 
Panelists:
R. Russell Bailey, Fair Lending Team Leader, Comptroller of the Currency
 
Robert Cook, Senior Fair Lending Specialist, Federal Reserve Board
 
Debby Goldberg, Acting Dir., Neighborhood Revitalization Project, Center for Community Change
 
Peter McCorkell, Senior Vice President and General Counsel, Fair, Isaac and Company, Inc.
 
Raj Mehra, Director, Financial Risk Management, PricewaterhouseCoopers
 
Margot Saunders, Managing Attorney, National Consumer Law Center

3:10 p.m. - BREAK

3:25 p.m.

What Information Should Consumers Receive About Credit Scoring? Who Should Provide That Information?

This panel will address, among other issues, whether consumers should have access to credit scores, what information is currently available to consumers concerning scores, what additional information could be made available, and the appropriate source to provide the necessary information (e.g.,consumer reporting agencies, housing counselors, real estate agents, loan officers).
 
Moderator:
Peggy Twohig, Assistant Director for Financial Practices, Federal Trade Commission
 
Panelists:
Ray Crescenzo, Vice President, Associated Credit Bureaus, Inc.
 
Virginia Ferguson, Vice President, National Association of Mortgage Brokers
 
Don M. "Dusty" Lashbrook, Executive Vice President, Mortgage.com
 
Forrest Pafenberg, Director of Real Estate Finance Research, National Association of Realtors
 
Elisabeth Prentice, New York/Puerto Rico District Director, Neighborhood Reinvestment Corporation
 
Kirk Willison, Sr. Vice President, Director of Government Affairs, Countrywide Home Loans, Inc.

4:45 p.m.

Concluding Remarks
David Medine, Associate Director for Financial Practices, Federal Trade Commission

TRANSCRIPT STARTS HERE

Document Body
Page Navigation Panel
Page 1 2
1
1
2
3
4
5
6
7
8
9
10 PUBLIC FORUM:
11 THE CONSUMER AND CREDIT SCORING
12 July 22, 1999
13 Matter No. P994810
14
15
16 Federal Trade Commission
17 Room 432
18 600 Pennsylvania Avenue, NW
19 Washington, DC
20
21
22
23

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
1.
1 Page 2 3
2
1
2
P R O C E E D I N G S
3 -----
4 (9: 02 a. m.)
5 MR. MEDINE: Good morning, and thank you all for
6 coming. This is a tremendous turn out in the middle of
7 the summer on a very, very important issue to consumers.
8 And I would like to begin by introducing the Director of
9 the Bureau of Consumer Protection of the Federal Trade
10 Commission, Jodie Bernstein.
11 MS. BERNSTEIN: Thank you, David. Thank you for
12 the introduction. And I guess they thought I needed it,
13 because I see they've made two name tags for me, which
14 either means that I won't remember my name, or we need a
15 special introduction.
16 I'm delighted to be here this morning to welcome
17 all of you to what's going to be a very important forum.
18 And it's particularly important, as David said, that so
19 many of you were able to get here. I know a lot of you
20 travelled from out of town. And we particularly
21 appreciate it, because the diversity here in the room of
22 views and expressions is going to be critical to us.
23 I can't think of a subject that more consumers

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
2.
2 Page 3 4
3
1 have expressed to us their desire to understand exactly
2 what it is. As someone said, when we get a call from
3 consumers --and we do get lots of calls about a variety
4 of subjects --it's as if there's a black box somewhere
5 and people --somebody somewhere is dropping information
6 into the black box and a score emerges. It's just almost
7 that mysterious to people, they tell us.
8 And then they want to know --and it really does
9 sound like the SATs --how can I improve my score. I
10 know that we won't --we won't produce a list that we can
11 tell consumers how to do that. It's much more complex
12 than that. But I do hope that with this group of experts
13 and the discussion we hope will occur today, that we will
14 all emerge with a good deal more knowledge and
15 understanding of what credit scoring is about. We know
16 it can be a huge benefit to consumers, and we will to be
17 able principally to be able to explain it to them and
18 ensure that they really do understand it and that they're
19 fairly treated.
20 So with the excellent panelists we have today,
21 again, I want to thank all of you for coming and for
22 agreeing to guide our discussion. I could not close
23 without noting that we have true experts to work with us

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
3.
3 Page 4 5
4
1 this morning. And I believe our first speaker this
2 morning is the gentleman sitting to my right, am I right?
3 MR. MEDINE: He is. I wanted to make a
4 housekeeping remarks first. I'm David Medine, the
5 Associate Director for Financial Practices here at the
6 FTC. This forum, first of all, would not have been
7 possible without the two people to my left, Kellie
8 Cosgrove and Peggy Twohig. And I want to congratulate
9 them for all their tremendous effort in putting this
10 together.
11
(Applause.)
12 MR. MEDINE: Cynthia Lamb also was a tremendous
13 assistant in organizing this event, and she's sitting in
14 the back.
15 A couple of just logistical points before we get
16 started. First, there will be a transcript of today's
17 meeting, and as per Commission practice, a copy of the
18 transcript will be available on our web site at
19 www. ftc. gov once it's been prepared.
20 As Jodie mentioned, the purpose of today's forum
21 is to discuss a wide range of issues associated with
22 credit scoring. We want this to be a discussion in which
23 everyone has an opportunity to participate, although the

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
4.
4 Page 5 6
5
1 first couple of sessions will be laying out some of the
2 background. But as the morning progresses, we're going
3 to encourage people to ask questions and to make comments
4 as we proceed, from the audience as well as from the
5 panel members.
6 And for those who are in the overflow rooms,
7 feel free to come by and join us with questions and
8 comments, although we may ask you to return to the
9 overflow rooms, given space considerations. But we don't
10 want people not in Room 432 to be denied the opportunity
11 to participate in today's discussion.
12 If you have any questions or need directions,
13 keep an eye out for the folks with yellow tags on them.
14 They're Commission staff, and they will be happy to
15 direct you to rest rooms, which are just outside and to
16 the left, or to our wonderful Top of the Trade, if you
17 are hungry and need a snack later on, on the seventh
18 floor.
19 It's my pleasure to start the morning off with a
20 good friend of ours, Pete McCorkell, to give us an
21 overview of credit scoring. Pete is the Senior Vice
22 President and General Counsel of Fair, Isaac and Company.
23 Jodie talked about the black box. Well, Fair, Isaac

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
5.
5 Page 6 7
6
1 makes the box, and so this will be a wonderful
2 opportunity for us to get a peak inside the box and get a
3 better sense of how the scores --credit scores are
4 prepared.
5 I think we all discovered that credit scoring
6 had entered the mainstream when in an episode of Murder
7 One a couple of years ago, someone talked about their
8 FICO score as if everyone knew what that meant. And so I
9 think credit scoring moved into the mainstream at that
10 point.
11 Fair, Isaac, of course Pete's company, pioneered
12 the commercial development of empirically derived,
13 predictive models for the credit industry and has really
14 popularized and expanded their use. They are located in
15 California, and they have subsidiaries dealing with the
16 whole issue of decisions using scoring methodology.
17 Pete supervises Fair, Isaac's legal affairs. He
18 joined them in 1987, and has provided advice and
19 assistance to credit grantors in connection with consumer
20 litigation and regulatory proceedings. He is the
21 company's primary liaison with the federal credit
22 regulatory and enforcement agencies, including the FTC,
23 and has conducted seminars for us and other agencies to

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
6.
6 Page 7 8
7
1 educate us about the scoring. And today he'll educate
2 you about the scoring as well.
3 Pete?
4 MR. MCCORKELL: Thanks, David. When I came in,
5 I used a line from Butch Cassidy and the Sundance Kid. I
6 said, can I move. So I'm going to get up and move.
7 Okay. I'm going to sort of approach credit
8 scoring a little bit like peeling back an onion. We've
9 got a lot to cover in the next hour, and so kind of put
10 on your seat belts and try to stay with me.
11 We're going to cover a number of topics, and
12 really in some sense the most important part of what I've
13 got to say is really bringing everybody's awareness about
14 what the nature of the credit decision is. Because an
15 awful lot of the alleged shortcomings of credit scoring
16 turn out to be problems that are inherent in the nature
17 of the credit decision and applied other ways of making
18 credit decisions.
19 I'm also going to talk about sort of the
20 differences between statistically based decisions and
21 judgmental decisions. I'm going to then go into a little
22 bit, hopefully in a very nontechnical way, about how
23 credit scoring systems are developed. I'm going to talk

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
7.
7 Page 8 9
8
1 about different types of credit scoring systems. I'll
2 focus a little bit on the credit bureau scoring systems,
3 because I think that's probably what most people have in
4 mind when they hear the term credit scoring or FICO
5 scores. And then finally, if there's enough time, touch
6 very briefly on some legal and public policy issues
7 related to credit scoring.
8 Actually, one more. Casey Stengel actually had
9 it about right. Making predictions about the future can
10 be kind of dangerous. But in fact, that's what credit
11 grantors have to do. They have to make predictions about
12 how borrowers will behalf, how borrowers will repay that
13 obligation. And when you come right down to it, there is
14 no way to predict with certainty how any individual
15 borrower is going to behalf, how they're going to repay.
16 But what you can do, is that you can predict how
17 groups of borrowers will repay with reasonable certainty.
18 You can predict that if I have a thousand customers with
19 these characteristics, 900 of them, or 950 of them, or
20 990 of them will repay in a satisfactory manner and the
21 others won't. But nobody can predict that this
22 particular borrower will or will not repay in a
23 satisfactory manner.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
8.
8 Page 9 10
9
1 And so what we're talking about is not
2 predicting with certainty the behavior of an individual,
3 but really trying to estimate the risk that different
4 borrowers pose. And risk by its very nature is assuming
5 that you're looking at a group of people, because
6 ultimately every borrower is either going to behave in a
7 satisfactory manner or they're not. They're either going
8 to repay in a reasonably satisfactory manner, or they're
9 going to go delinquent. Maybe go all the way into charge
10 off.
11 So at the end of the day with any particular
12 borrower, you can say, yeah, this was an absolutely great
13 borrower, or this was a terrible borrower. I should
14 never have extended that loan. But when you're making
15 that decision, you can't say for sure that this borrower
16 will or will not repay. And that really gets into kind
17 of the meat of credit scoring and what it's all about.
18 Basically there are fundamentally two ways of
19 making credit decisions. The sort of old fashion
20 judgmental type of decisions, where somebody takes a look
21 at whatever information is available, and sits down and
22 scratches his or her head, or pats his or her stomach or
23 whatever they do to think great thoughts, and comes to a

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
9.
9 Page 10 11
10
1 conclusion that this is a loan that I'm willing to make.
2 The other type of decision process is --and in
3 many cases the two are actually used jointly in any given
4 case. But fundamentally the other type is using the
5 statistical method of making decisions, and that broadly
6 is what I'm going to refer to as credit scoring.
7 Now, a little bit later on I'm going to get into
8 some terminology. In the mortgage industry, at least,
9 credit scoring has come to have, for at least some
10 people, a somewhat more narrower definition. But at
11 least for right now, I'm using credit scoring in kind of
12 the broadest sense of any kind of statistically based
13 credit decision system.
14 So what are the differences between those two
15 types of decisions? Well, actually what I really want to
16 talk about first is what are the similarities. Both
17 methods assume that the future is in some large measure
18 going to resemble the past. Both methods, you're taking
19 a look at the applicant that you've got to make a
20 decision on today, and comparing that applicant with
21 qualities or characteristics of borrowers that you've had
22 experience with in the past, and making the assumption
23 that if today's applicant looks like the people who

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
10.
10 Page 11 12
11
1 generally repaid in a satisfactory manner in the past,
2 then this applicant is likely to repay in a satisfactory
3 manner.
4 The days of Mr. Bailey and the Savings and Loan
5 from It's a Wonderful Life are long gone. Very few of us
6 are known personally to loan officers. Most credit today
7 is extended by mail, sometimes over the phone and
8 increasingly over the Internet. It's probably been 15
9 years since I've had an up front personal encounter with
10 a loan officer, and I suspect that the same is true for
11 most consumers.
12 Today you just don't have that friendly
13 interaction across the desk, and even when you do, you're
14 likely to have --you're not going to have it with Mr.
15 Bailey who knows everybody in town and has known you and
16 your parents and your grandparents. You're going to have
17 it with somebody that doesn't know you from a hole in the
18 wall.
19 And so while your own past history is certainly
20 an important part of the credit decision, in many ways
21 whether somebody is making judgmental or statistical
22 decisions, they're comparing what they know about you to
23 what they know about other applicants and how those other

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
11.
11 Page 12 13
12
1 applicants behaved as borrowers.
2 And then finally, it gets back to the idea of
3 trying to grant credit to acceptable risks. Now, a lot
4 of people making judgmental decisions may sort of fool
5 themselves into thinking well, I know for sure this
6 borrower won't repay, or I know that if I extended credit
7 to this guy, he would not repay. But in fact, you just
8 can't make those individual decisions with any degree of
9 certainty, so in either case you're really trying to hit
10 that acceptable degree of risk.
11 Okay. Now, what are the differences? Well,
12 credit scoring really defines that degree of risk. It
13 allows you to come up with a numerical measure of the
14 degree of risk that any borrower presents. It lets you
15 rank order those borrowers with respect to one another.
16 And I'll get into that a little bit later.
17 It explicitly lets you make those decisions
18 based on a degree of risk. A lender can say, I can make
19 money if I extend credit on this product to people where
20 10 of them will repay for every one that won't, and
21 therefore I have to have at least a 10 to one odds of
22 repayment in order to make money. So I'll extend credit
23 to everybody whose odds are 10 to one or better. If

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
12.
12 Page 13 14
13
1 they're worse than 10 to one, I'm not going to extend
2 credit to those people.
3 It lets you track how those borrowers perform
4 based on the criteria that you're using in the credit
5 scoring system. It lets you make adjustments in your
6 policy. If somebody is making judgmental decisions and
7 the boss comes in and says, our losses are too high,
8 we've got to tighten up, every credit officer is going to
9 have a different idea of what tighten up means. Or if
10 the boss says, gee, our market share is slipping, we're
11 losing business to the competition, we've got to loosen
12 up, every loan officer has to figure out what loosen up
13 really means.
14 With credit scoring somebody could come in and
15 say, we need to tighten up. Let's raise the cutoff by
16 five points. And because of the statistical work that
17 goes into the system, the management knows that if they
18 raise the cutoff by five points, they'll reduce the
19 acceptance rate by X amount and also reduce the
20 delinquency rate by Y amount.
21 And then finally, credit scoring allows the
22 automation of a lot of credit decisions, which I think
23 some of the other folks are going to talk about, but

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
13.
13 Page 14 15
14
1 which has really been a big part of the acceptance of
2 credit scoring. Certainly in small business lending, for
3 example, probably more people have gone to credit scoring
4 for the benefits of automation.
5 In small business lending, for example, the
6 typical statistics we hear are that it takes 10 to 12
7 hours of loan officer time to process a small business
8 loan judgmentally. Using credit scoring and automated
9 underwriting, it can be done in 15 to 30 minutes. Now,
10 if you can take 10 or 11 and a half hours out of the
11 process --and that's not data entry time. That's fairly
12 high paid time. If you take that much time and salary
13 expense out of the process of making a small business
14 loan, you can make a lot more small business loans,
15 especially small dollar small business loans, than you
16 could make if you were doing it judgmentally.
17 One way to sort of think about the differences
18 is that probably in most cases, even somebody making a
19 judgmental decision has a list of factors that they're
20 going to look at when they are reevaluating an applicant.
21 And, again, to just get some lingo in here, what we call
22 characteristics are really the different types of
23 information that are looked at the decision process.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
14.
14 Page 15 16
15
1 Another way that you can think about
2 characteristics is that you might think that each
3 characteristic implies a question that you could ask and
4 then get the answer, either from a credit application or
5 from a credit report.
6 Next slide, please.
7 And so in this case, at the top of the page
8 we've got some information that you might get from a
9 typical credit application, and then down towards the
10 bottom below the blanks are a number of characteristics
11 that you might get off of a credit bureau report. And
12 somebody making a judgmental decision is going to be able
13 to go through that list of characteristics, and probably
14 about the best that they're going to do is be able to say
15 well, this guy looks pretty good on this, but not so good
16 on that.
17 In this case, time at present address looks
18 good, and time on job looks good. Oh, but this guy is a
19 renter and not a homeowner, so that's not so good. Debt
20 ratio is good. He's got a good bank reference. The N/ A
21 after age doesn't mean that there's no age. It means
22 that legally age is not supposed to be used in judgmental
23 credit decisions because of ECOA in Rate B. But they do

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
15.
15 Page 16 17
16
1 allow it to be used in scored decisions.
2 And so you just go through the list like that,
3 and at the end of the process judgmentally you say, oh,
4 this is an acceptable borrower, and that again implies
5 that this is a borrower who poses an acceptable level of
6 risk. You couldn't begin to site the odds of that
7 borrower if you evaluate that borrower using judgmental
8 processes.
9 With credit scoring --and again here is another
10 thing to keep in mind. You're probably going to go
11 through very much the same list of factors that somebody
12 would go through judgmentally. And that's not just sort
13 of, you know, conjecture on our part. One of the things
14 that Fair, Isaac does, is when we build a credit scoring
15 system for somebody that hasn't used credit scoring
16 before, we also build a statistical model of their prior
17 judgmental decision process.
18 And so we can actually find out how --what
19 factors at least appear to be important in that
20 judgmental decision process, and almost invariably will
21 find that very much the same factors are being considered
22 --at least apparently considered --in the judgmental
23 process as wind up being considered in a credit scoring

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
16.
16 Page 17 18
17
1 system.
2 And so the list of factors, the types of
3 information that are considered, are very similar.
4 They're not --again, other than age, which for the legal
5 reasons can't be --or least maybe more accurately, isn't
6 supposed to be used in judgmental decisions. Other than
7 age, the types of information considered are going to be
8 pretty much the same.
9 But with a credit scoring system, for each of
10 those questions that are implied by the characteristics,
11 each possible answer to that question is going to have a
12 number of points assigned to it. And so you could add up
13 all those points and you get the overall score. In this
14 case, you get the same decision, that is an acceptable
15 borrower. In other words, a borrower that poses an
16 acceptable level of risk.
17 But you get one very important additional piece
18 of information, and that is, you get told this borrower
19 is 11 to one. In my example earlier, I said we could
20 afford to extend credit to people at 10 to one odds.
21 This guy is 11 to one, so we'll take him.
22 Next slide, please. This is a sample --and let
23 me stress a sample --of what a credit --the guts of a

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
17.
17 Page 18 19
18
1 credit scoring system, the credit scorecard itself, might
2 look like. It is not an actual credit scoring scorecard,
3 so don't go home and try to score yourself on this and
4 get into a major tiz about your score.
5 I'll just use it to illustrate a couple of
6 points here. First of all, in this case we've got the
7 characteristics down the left hand column. Again, those
8 questions --those types of information. And now across
9 in each row next to the characteristics, we've got the
10 different attributes. The different possible answers for
11 each of those questions, and you'll see that each of them
12 have a different number of points assigned.
13 A couple of things that I want to point out on
14 here, you'll see --and again above the jagged line are
15 pieces of information typically that you get from a
16 credit application, and below the line are
17 characteristics that you would typically get from a
18 credit report.
19 For all of those application characteristics,
20 and actually implicitly also for the bureau
21 characteristics, at the very far right of each column --
22 each row --there is a box labelled NI. That stands for
23 no inform, which is Fair, Isaac speak for no information

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
18.
18 Page 19 20
19
1 value. That may be different than the questions not
2 answered.
3 If you look at the fifth line down, Department
4 Store/ Major Charge Card, you'll see there is a box for no
5 answer. Sometimes when a question is asked on an
6 application and the applicant leaves that blank, the fact
7 that they left it blank tells you something. On the
8 other hand, in a lot of cases credit lenders change
9 applications, or for one reason or another a particular
10 piece of information may not be available for a given
11 individual.
12 We have figured out that it would not be fair to
13 assume that that means that person should get zero points
14 for that piece of information. What they get is what we
15 call the no inform points, which if you go to the next
16 slide, is really can of the neutral value. No inform
17 means that there is no information value to say that this
18 applicant is either better or worse than the average
19 applicant based on that particular piece of information.
20 So if you have a characteristic like age, where
21 you've got, you know, kind of a nice, smooth numerical
22 scale in most cases, there is going to be some age value
23 that's exactly equal to the no inform points. For a

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
19.
19 Page 20 21
20
1 given applicant, not because the applicant refused to
2 answer something, but because you didn't ask the question
3 on an application, if you're missing a piece of
4 information, it will go to the no inform points.
5 In other words, it says we don't have a piece of
6 information on this factor that tells us that this
7 applicant is likely to be risky or less risky than the
8 average applicant who walks through the door. So that's
9 what the NI or no inform points are about.
10 Okay. We saw all of those numbers, so where do
11 they come from? Well, we do not sit around the water
12 cooler at Fair, Isaac and kind of pull them out of our
13 ears, which again may come as a surprise to people. I
14 read something just last week that talked about the
15 scoring model assumes that. Well, scoring models don't
16 assume anything. All of the numbers that are in those
17 scoring models are the result of very painstaking,
18 detailed statistical analysis of real data.
19 The first thing that we have to do in developing
20 a scoring model is to get a sample of borrowers where we
21 can say, this borrower's performance is known to the
22 extent where I'm willing to say that if I knew when I
23 made the decision what I know now about how that borrower

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
20.
20 Page 21 22
21
1 has actually performed, I could say either I would make
2 the same decision to extend them credit, or if I had
3 known then what I know now, I wouldn't have extended
4 credit.
5 And that's what we mean by goods and bads here.
6 I'll use that term again probably throughout the rest of
7 my hour here. That's not a moral judgment or anything
8 else. In some senses, the good and bad refers to the
9 credit grantor's own decision. Did I make a good
10 decision in extending credit to that borrower, or did I
11 make a bad decision. Or in a sense in kind of the way we
12 define it, if, again, knowing how that borrower actually
13 performed you would still extend them credit, then that
14 counts as a good. If you knew how they would perform and
15 you wish you hadn't extended them credit, then that
16 counts as a bad.
17 For the people that you've been accepting under
18 --and let's assume this is somebody that's been using
19 judgmental decisions. For the people that have been
20 accepted before, of course it's very easy to figure out
21 who the goods and the bads are. You just wait a couple
22 of years, depending on the type of credit product, and
23 you see how they perform. You ask yourself, am I glad I

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
21.
21 Page 22 23
22
1 extended credit or not.
2 And so you can count the known goods and bads
3 out of that Accepts on the right hand side there. Again,
4 just to use nice round numbers here, we're assuming we
5 start out with 10,000 people. Under the old decision
6 process, 7,000 were accepted and 3,000 were turned down.
7 For the 7,000 Accepts, figuring out the goods and bads is
8 just a matter of counting based on how they actually
9 performed.
10 But in building a scoring system, we don't want
11 to ignore these 3,000 rejects. The example I gave of
12 somebody saying I've got to have 10 to one odds to make
13 money means that somehow if you could improve your
14 decision process, if you could make sharper decisions
15 about individual risk level, for every 11 people you turn
16 --if you were turning down people that were below 10 to
17 one, for every 10 people you turn down, in certain groups
18 at least, as many as nine of them would have turned out
19 to be good customers if you would have accepted them.
20 And, again, that's an important concept to grasp
21 for lots of types of --in fact, virtually any kind of
22 credit product. To set a cutoff score for economic
23 efficiency, the credit grantor is going to wind up

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
22.
22 Page 23 24
23
1 setting the cutoff score at a point where more of the
2 people that are turned down would have been good, and
3 sometimes by a very large majority. Sometimes it may be
4 10 to one. It may be even more than ten to one. Odds
5 are the point where it becomes unprofitable, because the
6 loss for one bad customer is a lot greater than the
7 profit from any single good customer.
8 And so when somebody says, you know, our cutoff
9 is 415, that doesn't mean that everybody who scores 414
10 and lower is going to go delinquent, and in fact at 414,
11 an awful lot of those people would probably have
12 performed just fine if you had accepted them. But it's
13 not economical given the facts of that particular
14 portfolio to extend credit to those people whose odds are
15 worse than that.
16 We don't want to ignore them, because in fact we
17 know that there is good business in those prior rejects.
18 And so what we go through is a process that has a number
19 of names. The one that I think is most descriptive is
20 Reject Inference. For those of you who paid attention to
21 your high school math classes, you will remember a
22 process called extrapolation, where you could infer where
23 a point was on a line where you didn't actually have any

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
23.
23 Page 24 25
24
1 sample points. It went beyond the ends of the line when
2 you're putting stuff on graph paper.
3 Remember when you were a junior or sophomore in
4 high school, you were putting stuff on graph paper, and
5 if you only had samples, they were sort of in the middle
6 of your graph. You could nevertheless extrapolate to
7 where some points would be outside where you had actual
8 samples.
9 Well, Reject Inference, you can think of that as
10 kind of a multi dimensional form of extrapolation, where
11 again we look at the rejects. We see how they compared
12 on all of the relevant characteristics to the Accepts,
13 and we infer from that, we estimate from that, we
14 extrapolate from that, how they would have performed.
15 And indeed we see here, that of those rejects,
16 the odds were about nine to one. They're almost --
17 they're about nine times as many inferred goods as there
18 are inferred bads in the Rejects. So to build a scoring
19 system, we fold them back in with the known goods and the
20 bads, and, again, hopefully the prior decision process
21 made some sense.
22 And so the odds from actual performance of the
23 people that were accepted, the overall odds for that

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
24.
24 Page 25 26
25
1 group --not the odds at cutoff. But the odds for that
2 group as a whole are 24 to one. The odds for the Rejects
3 are 8.7 to one. And again, I sort of use these numbers
4 because they will work out well for my next example. The
5 overall odds for that population turn out to be 16 to
6 one.
7 So if all that I know is that somebody has
8 walked in the door in this population, I can say, well,
9 if I extended credit to everybody that walks in the door,
10 for every 16 people that paid in a satisfactory way, one
11 of them wouldn't. And that's actually pretty --that's
12 pretty high odds. More typically, in most populations
13 we're going to see something --in the population odds,
14 we're going to see something on the order of ten to one,
15 eight to one and something like that.
16 But as I said, that 16 to one number happens to
17 work well for my next slide and actually several slides
18 after this. So we've now sort of divided the world into
19 the goods and the bads. There aren't any uglies, just
20 goods and bads. So now we continue the process of a very
21 careful analysis of actual data, and we see how people --
22 look at each characteristic, and we look then at each
23 attribute of each characteristic, and we see how those

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
25.
25 Page 26 27
26
1 people perform.
2 And my first example here is age. And we'll see
3 that for people under 30, that group had 10 percent of
4 the goods, but 40 percent of the bads. And so the
5 information value for that attribute is one to four. And
6 again, these numbers here are in percentages, not in
7 absolute numbers. But if all I knew about somebody is
8 that they're in this population and that they are under
9 30, I can multiple the population odds of 16 to one by
10 the information odds for that one isolated piece of
11 information of one to four, and the first line at the
12 bottom there, I find out that this person is four to one.
13 It's the 16 to one multiplied by one to four.
14 On the other hand, if I have somebody who is
15 over 50, you've got exactly the reverse situation.
16 Again, this is just a sample --an example. There 40
17 percent of the goods are over 50 and only 10 percent of
18 the bads. So the information odds there are the other
19 way around. It's four to one. Multiple that by the
20 population odds of 16 to one, and I can say people over
21 50 --my best guess, again, just based on this one piece
22 of information, age, are that they're 64 to one.
23 Well, now I've started spreading people out away

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
26.
26 Page 27 28
27
1 from that overall population odds of 16 to one. I've got
2 the younger borrowers over here riskier than the general
3 population, and the older borrowers a lot less risky than
4 the general population.
5 Of course, we don't look at just one factor. We
6 may throw in another factor, on the next slide, such as
7 own or rent, or residential status. And again, not
8 terribly surprising, the owners are better than average.
9 They're two to one. The renters are worse than average.
10 They're one to two.
11 And then on the next slide, I put those two
12 factors together and things started to get really
13 interesting, because now if I have an under 30 renter,
14 I'm taking again the population odds of 16 to one, the
15 information odds for under 30 of one to four, the
16 information odds for renter of one to two, and I multiple
17 that out, and now that under 30 renter is only two to one
18 odds. So if I lend money to three under 30 renters, I
19 can expect two of them to repay in a satisfactory manner
20 for every one that doesn't.
21 On the other hand, if I take my over 50
22 homeowner and multiple that out, now I've got odds of 128
23 to one. So I'm pulling different groups of that

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
27.
27 Page 28 29
28
1 population sort of further away from that center line by
2 adding the information value from these various factors.
3 Now, this example makes the assumption that all
4 of these factors are independent. Those of you who
5 remember the O. J. Simpson trial and the DNA evidence,
6 you'll remember there were some pretty staggering numbers
7 that the DNA experts cited of 9.5 billion to one odds
8 that these DNA markers could have belonged to anybody
9 other than O. J. Simpson.
10 Well, they were going through very much the same
11 process with a large number of individual DNA factors --
12 DNA characteristics, if you will. And based on the
13 assumption that they were independent, that's how you get
14 those kind of staggering numbers like 9.5 billion to one.
15 And if all of these factors in a credit decision were
16 also independent, you might get some pretty staggering
17 numbers.
18 That's not the way it works in credit. And in
19 fact, what we'll find is that there are correlations --
20 in some cases very high degrees of correlations --among
21 the different factors that we look at, so we've got to
22 correct for that. We don't assume that they're all
23 separate. We certainly wouldn't ever make a credit

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
28.
28 Page 29 30
29
1 decision using just age, or just own or rent, or just
2 those two factors, because in fact one of the things we
3 know --and it's pretty intuitive if you think about it
4 --is that as people get older, they're more likely to
5 become homeowners rather than renters.
6 You know, those of you who are over 30, you
7 know, remember back when you were in your 20's. Probably
8 most of us spent a few years as renters in our 20's, and
9 the chances of becoming homeowners get a lot higher as we
10 get older. And also, except for those who are in their
11 40's and still living with mom and dad, the time at
12 address is also going to go up as you get older.
13 And, you know, in a lot of cases, about the same
14 time that you get old enough to be legally able to borrow
15 anything, you're likely to leave home to go to college or
16 join the Navy and see the world or whatever, and so your
17 time at address takes a big hit. And then as you get
18 older, you're likely to be at the same address longer.
19 Homeowners, of course, are less likely to move than
20 renters, so there is a large correlation between time at
21 address and home ownership, as well as with age.
22 So these overlapping circles here, if you think
23 of the size of each circle independently as the

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
29.
29 Page 30 31
30
1 information value in that particular characteristic, the
2 overlap just suggests that because of these correlations,
3 while each of those factors taken individually may have a
4 fairly large information value, as you start adding these
5 correlated characteristics, you then have to correct for
6 that overlap.
7 And on the next slide, I throw in one more
8 factor. In this case, time on job. Well, of course,
9 time on job is likely to be correlated with those other
10 three factors. Again, unless your parents forced you
11 into child labor in the family business, if you're in
12 your 20's you're probably not going to have been at the
13 same job for very long. Your time at address and time on
14 job frequently go together, because if you get a new job,
15 in a lot of cases you've got to move, or maybe you have
16 an opportunity to move, and again, the home ownership.
17 So while time on job standing on its own has a
18 reasonable information value, what this represents simply
19 is that where you've already got those other factors --
20 kind of stability characteristics, age related
21 characteristics in the system --adding one more factor
22 of the same kind just brings very little additional
23 predictive value.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
30.
30 Page 31 32
31
1 And that's important for a number of reasons,
2 because typically in building a scoring system, we'll
3 start out with maybe as many as 50, and maybe as many as
4 a hundred, characteristics that independently have a
5 reasonable amount of information value. But because of
6 these correlations --these overlaps --as we actually
7 start building the system and we start getting a few
8 characteristics with a lot of information value in there,
9 a lot of those additional characteristics won't get into
10 the final system because of these overlaps.
11 Nevertheless, all of the information value there is still
12 captured by the characteristics that are in the system.
13 And so frequently somebody that's not sort of
14 familiar with the concepts of credit scoring comes along
15 and says, well, gee, this system doesn't have time on job
16 in it. I know that's predictive with credit performance.
17 There is something wrong with your credit scoring system.
18 No, actually the information value is in there, but it's
19 just picked up by those other factors that have
20 correlated with time on job. And so sometimes somebody
21 will come along and say, well, but you didn't consider
22 such and such. It's not in your scoring system. You
23 didn't consider such and such about me.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
31.
31 Page 32 33
32
1 Well, we may have considered that implicitly
2 because of all of these other factors that are in the
3 scoring system, and so for a number of reasons, they're
4 characteristics that people try to keep out of scoring
5 systems. One of them --something that, you know, sort
6 of might be a very obvious characteristic is income. It
7 sounds like a nice numerical characteristic. It's
8 probably got something to do with how people are going to
9 be able to pay their bills.
10 But it turns out that in practice income is
11 difficult to work with, because no matter how explicitly
12 the application asks the question, people insist on
13 answering it in different ways. You may ask for monthly
14 pre-tax income and you get semi-weekly, before tax and
15 after tax, and you get with and without child support and
16 alimony. And you get the rental property, even though
17 there is another box on the application for that. And it
18 actually turns out to be a very messy characteristic.
19 And so somebody says, well, you didn't ask for my income.
20 How could you possibly make an intelligent credit
21 decision. Well, it's because of these correlations.
22 If we go to the next slide, sort of the goal of
23 all of this, as I said, we're trying to separate --move

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
32.
32 Page 33 34
33
1 those people apart on the risk scale away from that
2 average that you know just because they walk in the door.
3 And what we really want to do in building this credit
4 scoring system is come up with a system that will
5 separate the eventual goods from the eventual bads by as
6 much as possible, so that --now there will always be
7 some overlap in the distributions.
8 Actually this doesn't quite do it justice,
9 because in reality the distribution of goods would go all
10 the way down here to the bottom of the scale, and the
11 distribution of bads would go all the way up to the top
12 of the scale. But what we're trying to do is move those
13 two distributions as far apart as we can, so that the
14 bulk of them in my nice bell shaped curve here --and of
15 course life isn't usually quite that perfectly drawn, but
16 it kind of approximates that.
17 What we're trying to do is move those two curves
18 as far apart as possible, so that we can set a cutoff
19 score that turns away as many of the bads as possible,
20 while turning away as few of the goods as possible. Now,
21 these are in percentages. Hopefully in any population
22 the goods would be a lot smaller in absolute numbers --
23 the bads would be a lot smaller than the goods in

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
33.
33 Page 34 35
34
1 absolute numbers. So these are in percentages, which is
2 why the curves are the same size.
3 But you can see at the top where the two curves
4 are a little further apart, you can set a cutoff that
5 cutoffs most of the bads and only a very small percentage
6 of the goods. Down at the bottom, the two distributions
7 are closer together, and so to turn away the same
8 percentage of bads, you've got to cut off twice as many
9 of the goods.
10 Now, that's the goal of the credit scoring
11 system. There is no absolute right or wrong, and we call
12 that measure divergence, just how far apart those two
13 distributions are. There is no right or wrong answer for
14 that. The goal is to get those two distributions as far
15 apart as possible in any given case, because then it lets
16 you make more efficient decisions at cutoff.
17 For example, for a college student population,
18 college students, you know, all look at lot alike in
19 terms of credit factors. Again, unless they're still
20 living at home, they're all renters. They've all been at
21 the address for a short time. If they've got a job at
22 all, they've been on the job for a very short time. They
23 have very little credit history, etc. So you might --if

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
34.
34 Page 35 36
35
1 you had a college student population, you simply couldn't
2 get the two curves as far apart as you could for a more
3 general population.
4 Next slide. Now, technically what credit
5 scoring systems do is to rank the order of consumers by
6 risk, because they don't consider things like economic
7 changes. If the economy goes south, people that scored
8 200 in this example that equaled odds of 20 to one, if
9 the economy takes a real downturn, those people that were
10 scoring 200 may find themselves with odds of 18 to one or
11 15 to one.
12 And so we don't purport to provide an absolute
13 odds quote, but, of course, what every credit grantor
14 wants to do is to turn that score into an odds quote.
15 And there are ways that they can do that, both from the
16 development statistics and also in terms of watching how
17 their own population performs.
18 So we get through all of this, and somebody is
19 probably still out there saying, oh, gee, but why do you
20 think scoring can do a better job of predicting credit
21 behavior of estimating risks than judgmental decisions.
22 It's real simple. It's better to count than to guess.
23 And fundamentally that's the difference between

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
35.
35 Page 36 37
36
1 how somebody using a credit scoring system is making
2 decisions. They're counting very carefully the prior
3 performance. They're taking a very hard look at the
4 prior performance of different kinds of people,
5 separating that into individual characteristics, and then
6 comparing that in a very disciplined way to what you know
7 about the next applicant through the door.
8 If somebody is making judgmental decisions,
9 they're trying to keep all of that in their head and, you
10 know, there is no way that you can juggle 10 or 12
11 different characteristics, each of which has three to six
12 different attributes. There is simply no way that you
13 can juggle that much information in your head, because if
14 you have 10 or 12 characteristics with three to six
15 attributes each, you have 10 to 20,000 possible
16 combinations in there. There is no way that somebody
17 judgmentally can keep straight 10 or 20,000 different
18 combinations of attributes.
19 And so to a large extent, they're guessing.
20 They're taking --making an informed guess, hopefully in
21 most cases. But really they're guessing at how today's
22 applicant compares to the applicants that they have seen
23 in the past and for whom they have actual performance.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
36.
36 Page 37 38
37
1 Okay. Real briefly I want to talk a little bit
2 about different types of credit scoring systems.
3 Historically most credit scoring systems were built for a
4 particular credit grantor, what we call custom
5 scorecards. The next category of borrowed scorecards is
6 really not terribly relevant these days, although in a
7 few cases it might be possible for one credit grantor to
8 borrow a system from a different credit grantor.
9 What is becoming very common, and again probably
10 what most people think of when they hear about a credit
11 scoring system, are credit scoring systems built from
12 data across multiple credit grantors. Now, of course,
13 the biggest example of that are the credit bureau
14 systems, that I'll talk about in more detail, and what we
15 call pooled system.
16 And then in a very few cases, somebody may come
17 along and want to do something kind of entirely new and
18 different. It doesn't happen in the United States,
19 again, very much these days. We do this more overseas.
20 Sometimes somebody will come along and want to do
21 something so new and different. They say well, we don't
22 have any data on this. We don't have any data on similar
23 portfolio. Fair, Isaac, you've been in this business for

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
37.
37 Page 38 39
38
1 40 years, what can you do for me.
2 And then in that case we can sit around the
3 water cooler and say, well, based on our 40 years of
4 experience in lots of different systems, here are some
5 guesses we can make about the kind of borrowers you're
6 going to see. We can build a judgmental --what's really
7 a judgmental scorecard, or what we call a launch
8 scorecard, just to get them going until they've got some
9 data.
10 Next one. The information used in a credit
11 scoring system can come from a variety of sources. Of
12 course, the most common are credit reports and credit
13 applications. A bank may have prior experience with a
14 particular customer. Sometimes demographic information
15 may be useful. For existing customers the issuer's
16 billing file has a lot of good information. And in
17 secured lending, the terms of the deal, the amount down
18 and things like that, can be very useful.
19 Next. Credit scoring can also be used to make a
20 variety of different types of decisions. Last year there
21 were about three billion pre-screened credit card
22 solicitations mailed in the United States. Probably 2.9
23 something billion of those were credit scored. Believe

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
38.
38 Page 39 40
39
1 it or not, not everybody in the world gets those pre-
2 screened solicitations, and typically one of the screens
3 they go through is credit scoring.
4 New applications, again, is probably what most
5 people think of in terms of credit scoring, making a
6 decision on an application for new credit. On the other
7 hand, if you pay attention if you own a credit card or
8 two, you'll notice that these days the issuers, for most
9 of you at least, try to do nice things once in a while.
10 They send you that letter that says because
11 you're such a wonderful customer, you don't have to make
12 the minimum payment this month. Or because you're such a
13 wonderful customer, we're going to increase your limit.
14 Or if you should happen to be standing at the cash
15 register with a purchase that's going to put you over the
16 limit, they want to know whether or not they're going to
17 authorize that transaction.
18 And finally when that piece of plastic expires,
19 they want to decide should we reissue it at all. If so,
20 should we reissue it for six months or for three years.
21 And so they're using decisions to manage those existing
22 --they're using credit scoring to manage those existing
23 accounts. They may also be using credit scoring to try

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
39.
39 Page 40 41
40
1 to decide whether or not to try to sell you another type
2 of product, based on what they know about your behavior
3 on the one you've got.
4 And finally, in certain kinds of lending,
5 especially mortgage lending, lenders frequently resell a
6 lot of the portfolio. The lender is using credit scoring
7 to decide which parts of the portfolio it wants to sell
8 versus which ones it wants to keep. And the buyer or
9 investor, on the other hand, may be using credit scoring
10 to decide which loans it is interested in buying.
11 There are also a number of different performance
12 definitions. Up until now I've sort of talked about
13 good-bad or really a scoring system designed to predict
14 any credit delinquency. We can --and there are systems
15 designed specifically to predict the risk of bankruptcy,
16 rather than other types of delinquency.
17 If somebody is making a mailing --a pre-screen
18 mailing --they want to know whether somebody is likely
19 to respond, as well as if they do respond and you issue
20 them a card, will they pay. There are a lot of people
21 who are absolutely great credit risks who are literally
22 not worth the price of a stamp for a pre-screen mailing,
23 because they are so unlikely to respond.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
40.
40 Page 41 42
41
1 And actually I'm kind of in that category. My
2 wife gets a lot more pre-screened mailings than I do,
3 because in fact we have sort of different patterns of
4 credit uses. And so somebody looking at her credit
5 bureau file is more likely to say this is somebody that
6 might be interested in another credit card. McCorkell
7 hasn't gotten another credit card in 15 years. We're not
8 going to waste a stamp on him.
9 Okay. Really just kind of an intro for the next
10 panel, scoring today in the mortgage industry, which is
11 probably as David and Jodie mentioned, is really how
12 scoring came to kind of the front of public
13 consciousness. Scoring is really used in a couple of
14 different ways. And, again, there is some terminology
15 here.
16 A lot of the mortgage industry refers --when
17 they say credit scores, they're referring to scores built
18 just on credit bureau data that are installed at the
19 credit bureau. When you order a credit report, a score
20 is generated. And when people in the mortgage industry
21 talk about credit scoring, they may mean what I would
22 probably refer to as bureau or credit bureau scores.
23 And most of those credit scores or credit bureau

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
41.
41 Page 42 43
42
1 scores used in mortgage lending, interestingly enough are
2 what I labelled here as generic scores, which are
3 designed to predict the risk of any credit delinquency.
4 There are also some mortgage-specific credit bureau
5 scores that were designed specifically to predict the
6 risk of mortgage delinquency. Actually what we found is
7 that the generic scores do a very good job of predicting
8 mortgage at risk, as well as other types of credit risks.
9 And then finally, what I've labelled
10 comprehensive mortgage scores, I think a lot of folks
11 just refer to as mortgage scores, where they are looking
12 not just at the credit bureau --the credit history piece
13 of information, but they're also looking at application
14 information and the deal terms --the amount down and
15 that kind of thing --in making a more comprehensive
16 decision. I think the next panel will get into that a
17 little more.
18 So what I would like to turn to next are
19 specifically the credit bureau scores. And here, first
20 of all just kind of broadly, are the types of information
21 that are considered by the credit bureau scores. Now,
22 these are scores that Fair, Isaac and others --there are
23 other scoring developers out there. I don't know why

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
42.
42 Page 43 44
43
1 anybody would use them, but some people do.
2 So there are other flavors of credit scores, and
3 occasionally somebody will use a FICO score to mean any
4 kind of credit score or any kind of credit bureau score,
5 even though FICO happens to be a trademark of ours. It's
6 a little like Kleenex. You don't always look at the
7 brand label when you walk into the store and buy a box of
8 Kleenex. You just want tissues to blow your nose. FICO
9 scores have sort of gotten the same kind of generic use
10 of the brand name.
11 These are scores that are running just on the
12 credit bureau data, and they're looking generally at five
13 types of information. And they're really listed here in
14 the order of importance. First is previous credit
15 performance. Has this person performed well, or have
16 they gone delinquent on other obligations. Not
17 surprisingly, that's the most predictive piece of
18 information we've got.
19 Secondly, now those for people where the answer
20 is yes, all of their prior credit they've performed in a
21 satisfactory way in the sense that they've met all of the
22 contractual obligations and they've never been late,
23 etc., that accounts for about 85 percent of the

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
43.
43 Page 44 45
44
1 population. And so there are only, again, rough numbers,
2 15 percent of the population that's got serious
3 delinquencies noted on their credit report at any given
4 time.
5 The other 85 percent of the population does not
6 all present the same degree of risk in terms of future
7 credit behavior. And so we've got to look at other types
8 of information as well to see how those folks are likely
9 to perform in the future.
10 And the next most predictive piece of
11 information is what's labelled, current level of
12 indebtedness. In one sense, you can sort of think of
13 that as how close to the edge are these folks. Is it
14 somebody --you know, somebody who has got five credit
15 cards, all of them they have borrowed right up to the max
16 and they're making minimum payments, are a lot riskier
17 than somebody that their balance --the limit ratio is
18 down in the 10 to 20 percent range.
19 The amount of time that credit has been in use.
20 And actually that line is a little bit mislabelled. It
21 probably ought to say sort of the amount of credit
22 history available, which is really a combination of the
23 time that credit has been in use and also the number of

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
44.
44 Page 45 46
45
1 trade lines that are on the file.
2 Now, trade lines can be on the file that aren't
3 active any more. In fact, again, those of us who have
4 been around for a while, sort of after we got into the
5 credit world, we probably have a lot of accounts that
6 we've paid off and closed. That good information, in
7 most cases, will stay on the file kind of indefinitely.
8 And so somebody that's got 20 or 30 years of
9 credit experience and any kind of reasonable number of
10 trade lines, maybe a lot of which have been paid off --
11 you know, student loans that have been paid off.
12 Mortgages that have been off. You know, your first
13 credit card may have been a gas card or a store card,
14 because that was the only place you could get a credit
15 card when you, you know, first got out of school.
16 So those have been paid off. All of those
17 historical pieces are likely to still be on the file, and
18 so we look at that. It's really the amount of credit
19 history available.
20 The next category we look at is the pursuit of
21 new credit. This is where the inquires come in and also
22 new account openings. Somebody that has a lot of
23 inquiries generated by the consumer --and I'll get into

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
45.
45 Page 46 47
46
1 that in a little bit --or has opened a lot of new
2 accounts recently, is more risky than somebody that
3 hasn't been out there actively pursuing a lot of new
4 credit.
5 If I could actually then skip down to I think
6 number --yeah, number 33. There are a few --I'm
7 skipping a few slides in the presentation, because we're
8 kind of running out of time here. Keep going. Keep
9 going. Bingo. Okay.
10 I mentioned inquiries. This has been a topic of
11 some controversy. Fair, Isaac about a year ago
12 redesigned our bureau systems to do what we call inquiry
13 de-duping. We know that in a lot of cases somebody
14 shopping for a mortgage or for an auto loan will wind up
15 with multiple inquiries from the same transaction. And
16 of course not many people buy more than one house in any
17 given month, or more than one car.
18 And so we've sort of been able to say, if we
19 count all of the auto loan inquiries and all of the
20 mortgage loan inquiries within any given 30 day period as
21 a single inquiry, because it's probably only one auto or
22 one house that they're shopping for, that that is really
23 a better representation of their credit behavior.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
46.
46 Page 47 48
47
1 And then also, again in mortgage lending
2 especially, people may wind up shopping around for some
3 period of time looking for the best rates, going to
4 different lenders or go through a mortgage broker who
5 shops at the different lenders. And so in fact what
6 we've said is, we don't count any inquiries in the 30
7 days before this particular score is computed.
8 So you can have a hundred inquiries in the month
9 before the score is computed and they won't count at all.
10 And you can have any number of mortgage or any number of
11 auto loans in any --I'm sorry. It's a 14 day period at
12 any time before the score is computer, and they'll only
13 count as one inquiry each.
14 Okay. If we could slip I think two slides.
15 It's just an example of the de-duping. One more. One of
16 the things we've done with the bureau scores is to
17 actually use multiple scorecards rather than a one size
18 fits all scorecard. And that will become relevant later
19 in the afternoon when we talk about what ought to be
20 disclosed.
21 And I think the question that Jodie posed is,
22 how can I raise my score. Because of the multiple
23 scorecard design, it's not possible to tell somebody, if

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
47.
47 Page 48 49
48
1 you do this, your score will go up by X, because in fact
2 the credit bureau information is a moving target.
3 There's always information coming into the credit bureau.
4 Essentially on a daily basis credit bureaus are getting
5 information from credit grantors.
6 Also, our systems consider the age of certain
7 types of information. A delinquency last week is a lot
8 worse than a delinquency five years ago. A new account
9 that was opened up last month is more significant than a
10 new account that was opened up three years ago in terms
11 of pursuit of new credit.
12 And so even if you don't do anything with your
13 credit. You don't use your credit card in a given period
14 of time and you don't make a payment. And of course if
15 you don't make a payment for too long, something will
16 happen, because the credit grantor will now start
17 reporting you as delinquent. But, you know, you don't
18 miss a payment, but you don't make a payment. Even if
19 you don't do anything with your credit history, just the
20 passage of time is going to affect some of those things.
21 And so I could sit there with somebody's credit
22 report and say well, if you close this account, or if you
23 paid off that account, here's what would happen to your

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
48.
48 Page 49 50
49
1 score if nothing else changed. But something else will
2 always change. And then one of the things that may
3 change, is that by doing something you may wind up on a
4 different scorecard, because the way that the system is
5 split, is first of all it looks for the presence of
6 serious delinquency information, either in the trade
7 lines or in the public record part of the credit file.
8 And so we have sort of the goods and the bads --
9 the previous goods and bads now separated using different
10 scorecards, because if we used a one size fits all
11 scoring system, anybody with any delinquency would get a
12 terrible score. But, in fact, not everybody with a prior
13 delinquency presents the same degree of risk, so we can
14 do a better job.
15 Now, the people with prior delinquencies as a
16 group are riskier than the people with no delinquencies
17 as a group. But we can do a better job by this split
18 scorecard design of finding the folks with prior
19 delinquencies that today pose a reasonable risk versus
20 the ones with prior delinquencies who are still just
21 terrible risks going forward.
22 And then depending --and then after we sort of
23 make the split based on the existence of delinquency,

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
49.
49 Page 50 51
50
1 then we also look at some other information: the
2 thickness of the file, in a sort of philosophical sense
3 how many trade lines, etc., there are, the age of the
4 file, how long has this person had established credit.
5 Again, somebody who is 50 years old and has 30
6 trade lines, most of which are now closed, that may be a
7 very typical pattern. Somebody who is 22 years old and
8 has 50 trade lines, and they've only been in the file for
9 24 months and they've got 50 trade lines already, that's
10 probably kind of a strange bird.
11 And so we're using different scorecards to
12 evaluate those different groups of people and the recency
13 of new trade lines opened. Again, somebody who is kind
14 of out there in a steady state mode who hasn't taken out
15 a lot of credit recently versus somebody who is kind of
16 out there shopping actively, we're using different
17 scorecards to evaluate them.
18 Now, the same score still means the same thing.
19 If you get a 700 on any of those score cards, it still
20 means the same degree of risk. But we get there by
21 looking at different information and maybe weighing
22 information differently, depending on which of those sub-
23 populations somebody is in.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
50.
50 Page 51 52
51
1 Okay. If you could again skip two. Just very
2 briefly, sort of the legal public policy issues, I guess
3 I'll save most of that for the panel this afternoon about
4 is it fair. I really just want to describe some of the
5 other slides that are in your packet.
6 We did two studies --and if you would go to the
7 next slide, please. We did two studies which are
8 labelled the LMI Study --low to moderate income. We
9 took data where we had --where we were developing custom
10 scorecards for specific credit grantors, and in many
11 cases where there were credit grantors who were going
12 from a judgmental decision process to a scoring
13 environment.
14 We looked at that data. We said sort of how do
15 the low to moderate income folks do when they're scored
16 versus how they were doing when the credit grantors were
17 making judgmental decisions. What is different about the
18 low to moderate income population versus the general
19 population, and sort of is scoring really an effective
20 way of making decisions on those populations.
21 A lot of people assume that for a lot of low to
22 moderate income borrowers, scoring just can't do a very
23 good job of evaluating credit risks, because they look

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
51.
51 Page 52 53
52
1 different than the rest of the population. Well, in
2 fact, what we found out was that not surprisingly the
3 lower income borrowers as a group scored lower. That's
4 why that yellow line is lower than the other lines in
5 this graph.
6 If you go to the next slide, what we also found
7 out is that at any given score, the risk of those low to
8 moderate income borrowers was in fact at least as great,
9 and usually in most cases just a little bit greater, than
10 the general population. So that if a score in this case
11 of 200 equaled the 20 to one for the general population,
12 for the low to moderate income population it probably
13 equalled 19 to one or 18 to one. In other words, they
14 were just a little more risky than the general population
15 in any score. So, in fact, the scores were doing a very
16 good job of predicting the risk for that population.
17 The other thing that we looked at --again, if
18 you would skip two. This is why they were different. If
19 you skip to the next one, there are two pages that look
20 very similar in your set, and unfortunately they are
21 mislabelled in the printed set. They're right on the
22 screen.
23 Slide number 42 should say "Maintain acceptance

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
52.
52 Page 53 54
53
1 rate; compare bad rates." And what we did here, is we
2 took the acceptance rate for the LMI population under the
3 prior judgmental decision process. We looked at the bad
4 rate that was coming up. And in fact, you see in some of
5 these cases --in fact, in a lot of these cases that bad
6 rate is terrible. It's horrible. The decisions that are
7 being made judgmentally on that LMI population are bad,
8 terrible decisions.
9 You know, the first couple of lines you've got
10 an acceptance rate of 26 percent and a bad rate of almost
11 11 percent. The next line you've got an acceptance rate
12 of 24 percent and a bad rate of 18 percent. Nobody is
13 going to live with an 11 or 18 percent bad rate in a bank
14 card portfolio. If you kept making judgmental decisions
15 there, what's going to happen is somebody is going to
16 come along and say, wait a minute. Our bad rate for low
17 income borrowers is totally unacceptable. We just can't
18 lend money to low income borrowers at all.
19 So what we did is say, well, okay. If you keep
20 the same acceptance rate and now you make those decisions
21 using scoring, what happens to the bad rate --and in
22 fact you'll see in every case the bad rate goes down, and
23 in many case very dramatically.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
53.
53 Page 54 55
54
1 The next slide, again, is mislabelled in the
2 printed set. It's okay up here. It should say "Maintain
3 bad rate; compare acceptance rates." Now, again,
4 maintain the bad rate for most of these is probably not a
5 realistic scenario. But using scoring, what you see is
6 if you were willing to live with that bad rate --and
7 actually the third and the fourth ones, the bad rates may
8 be acceptable, 2.3 and 4.3 percent.
9 But the acceptance rates judgmentally were very
10 low for those folks. If you evaluated that same
11 population using credit scoring and you're willing to
12 live with that two and four percent bad rate, you could
13 accept in the one case --in bank card number four you
14 could accept four times as many applicants as you were
15 accepting judgmentally. And that's again very consistent
16 across that.
17 The next study that we did was using credit
18 bureau information. And there, rather than looking at
19 the low to moderate income population, we looked at
20 consumers in zip codes with a high concentration of
21 minorities, blacks and hispanics. And I could get into a
22 lot detail about why we chose the various categories we
23 did. But basically we looked at zip codes with those

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
54.
54 Page 55 56
55
1 high minority concentrations. One of the things we found
2 was that those zip codes represented a little bit less of
3 the credit bureau database than they did of the entire
4 population, although certainly they weren't totally
5 unrepresented, as I think a lot of people assume.
6 And indeed there are folks in that population
7 that simply don't use mainstream credit. They are never
8 going to have a credit report. But they're never going
9 to go to a lender that would pull a credit report in the
10 first place, and secondly the degree of under
11 representation isn't so great that it would really affect
12 the ultimate scorecard design.
13 And we see there --if you would go to the next
14 slide --that in fact that zip code definition winds up
15 looking a lot like the low to moderate income definition,
16 because those zip codes have approximately --the average
17 household income in those zip codes --and this was a few
18 years ago now --is about two thirds of the average --
19 or, I'm sorry --of the household income of the remainder
20 of the population. So not surprisingly in economic
21 terms, those high minority zip code populations look a
22 lot like the low to moderate income populations, and so
23 not surprisingly we came up with very much the same

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
55.
55 Page 56 57
56
1 results.
2 If you go to the next slide, again you'll see
3 the score distribution is a little bit lower, which again
4 means that at any given cutoff score, you're going to
5 accept fewer of those folks. But if you go to the next
6 slide, you'll see again exactly the same pattern. In
7 this case the red line is the minority population. The
8 dotted yellow line is the rest of the population. At any
9 given score those folks represent a slightly higher
10 degree of risk.
11 So again, the scorecard is doing an effective
12 job of rank ordering risk, and to the extent there is any
13 discrepancy in the risk estimation, it actually favors
14 those high minority areas of borrowers just as it did the
15 low income borrowers. Again, for the credit grantors
16 listening, that difference is very slight. It's not
17 worth getting excited about. But to the extent there was
18 any difference, it was consistently in favor of the low
19 to moderate income and the minority zip code borrowers.
20 Again, we looked at how those two populations
21 differed, and finally we asked ourselves, well, could we
22 build --next slide. Could we build a scorecard
23 specifically for that minority population that worked

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
56.
56 Page 57 58
57
1 more effectively than the scoring systems that are being
2 used now.
3 Well --and next slide, please. What we found
4 out was that if we used a one size fits all credit bureau
5 score, in fact then we could build a more effective
6 minority zip code scoring system. Now, I suspect that
7 the folks from the Fed and some of the other regulatory
8 agencies might want to have a conversation with us if we
9 did that. But this was a research exercise. We found
10 that it would be possible.
11 But what we also found was that the existing
12 Fair, Isaac credit bureau score using the multiple
13 scorecard approach, where you're looking at different
14 sub-populations separately, where you're looking
15 separately at the folks with prior delinquency, and
16 you're looking separately at the folks that don't have a
17 lot of established credit, that that multiple scoring
18 system that we were using was in fact more effective in
19 predicting risk in that minority zip code population than
20 a scorecard --a single scorecard developed specifically
21 for that population.
22 I've run a little bit over my time. I thank
23 everybody for your indulgence, and I'll see you this

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
57.
57 Page 58 59
58
1 afternoon. Thank you.
2 MR. MEDINE: Thank you, Pete, for an excellent
3 overview of a very complex issue.
4
(Applause.)
5 MR. MEDINE: Now we're going to take a minute
6 and get our next panel set up here, and then we'll resume
7 in about two minutes.
8
(Brief recess.)
9 MR. MEDINE: We just heard from Pete McCorkell
10 on the issue of credit scoring generally. This next
11 panel is going to focus on the use of credit scoring
12 specifically in the mortgage industry. Obviously home
13 purchases are a consumer's largest purchase. And so how
14 credit scores interrelate with that decision is of
15 critical importance to consumers. And the presentations
16 will also include information about mortgage scoring and
17 the automated underwriting process.
18 We're going to start from my extreme right, so
19 I'll first introduce Pamela Johnson from Fannie Mae. Pam
20 is Vice President of Single Family Mortgage Business for
21 Fannie Mae. She is responsible for new product
22 enhancements to Fannie Mae's automated underwriting tool,
23 the Desktop Underwriter, and is responsible for credit

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
58.
58 Page 59 60
59
1 research and data acquisition for Fannie Mae.
2 We're thrilled to have her here this morning.
3 Thank you, Pam.
4 MS. JOHNSON: Thank you. Can everybody hear me
5 okay? All right. Are we ready? Let me start first of
6 all by stepping back for a minute, because I think I want
7 to clarify and make sure everybody understands where we
8 are here.
9 Peter did a fabulous job walking through how
10 does one develop a scorecard. But I want to be clear
11 that there are many different types of scorecards out
12 there. There are many different --they're used in a
13 variety of industries, not just in the mortgage industry.
14 In fact, the scorecards have been around a lot longer,
15 and used for a longer period of time, than they have in
16 the mortgage industry.
17 The consumer industry --for example, credit
18 cards, etc. --have been using scorecards, I don't know,
19 how long, Peter, 20, 30 or 40 years. But in the mortgage
20 industry it's only been recent that we've adopted the use
21 of credit scoring as a tool in underwriting.
22 And I want to be clear about which scorecard we
23 recommend, which is the use of the credit bureau

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
59.
59 Page 60 61
60
1 scorecards. So Peter spoke a little bit about this in
2 his presentation. And I wanted to point out that the
3 characteristics in the credit bureau scorecard are those
4 listed on page 24 of Peter's presentation, and not the
5 ones that he talks about early on, which could relate to
6 other scoring models.
7 So in case you missed it, he listed the five
8 characteristics that are most prominent in a credit
9 bureau scorecard. And those were previous credit
10 performance, current level of indebtedness, amount of
11 time credit has been in use, pursuit of new credit and
12 types of credit available. So just as a way of
13 clarification.
14 First of all, let me say I'm delighted to be
15 here. I thank David and team for inviting me. I think
16 this is a critically important issue, and I think this is
17 a terrific forum.
18 Credit scores and automated underwriting are a
19 very important tool in the mortgage industry. At Fannie
20 Mae we spent a considerable amount of time researching
21 and analyzing credit scores. And long before we
22 determined that they would be useful in mortgage
23 underwriting, we went through a lot of work to answer

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
60.
60 Page 61 62
61
1 some very important questions. Those questions are: do
2 they work, are they predictive of risk as it relates to
3 mortgage lending, are they fair, do they work for all
4 borrowers in the same way, and when used correctly, will
5 they expand home ownership opportunities.
6 We answered yes to all of those questions after
7 a lot of research, and I want to take you through some of
8 our research and some of the analysis we did to get
9 there.
10 The first slide, please. So today I'm going to
11 talk to you about credit scoring, mortgage scoring --
12 because they are different --and automated underwriting.
13 I'll also discuss how scores should be used to originate
14 and underwrite mortgage loans.
15 Next slide, please. As Peter indicated, credit
16 scores are empirically derived in a statistical method of
17 assessing risks. A credit score --and specifically,
18 again, we're talking about a FICO credit bureau score --
19 is based solely on the information in the borrower's
20 credit report. The evaluation of the information in the
21 borrower's credit report, meaning how the borrower
22 manages credit, has always been an important factor in
23 mortgage underwriting. Credit scores when used correctly

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
61.
61 Page 62 63
62
1 are based on information in the credit report that has
2 been proven to be predictive of credit risk and loan
3 performance.
4 Next slide, please. Mortgage scores, on the
5 other hand, are slightly different. A mortgage score is
6 also based on information that has proven to be
7 predictive of credit risk. A mortgage score is based on
8 information about the borrower obtained from the loan
9 application and the credit report, as well as information
10 about property value.
11 Mortgage scoring supports a comprehensive
12 analysis of both the borrower's ability to repay a loan
13 and the borrower's management of credit. A mortgage
14 score helps an underwriter perform a more consistent,
15 objective and accurate evaluation of this information
16 than traditional underwriting does.
17 Next slide, please. Fannie Mae recommends that
18 lenders use credit scores, specifically the FICO credit
19 bureau scores, in their manual underwriting process. In
20 addition, we have developed a more comprehensive
21 customized mortgage scoring model for use in Desktop
22 Underwriter, our automated underwriting system.
23 Next slide, please. Okay. So why should we use

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
62.
62 Page 63 64
63
1 credit scoring in mortgage lending.
2 Next slide, please. Credit scores are in fact
3 predictive of credit risks for all loans and all
4 borrowers. A borrower with a credit score below 620 is
5 in the neighborhood of two and a half and sometimes three
6 times more likely to default on a mortgage loan than
7 someone with a credit score between 660 and 699.
8 Next slide, please. Okay. This slide
9 illustrates this point. And if you'll look at the green
10 box in the middle, which indicates a score --a loan with
11 a score range of 660 to 699, and you compare that to a
12 loan with a score range of less than 620, you can see the
13 dramatic difference in the likelihood of default
14 performance.
15 Next slide, please. Similarly, this slide shows
16 how default rates on mortgage loans increase as credit
17 scores fall. Specifically, this chart shows that for
18 every 60 points in a decrease --a 60 point decrease in a
19 credit score, there risk of default doubles. If you'll
20 look at the blue line, the blue line is in fact a
21 geometric curve that shows a doubling of risk every 60
22 points, and the red lines are the actual default rates by
23 a FICO score. And you can see how closely correlated

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
63.
63 Page 64 65
64
1 they are.
2 Next slide, please. We have learned from the
3 analysis of loans delivered to Fannie Mae that a
4 borrower's management of credit, as measured by a credit
5 score, had little correlation with the borrower's income.
6 In other words, it's just as likely for a low income home
7 buyer to have a high credit score as a high income home
8 buyer.
9 Specifically, 39 percent of the borrowers whose
10 mortgage loans were delivered to Fannie Mae in 1998 and
11 who had incomes of 80 percent or less of the area median,
12 also had scores over 740. The percentage of borrowers
13 with incomes over 180 percent of area median who had
14 credit scores over 740 was 37 percent.
15 I think most of you probably know that a 740
16 credit score represents an excellent credit risk and an
17 excellent credit history, and can be used as a
18 compensating factor for other high risk factors in a
19 mortgage loan, such as a low down payment.
20 Next slide, please. Okay. The predictiveness
21 of the credit scores allows us to adopt a different and
22 more dynamic view of the risk of borrowers who make low
23 down payment loans --make low down payments. This is a

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
64.
64 Page 65 66
65
1 view of the relationship between FICO credit scores and
2 loan to value ratios. Traditionally, loan to value
3 ratios have served as an important indicator of mortgage
4 default risks with high LTV loans representing a higher
5 degree of default risk than low LTV loans.
6 This slide shows that when you combine credit
7 scores with loan to value ratio, there is an even greater
8 opportunity to expand home ownership. A borrower who
9 puts less than 10 percent down but has an excellent
10 credit history, as here defined as a credit score above
11 740, is less likely to default than a borrower who puts
12 30 to 40 percent down with a poor credit history.
13 These numbers in this chart --these are
14 relative numbers. We use as the number one what we call
15 an average loan, which is indicated --you can see the
16 one in red in the middle, which here is a loan with a 76
17 to 80 LTV and a credit score between 660 and 699. All of
18 the other numbers on the chart are related to that one,
19 meaning they perform either better than or worse than
20 that average loan.
21 Next slide, please. Credit scores are an
22 objective measure of credit risk based on information
23 from credit reports that have proven to be statistically

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
65.
65 Page 66 67
66
1 predictive of credit risk and loan performance. Fannie
2 Mae has verified that the credit scores are predictive of
3 mortgage loan performance for all loans and for all
4 groups of borrowers regardless of demographic
5 characteristics such as race or income.
6 Next slide, please. Okay. This slide
7 illustrates this last point. No matter what the race of
8 the borrower, the lower the credit score, the greater the
9 likelihood of default. And we've done this slide over
10 numerous different time periods, and the results continue
11 to be consistently the same.
12 Next slide, please. Let me tell you what Fannie
13 Mae's policy is on the use of credit scores in mortgage
14 underwriting. We advise lenders that credit scores
15 should be used as a careful --as part of a careful,
16 comprehensive analysis of each mortgage application.
17 They should not be used in isolation or as the sole
18 factor in making an underwriting decision.
19 This is the policy that we have communicated
20 with our customers --our lender customers --and we've
21 made that point clear in our lender letters that we've
22 issued on credit scoring and how to use it.
23 Next slide, please. Fannie Mae purchases loans

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
66.
66 Page 67 68
67
1 with credit scores below 620. We do not prevent lenders
2 from delivering these loans to us, nor do we make them
3 re-purchase these loans simply because the credit score
4 is below 620. Our mortgage scoring model development
5 data sets, the data sets for the scoring model, include
6 loans with scores below 620, making it possible for our
7 current mortgage scoring models that we use in Desktop
8 Underwriter to accommodate these loans and accurately
9 assess their risk.
10 Desktop Underwriter will underwrite and approve
11 loans with scores below 620 and does so. In our letters
12 to lenders on credit scoring, we have provided specific
13 examples of situations in which it would be acceptable to
14 sell us loans with credit scores below 620.
15 Next slide, please. Okay. Over the past
16 several years, Fannie Mae has developed and made
17 available to mortgage lenders our automated underwriting
18 system called Desktop Underwriter.
19 Next slide, please. Desktop Underwriter has
20 been designed as a decision support tool for lenders. It
21 helps reduce origination time and costs through
22 streamlined processes and opportunities to re-engineer
23 work flow.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
67.
67 Page 68 69
68
1 For example, Desktop Underwriter facilities
2 point of sale underwriting decisions. Desktop
3 Underwriter also helps lenders make more consistent,
4 objective and accurate underwriting decisions. However,
5 just as with credit scores, the use of Desktop
6 Underwriter is meant to compliment the underwriter's
7 judgment. It does not replace the underwriter's
8 judgment.
9 Desktop Underwriter performs a comprehensive
10 risk assessment of each loan application. This risk
11 assessment is much more comprehensive than a credit
12 score. It is based on the mortgage scoring model,
13 combined with the rules taken from Fannie Mae's selling
14 guide to determine loan and borrower eligibility.
15 Next slide, please. Desktop Underwriter, or DU
16 as we call it, issues one of three recommendations for
17 most loans. An approved recommendation, which means that
18 the loan meets Fannie Mae's credit underwriting
19 guidelines and does not need further review by an
20 underwriter. A referral recommendation, which means that
21 the loan may need Fannie Mae's credit underwriting
22 guidelines, but it needs to be reviewed by an underwriter
23 for that determine. Referral with caution

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
68.
68 Page 69 70
69
1 recommendation, which means that the loan does not appear
2 to meet Fannie Mae's credit underwriting guidelines, and
3 it must be reviewed by a human underwriter.
4 And it is in fact the referral and referral with
5 caution loans that we look to our lender partners and
6 underwriters to work with. That's the message we send.
7 Don't deny it; work with the loan. Those are the loans
8 that need extra attention.
9 Next slide, please. Desktop Underwriter does
10 not approve or decline loans. Underwriters make those
11 decisions. Fannie Mae has not used any prohibited
12 factors, such as race, age, gender, color, religion,
13 national origin or marital status, as characteristics or
14 risk factors in the mortgage scoring models or risk
15 assessment rules in Desktop Underwriter.
16 Next slide, please. As part of our ongoing
17 effort to validate that our scorecard works in our
18 mortgage scoring model, we consistently look at
19 performance. Now, this slide --this chart shows the
20 percentage of loans that become seriously delinquent, and
21 that means they've missed three payments, by the age of
22 the loan and the DU recommendation.
23 Now, the reason why we look at serious

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
69.
69 Page 70 71
70
1 delinquencies as opposed to defaults is that we don't
2 have enough time or history yet with our loans that have
3 been delivered after having been underwritten in Desktop
4 Underwriter. You need several years of history with the
5 loans --seasoning with the loans --to really track
6 default performance.
7 So we're tracking delinquency performance right
8 now, and a loan that misses three payments is considered
9 a high risk loan for foreclosure and default. After 12
10 months the percentage of the serious delinquencies is
11 more than six times greater for the loans designed refer
12 with caution than for the loans recommended for approval
13 in Desktop Underwriter.
14 Next slide, please. Desktop Underwriter
15 evaluates the information in a mortgage application based
16 on factors that have proven to be statistically
17 predictive of credit risk and loan performance. DU
18 evaluates the information in each mortgage application
19 the same way each time, consistently, objectively and
20 accurately, and without regard to the demographic
21 characteristics of the borrower.
22 Next slide, please. Desktop Underwriter does
23 not use --and I mentioned this before --any of the

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
70.
70 Page 71 72
71
1 characteristics such as race, age, color, religion,
2 national origin or marital status as factors in the
3 scoring model or the risk assessment rules in DU. In
4 addition, where the borrower lives and where the borrower
5 is purchasing a home are factors that are also not used
6 in the mortgage scoring model.
7 Fannie Mae has verified that DU's
8 recommendations are predictive of mortgage loan
9 performance for all borrowers --for all groups of
10 borrowers --for all loans and regardless of
11 characteristics such as race and income.
12 DU provides opportunity for expansion or
13 broadening of the market by looking at an array of
14 factors that can compensate for other high risk factors
15 in the loan. When used properly, and when the
16 information in the credit report is accurate --and I
17 want to emphasis that --credit scores and mortgage
18 scores and automated underwriting support consistent and
19 objective underwriting decisions.
20 And when you look back at where we've come, for
21 years we've used human underwriting and manual
22 underwriting. And it is my opinion that manual
23 underwriting creates a greater opportunity for

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
71.
71 Page 72 73
72
1 discrimination because of the subjectivity associated
2 with manual underwriting. One of the benefits of using
3 credit scoring and using mortgage scoring is it does
4 provide an objective analysis of the likelihood of
5 default or the credit risk of that particular borrower or
6 loan.
7 And that concludes my remarks.
8 MR. MEDINE: Thank you very, very much. Let me
9 just --if I could just ask a quick question --and Pete
10 eluded to this earlier --about the comparative benefits
11 of judgmental versus credit scoring.
12 You mentioned one benefit in terms of the
13 possibility of injecting discrimination into the process?
14 MS. JOHNSON: Um-hum.
15 MR. MEDINE: Have you also analyzed whether
16 certain people are more likely to qualify for, say, a
17 loan under a credit scoring system than under a
18 judgmental system? In other words, does it benefit more
19 borrowers to be scored than to have been subject to a
20 judgmental system?
21 MS. JOHNSON: I think if used correctly. Again,
22 the way we recommend that scores and mortgage scoring is
23 used, is to really take the loans that are sort of the

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
72.
72 Page 73 74
73
1 easier ones to do and put those aside, and really work
2 with the loans that are the harder loans. And what we've
3 found is that when you combine credit scoring and other
4 factors to get a mortgage score, that you actually have
5 increased your opportunities for making loans to more
6 borrowers.
7 Specifically to your question, David, when we
8 looked at the slides around the combination of LTV and
9 the FICO score, and that, you know, a 10 percent down
10 mortgage was in fact performing better than some that --
11 where someone may have put down 20 or 30 or 40 percent,
12 when you overlay credit scores on top, I think that gives
13 you an example of the opportunity that exists to expand
14 the numbers of loans one can make.
15 Because no longer is a low down payment loan by
16 itself the highest risk alone in the bucket. And in
17 fact, it may be --there may be plenty of opportunity to
18 expand into, you know, a greater population of low down
19 payment loans by using credit scores as a compensating
20 factor.
21 MR. MEDINE: Thank you very much. And Peter is
22 nodding, so why don't we turn next to Peter Mahoney, who
23 is the Associate General Counsel at Freddie Mac. Peter

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
73.
73 Page 74 75
74
1 is responsible for legislative and regulatory matters,
2 with a particular emphasis on housing policy and fair
3 lending issues at Freddie Mac.
4 He's a frequent speaker on housing, fair lending
5 and community reinvestment issues, and has published
6 several articles on community reinvestment and housing
7 issues.
8 MR. MAHONEY: You can stop reading now.
9 MR. MEDINE: Thank you for joining us.
10 MR. MAHONEY: That's the whole version there.
11 It goes on. I've got a home page that's got the whole
12 thing, rotating head and everything.
13
(Laughter.)
14 MR. MAHONEY: I'm going to stand up, because I
15 tend to have a lot of nervous tics, and so this helps me
16 get rid of it. I ran for an hour this morning. But this
17 is a pretty big group, and it's also a very impressive
18 group. I recognize a lot of the experts in the industry.
19 A lot of regulators. So I'll stand up and move around a
20 little bit and give the cameraman something to do here.
21 It doesn't surprise me. Pam and I didn't really
22 get a chance to talk about our presentations too much,
23 because we are fieriest competitors and this is the FTC,

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
74.
74 Page 75 76
75
1 and I didn't run the risk of an antitrust violation. But
2 I doesn't surprise me that Fannie and Freddie are saying
3 some of the same things. So my presentation may trod on
4 a little bit of the same ground, so I'll try to go
5 quickly over those issues.
6 I thought since I knew I was going second and
7 that Pam is so confident and well spoken and does such a
8 great job, that she would hit some of these highlights,
9 so that I could maybe concentrate on some of the touchier
10 questions regarding the use of credit scoring in mortgage
11 lending, and particularly focus on what we think of as a
12 lot of the questions we've heard since 1995 when we first
13 started talking about this, a lot of the questions that
14 we've heard from the community.
15 We do do a lot of talking and then listening
16 about credit scoring. And we've heard kind of the same
17 questions over and over, so I thought I would put
18 together for you, as soon as Mr. Clark is ready --you
19 can roll the next one. This is a logo that I have to put
20 on everything I do.
21 The myths associated with credit scoring --and
22 I didn't bring copies of this handout, but David Jamison,
23 who is a legal intern with us this summer from Howard Law

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
75.
75 Page 76 77
76
1 School, is the designated person to whom you give your
2 business card or your web address or whatever, and we'll
3 get a copy of this to you. In addition, we had some
4 materials out front. There is a report that we did in
5 1996 on automated underwriting, and issues regarding
6 fairness and consumer issues. I think those copies are
7 all gone, so we can get a copy of that to you if folks
8 have not already seen that.
9 Let's go through these myths as quickly as
10 possible. You can go ahead and roll through. These are
11 the ten myths, but I think I'll just get to them as we
12 go, because I know we're constrained for time. And I do
13 give you an instruction to cut me off, because I am
14 Irish.
15 Go to the next slide. Myth. Only one type of
16 credit score is used in mortgage lending. I think, as
17 Pam aptly pointed out, and I think Pete pointed out the
18 same thing, there are a lot of different types of credit
19 scoring models that are used in the mortgage market.
20 Pete talked at length about FICO scores. The Fair, Isaac
21 score uses only what is in one of the three national
22 credit repositories. So it is only going to deal with
23 credit performance of a consumer. It doesn't include

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
76.
76 Page 77 78
77
1 kind of personal type of information, etc. It also
2 doesn't include attributes that are uniquely associated
3 with a mortgage loan.
4 Most mortgage lenders want to know how much
5 you're going to put down for the house and what the
6 collateral value of the house is. They're going to want
7 to know what type of loan you're looking for. Is it a
8 one year ARM? That's riskier. Is it a 30 year fixed
9 rate mortgage loan? That's less risky. They're looking
10 for characteristics beyond your credit and payment
11 record.
12 FICO only looks at credit and the payment
13 record. We have our own team of crack scientists at
14 Freddie Mac. We have just --our Chairman is an
15 economist. He buys economists. We have so many
16 economists, you wouldn't believe it. We have a huge
17 group of folks who do nothing but estimate scorecards all
18 day at Freddie Mac.
19 And we have built our own custom mortgage
20 scoring model, and we have deployed it in a system called
21 Loan Prospector. It was the first automated underwriting
22 system made available broadly for mortgage lenders, and
23 it is a service that uses a custom mortgage scoring model

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
77.
77 Page 78 79
78
1 that is based not only on the information that is in a
2 credit record kept at the repositories. We go and get
3 those and bring them in. But it's also based on a host
4 of other variables that I'll discuss with you in a few
5 moments.
6 Okay. We strongly endorse the use of automated
7 underwriting. We think that's kind of the A way to
8 underwrite mortgage loans today. However, for lenders
9 who are still using manual underwriting systems --in
10 other words, they're still going down their kind of
11 written checklist and qualifying people kind of the old
12 fashion way today --we strongly recommend that they also
13 pull a generic FICO score from each of the three
14 repositories.
15 And then we tell them to throw out the high one,
16 throw out the low one, and use that middle one as a good
17 benchmark for getting you into assessing someone's credit
18 worthiness. Like Pam said, we tell people don't use it
19 exclusively. But we think it can be used as one tool in
20 the manual underwriting process.
21 Okay. Second, credit scores make lending
22 decisions. And the question David asked I was interested
23 in, because the fact of the matter is, credit scores are

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
78.
78 Page 79 80
79
1 a tool. It's like going to the gas station and saying,
2 you know, is my battery fluid low. And they put
3 something in and a little dial shuts up and says
4 zzzzzzzz. You know, that's the score of your battery.
5 That's not going to fix anything. That's not
6 going to make a lending decision. That's not going to do
7 anything for anybody. It is simply a ranking tool. And
8 once the result of the credit score is made available to
9 someone who has money in their pocket, then the lender
10 makes a decision on the basis of the information they've
11 learned from the credit score and from other things.
12 How lenders --and I include Freddie Mac and
13 Fannie Mae as a lender. How we apply credit scores, what
14 decisions we make in using credit scores, and how we
15 deploy them, that is where you get all of the questions
16 about what is the effective impact of credit scores. Is
17 it constricting credit, or is it going to expand markets.
18 And I'll talk a little bit more about that later, but I
19 did want to drive that point home.
20 Now, what it does do, is it permits you when
21 you're making decisions to do them consistently and
22 objectively without bringing into account other factors
23 that may raise significant issues about intentional bad

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
79.
79 Page 80 81
80
1 will toward a particular applicant.
2 Credit scoring models are a black box and we
3 don't know how they work. Well, I would beg to differ
4 with that, and I think this is probably one of the most
5 controversial myths. A lot of people are going, well, I
6 still don't know how they work, okay. So credit scoring
7 models rely on common and publicly available variables.
8 Pete went through, and Pam told you, page 24 of
9 Pete's slide presentation, that shows you the type of
10 variables that are used. And by variable I mean just an
11 input in the FICO generic bureau score. Okay. Now, you
12 can get better than that. You can contact each of the
13 three repositories, and they will give you a list of the
14 feedback messages --and there's a long laundry list of
15 them --that each of the three repositories use when they
16 send a credit score. That gives you a very, very good
17 idea of the type of information that the Fair, Isaac
18 model is looking at when it creates one of these generic
19 credit scores.
20 In addition, Freddie Mac has published
21 information about the variables that we rely on. We rely
22 on all those credit inputs. We also use loan devalue,
23 which is how much money is someone putting down on a

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
80.
80 Page 81 82
81
1 house. That's a very, very powerful indication. If
2 someone has one percent down on a house versus 40 percent
3 down on a house, they're going to have a very different
4 risk profile. And we use a lot of other kind of mortgage
5 related factors.
6 Let me point out one thing, though. With all
7 due respect to Pete, on page seven and eight of his
8 presentation, he listed a bunch of potential application
9 type variables. And I just want to state categorically
10 for the record that Freddie Mac does not use in any of
11 its custom scoring models time at present address, time
12 at present job and residential status. We don't use age.
13 We don't use income. We do not use, do you own or rent,
14 years at address, occupation or any of that stuff.
15 MS. JOHNSON: Ditto for Fannie Mae.
16 MR. MAHONEY: Okay. I just want everyone to get
17 that down. We're aware of some factors. A first time
18 home buyer is a predictive factor. We don't use it,
19 because our statutory purpose --one of them --is to
20 expand home ownership opportunities for people who want
21 to go from renting to owning. So we don't use that.
22 And even our written underwriting guidelines
23 have been modified heavily in the last ten years to

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
81.
81 Page 82 83
82
1 eliminate some of those old customs and practices of
2 mortgage underwriting, and to say, don't look at how long
3 someone has been on a job. Look at what their income
4 projection is for the next couple of years. That's kind
5 of the way we look at things.
6 Okay. We can go to the next one now. This
7 gives you a little snapshot of how the credit score model
8 that we deploy in Loan Prospector works and how it's kind
9 of related to standard, traditional underwriting criteria
10 for mortgages. We use collateral, down payment. We use
11 credit reputation, history of repayment, current account
12 balances, inquiries and new accounts. The same kind of
13 thing that a FICO score measures.
14 We use a debt to income ratio. We look at how
15 much cash reserves you have when your down payment is
16 done and you've closed your deal. Those factors are then
17 cooked in a big mathematical equation, with guys and gals
18 in white lab coats. And I don't understand them and I
19 don't get any of it.
20 But I do understand what the variables are, and
21 I understand that what they produce then is a set of
22 inputs that creates a very comprehensive statistical
23 profile of how --what your odds are of defaulting on

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
82.
82 Page 83 84
83
1 your mortgage in the future, based on the information
2 going in compared with this very wide range of other
3 folks. And it's as simple as that.
4 Now, credit scores are not based upon a
5 representative population sample. I think this is an
6 issue that's largely going to be going away, because we
7 have confirmed that generic credit scores --the stuff
8 that's in FICO scores that we've recommended --are based
9 upon a representative sample based on some of the stuff
10 that Pete showed you, which is that there is a 6.7
11 percent on minority area representation versus an 8.3
12 percent of the overall population. We think that's very
13 good, given how many folks have credit records, etc.
14 For our own custom mortgage scoring models, we
15 have ensured that there is a representative sample used.
16 We look at demographic characteristics. We look at
17 geographic characteristics to try to make sure that when
18 we're kind of creating the sample from which we build the
19 scorecard that those kind of factors are used.
20 But the proof of the pudding is in the tasting.
21 And the representative --we believe representativeness
22 is best demonstrated by the fact that, as I'll show you,
23 credit scoring models that we use are predictive for

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
83.
83 Page 84 85
84
1 those neighborhoods and for those folks, and therefore we
2 think that that is a very strong indication of fairness
3 in the fact that it's taken into account on enough
4 different folks' behavior.
5 Credit scoring has not been independently
6 validated. That is not an accurate statement. The
7 Federal Reserve Board --actually Bob Avery is here, and
8 you can ask him about this. They obtained some credit
9 scores. They tested it against low income populations,
10 as well as a broad sample of populations. Their findings
11 were published in the July 1996 Federal Reserve Bulletin.
12 I'm still looking for my autographed copy.
13 Recently --very recently HUD economists have
14 been working for a couple of years with Fair, Isaac on
15 estimating their own scorecard based on FHA borrowers.
16 FHA borrowers, as you know, tend to be much lower
17 income and to have a higher concentration of minority
18 households, particularly African American and Hispanic
19 households.
20 They recently reported at a Fannie Mae research
21 round table that mortgage scorecards are successful in
22 predicting risks for FHA borrowers. There is also
23 evidence that mortgage scorecards are predictive within

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
84.
84 Page 85 86
85
1 subgroups such as low income and minority borrowers. And
2 that's an article that was put together by three
3 economists at HUD, so that's a very strong, independent
4 corroboration.
5 And then we've published our own stuff, which I
6 told you about, this automated underwriting report which
7 came out in 1996. And I'm about to show you some
8 additional information.
9 Credit scoring is unfair to low and moderate
10 income borrowers. Now, Pam showed you her statistics,
11 and mine don't look much different. We both buy millions
12 of loans, so they probably wouldn't look different.
13 Actually, FICO scores and Loan Prospector scores that we
14 have do not vary much by income. And I think the reason
15 why is that income is a very lousy proxy for wealth.
16 There are people who live in Beverly Hills who are low
17 and moderate income, because they live in a census track
18 that happens to be very high income.
19 So income is a --as Pete said is a notoriously
20 difficult thing to measure. So while we take some
21 comfort from the fact that we think across --you know,
22 we try to drill down into that and confirm that our
23 scoring models work for truly low income neighborhoods

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
85.
85 Page 86 87
86
1 where the median income is very low, and they do. But I
2 guess what I'm suggesting is they're flat across the
3 board, simply because income is a notoriously difficult
4 thing to track.
5 And now I'm going to show you a picture. Yeah.
6 This is based on our 1994 mortgage purchases. This is a
7 Loan Prospector scoring model, okay, and what we did is,
8 we just cut this thing up into tenths. And we're showing
9 you, these are the high risk guys over here, and these
10 are the low risk guys over here as identified by our
11 scoring model, and these are the actual default rates.
12 Okay. This is from 1994, and we measured this earlier in
13 1999. Over a million loans. And these are low income,
14 middle income and high income folks. And I can define
15 that for you if you want me to, but trust me.
16 And what you're seeing is a very, very similar
17 distribution, okay. So the folks that are identified by
18 our model as low risk, actually are barely defaulting.
19 Those folks in the last three buckets particularly, very
20 high risk and very high default rates. Anyone in the
21 mortgage business knows if you're approaching an eight or
22 10 percent default rate, you're in a lot of trouble.
23 Next one. The next myth is that credit scoring

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
86.
86 Page 87 88
87
1 is unfair to minority borrowers. I would like to put on
2 the table once again and emphasize something I think we
3 looked past. Over the last ten to 15 years the industry,
4 with a lot of help from public policy and housing
5 advocates, has made an enormous amount of progress in the
6 way we treat a broad cross section of borrowers who want
7 to purchase a home. We have made enormous strides, and
8 there is a lot of statistical evidence to back that up.
9 And I think that one of the things we loose
10 sight of when we talk about credit scoring and the
11 progress we've made, and how rapid technological change
12 is, is that we're saying, you know, oh gosh, you know, I
13 have problems with credit scoring. Well, the solution
14 would be to go back to the way we used to be doing it,
15 where you had a single loan officer kind of deciding
16 whether a person got a loan. I don't think anyone wants
17 that system.
18 Now, as to the statistical results, again,
19 credit scores very accurately predict borrower default,
20 regardless of race or ethnicity, and in our view, again,
21 ensure a level playing field.
22 We can go to the next one. Again, this shows
23 you the demographic results. This is 1994 borrowers.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
87.
87 Page 88 89
88
1 This is just a sample. But I can tell you that I've seen
2 a lot of samples and this is quite representative of the
3 African American households, Hispanic households and non-
4 minority households.
5 Now, you do see a slightly higher default rate
6 as you get really all across here. And again, what
7 accounts for that, I'm not completely sure. I believe
8 it's a wealth factor. If I have an income of $150,000
9 and you have an income of $150,000, but I'm the first
10 person in my family that ever went to college and I'm
11 supporting my grandmother, my mother and I'm helping out
12 over family members, if I lose my job, I'm going to have
13 a higher likelihood of defaulting.
14 If I make 150 grand and I lose --I have the
15 same income as this other person. I lose my job, but I
16 can call up my mother and say, hey, can you give me
17 $5,000, I need to make some mortgage payments and pay my
18 credit cards, those are the kinds of things that we
19 really can't measure by income statistics. But I think
20 they do have a profound effect on some of these results,
21 and it is a generational issue that, you know, God
22 willing will obviate over time.
23 Go to the next one. Here is a myth that I --

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
88.
88 Page 89 90
89
1 you know, I hope people recognize. Credit scores are
2 always predictive for every single person. That is not
3 true. And so I know we're going to hear from some folks
4 who believe that credit scores were not fair and that
5 there were perverse results created by use of credit
6 scoring.
7 I am not up here to attempt to deny that there
8 are folks who could pay off a home mortgage loan or some
9 other trade account, and the credit score may indicate
10 that they are higher risk than they believe themselves to
11 be or that some other piece of evidence they have would
12 indicate. There are inaccurate credit records out there,
13 and the FCRA and other statutes are there to correct
14 that.
15 But even beyond that, a credit score is simply
16 going to say your odds versus a bunch of other people
17 across the spectrum or that you're a higher risk. And
18 that is not a moral condemnation of those folks. It is
19 what we believe to be a very accurate picture. And so
20 again, all that we're saying is that credit scoring
21 improves the odds that we're going to get it right, okay,
22 in predicting mortgage performance. Moreover, we believe
23 that over time our models will get better. We're

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
89.
89 Page 90 91
90
1 collecting more and better information on folks, and this
2 first generation of models should not be thought of as
3 fixed in stone.
4 People without credit scores cannot obtain a
5 mortgage. As Pam said, that's just not true. We buy a
6 lot of below 620 FICO scores. That number comes from the
7 industry letters that Fannie and Freddie published in
8 1995, and some folks had believed that lenders were going
9 to turn that into kind of a bright line below which you
10 are sent into the darkness to wail and gnash your teeth.
11 But we buy a lot of particular loans that have
12 gone through automated underwriting. We will balance off
13 the risks, and if someone has a dinged up credit record
14 indicated by a credit score, that can be offset by
15 compensating factors. In addition, if there are
16 particular life circumstances when we refer the loan back
17 that a borrower can work through with the lender, we will
18 buy those loans.
19 We have a lot of underwriting guidelines
20 designed to deal with extenuating circumstances and
21 difficulties in life history that mean that a credit
22 score is not indicative of the actual future mortgage
23 performance. And we take those seriously, and we buy a

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
90.
90 Page 91 92
91
1 lot of mortgage loans on that basis.
2 Now, we also work with a lot of people. I know
3 Marcia Griffin is here. I don't see her, but I know
4 she's here from Home Free USA. We work with a lot of
5 credit counseling agencies and housing counseling
6 agencies. And again, people can reestablish credit or
7 establish credit for the first time within a couple of
8 years, and that to me is one of the best roads to go,
9 which is get a good track record and then get the
10 responsibility of home ownership under your belt.
11 Next slide, please. Okay. And finally, credit
12 scoring reduces the availability of low cost mortgage
13 credit. I can tell you that there is an article right
14 here that David actually pulled off the Lexis/ Nexis for
15 me, which estimated that the efficiencies of automated
16 underwriting systems have cut costs for originators which
17 have been passed onto consumers in the form of lower fees
18 and points at closing.
19 And this guy from Nomura Securities, whom I've
20 never met, said that closing points and fees in July of
21 last year on a 30 year mortgage averaged 1.1 percent
22 versus 1.6 percent in 1993. Okay. So points have gone
23 down. Closing points have gone down. It takes four

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
91.
91 Page 92 93
92
1 minutes to approve a mortgage loan application now,
2 whereas in 1993 we were still dealing with a couple of
3 weeks for many applicants, okay.
4 We're talking about efficiencies. We're a macro
5 economic deal, Freddie and Fannie are, and what we're
6 trying to do is make the system more efficient. And
7 scoring allows us to bring technological efficiencies to
8 bear. And we can predict risks, which means we can creep
9 down on that risk spectrum. We can offer our low cost
10 financing to more people because we understand their
11 risks.
12 And we're doing that right now. We're buying
13 loans right now today that are under automated
14 underwriting that are scored that a year ago or two years
15 ago would have been called A minus. And those people
16 have a lower mortgage interest rate because of our
17 ability to tap low cost funds.
18 So we believe that over time credit scoring will
19 dramatically --dramatically --reduce costs, and that
20 most of the folks who have been shut out of the prime
21 market and into the sub-prime market are a
22 disproportionately higher share of those are low income
23 and minority folks. So our fervent belief is that this

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
92.
92 Page 93 94
93
1 is an excellent and promising development for mortgage
2 markets and that scoring properly understood is a great
3 tool to achieve our objective of expanding markets.
4 And that's all I got. Thank you.
5 MR. MEDINE: Thank you very much for your high
6 energy presentation. We next turn to Carroll Justice
7 from FT Mortgage. Carroll began his mortgage banking
8 career in 1984 and is currently Executive Vice President
9 at FT Mortgage Companies, where since 1992 he has served
10 as head of Secondary Marketing with responsibility for
11 interest rate management, private investor applications,
12 and private investor and agency relations.
13 Carroll?
14 MR. JUSTICE: Thanks, David. Good morning. I
15 want to thank you all for the opportunity to allow FT to
16 talk about an issue that has become very critical and
17 very beneficial to our industry.
18 FT Mortgage Companies is a Dallas-based
19 residential A quality mortgage lender. In 1998 we
20 originated roughly 24 billion dollars in mortgage loans.
21 We have 167 offices in 31 states. Our originations are
22 55 percent retail and 45 percent wholesale. We service
23 loans in all 50 states. We have over 450,000 loans in

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
93.
93 Page 94 95
94
1 our servicing portfolio.
2 Sixty percent of our loans are sold to Fannie
3 Mae and Freddie Mac. Twenty percent of our loans are FHA
4 and VA loans and are securitized in Ginny Mae Mortgage
5 Back Securities. And the remaining loans are sold to
6 private investors. As a typical mortgage banker, 100
7 percent of our loans are sold into the secondary market.
8 We portfolio no loans that we originate.
9 We began using credit scores, and in our case
10 FICO scores that we talked about this morning, with the
11 release of Freddie Mac's Bulletin 96.6 in September of
12 1996. With this Bulletin Freddie Mac strongly --and
13 I'll emphasize that --strongly recommended the use of
14 credit scores on all loans we sold to them. Fannie Mae
15 quickly followed suit, and now all of our significant
16 investors require some form of credit score usage. That
17 is, with the exception of FHA and VA.
18 Now, to facilitate loans moving from one program
19 to another, even though all of our investors don't
20 require them, we require a FICO score as part of the
21 credit report on the loans we originate. Now, if a
22 credit score is in dispute or not available, we use other
23 credit information and totally disregard the credit

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
94.
94 Page 95 96
95
1 score.
2 Disputed credit scores represent basically one
3 percent of our originations. Loans with no credit scores
4 are roughly one to two percent. So in effect we're able
5 to generate a credit score of somewhere between 97 and 98
6 percent of the loans that we originate.
7 Now, besides meeting investor requirements, we
8 also use credit scores because we found them to be a very
9 good predictor of credit quality. To validate this, we
10 took a look at our 1998 originations. And what we did,
11 is we looked at those loans that went 90 days into
12 default within the first 12 months or origination. Now,
13 we picked 90 days, because it's a very significant level
14 of default in a very short period of time.
15 In looking at these loans, we went into our
16 servicing system and tried to make sure --or I shouldn't
17 say tried. We made sure there was no noise coming out of
18 our servicing system, such as loan not being set up
19 properly or the borrower not receiving coupons. Any
20 reason or any time the loan went into suspense in our
21 servicing system, we threw that loan out.
22 Looking at that population and that criteria, we
23 found that loans with FICO scores below 585 were ten

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
95.
95 Page 96 97
96
1 times greater to go into a 90 day default rate in the
2 first 12 months than loans with FICO's of 660 or higher.
3 Loans with FICO scores between 585 and 620 had a seven
4 times greater probability of going into default.
5 Now, when you consider the cost of default
6 services for a mortgage banker like us, it can be as high
7 as 40 percent of the cost of servicing. The average
8 servicing cost probably runs between 60 and 70 dollars a
9 loan per year. We, with over 450,000 loans in our
10 servicing portfolio, think that credit quality is a very
11 important item in our originations.
12 A third reason for using credit scores is that
13 they are a critical component of automated underwriting
14 systems. With the introduction of Fannie Mae's Desktop
15 Underwriter and Freddie Mac's Loan Prospector, mortgage
16 lenders had new tools that allowed them to use the power
17 of the computer to analyze a very complex credit and
18 collateral decision that was required to make a sound
19 credit decision.
20 Before automated underwriting, manual processes
21 followed rules based models. If the loan didn't fit the
22 rules, it could get rejected or at least covered up with
23 paper to cover the underwriter's decision. This not only

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
96.
96 Page 97 98
97
1 increased the time and the cost to originate a loan, it
2 also increased the probability the loan could get
3 rejected.
4 After using automating underwriting systems for
5 two years, it's our experience that AU --or automated
6 underwriting --makes it easier to get a loan. It not
7 only reduces processing and underwriting times. It
8 reduces documentation requirements and also makes a
9 better credit decision than manual processing and
10 underwriting. You know, the additional documentation
11 that's required of traditional underwriting by itself can
12 cause borrowers to become discouraged and perhaps even
13 give up on the process.
14 In 1989 we pioneered a limited documentation --
15 alternative documentation is probably a better term --an
16 alternative documentation quick closing system. We did
17 this with the approval of Fannie Mae and Freddie Mac.
18 What we had found that caused us to do this, is that our
19 industry typically made the loan process and underwriting
20 requirements a very onerous process for most borrowers.
21 And we decided that whatever we could do to reduce the
22 documentation and the time required would improve the
23 likelihood that more borrowers would get loans.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
97.
97 Page 98 99
98
1 Roughly --or currently, I should say.
2 Currently roughly 80 percent of our loans --of our
3 conventional loans --go through DU or LP. We do not
4 have a proprietary system of our own. We are currently
5 using LP on our FHA and VA loans, and we were one of the
6 pilots for the Fannie Mae DU system for FHA loans.
7 Now, in addition to the agencies, many of our
8 nonconforming investors accept DU or LP for loans sold to
9 them, or they at least have their own proprietary system.
10 What we've seen over the last two years is a widespread
11 expansion of automated underwriting throughout the
12 mortgage finance industry. And a critical component of
13 that automated underwriting has been the use of credit
14 scores.
15 I want to stop here just a moment and kind of
16 explain how we use credit scores, or how we use DU and
17 LP. We gather the information. If a loan is an accept
18 on either of the systems that we use, the loan is
19 automatically approved. If a loan is a caution or a
20 reject, it then goes through our manual underwriting
21 process. What the automated underwriting system does for
22 us on the caution loans, is that it allows us to focus on
23 a specific area and makes the process much easier to

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
98.
98 Page 99 100
99
1 decide whether or not we have a good credit or we have a
2 bad credit.
3 Now, I understand that many of you question the
4 use of automated underwriting and credit scores because
5 of their potential for disparate treatment. Our
6 experience does not support your concern. We found a
7 five to ten percent increase in the number of high risk
8 loans that get approved with the use of automated
9 underwriting. We believe this is due to the system's
10 ability to examine all data available and quickly compare
11 compensating factors. A human underwriter using
12 traditional rules based systems can't make the same
13 complex decision on what has historically been a high
14 risk loan.
15 At FT the credit score is usually the --the
16 credit score is usually first available at the time of
17 underwriting when the credit report is ordered.
18 Technology constraints currently restrict us from viewing
19 it at the point of sale. In spite of this, we are,
20 however, moving towards a point of sale automated
21 origination system. In the future many of our customers
22 will know their rate, price and underwriting decision
23 when they meet with a loan officer.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
99.
99 Page 100 101
100
1 Now, the important point to remember here is not
2 how technology is going to be used. However, it is that
3 we don't use credit scores to pre-qualify, but rather as
4 a part of the total credit picture. Credit scores are
5 used at the underwriting decision point, not to pre-
6 screen applications. No borrower at FT is denied a loan
7 because of a credit score. As I said earlier, if there
8 is any question about the validity of a score, or no
9 score exists, we manually underwrite the loan using
10 traditional credit information.
11 Now, to answer how these disputed loans or no
12 credit score loans perform, we've also looked at that.
13 And what we have found is that borrowers with no credit
14 score have this 90 day, early payment default rate,
15 slightly worse than our average EPD rate. Borrowers with
16 incorrect or disputed scores have default rates roughly
17 equal to those default rates on borrowers with FICO
18 scores less than 620, or roughly seven times our 660
19 rate.
20 Now, in addition to not using disputed credit
21 scores, we don't have cutoff levels on FICO scores for
22 our loan products unless the investor requires it. For
23 conventional or FHA loans with FICO's under 585 that are

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
100.
100 Page 101 102
101
1 manually underwritten, we do require two levels of
2 underwriting approval. That's not only to make sure
3 we're getting a good loan, but it's to make sure that we
4 don't have disparate treatment against somebody who has a
5 low FICO score. If DU or LP approves the loan, we accept
6 that approval, even though it may have a low FICO score,
7 without any further underwriting.
8 After using credit scores for roughly three
9 years, we have made the following conclusions about
10 credit scores and automated underwriting. Credit scores
11 have helped us identify superior quality loans, and
12 therefore will help control future servicing and
13 origination costs. This will help us lower our overall
14 costs to produce and therefore our pricing to customers.
15 As Peter pointed out, we feel that the lower
16 costs --lower operating costs associated with credit
17 scores and automated underwriting will be passed on
18 directly to the consumer. For those of you who are very
19 familiar with mortgage bankers, you'll know there are no
20 subsidies that we don't like. If we have an opportunity
21 to pass it on to the borrower, we're going to pass it on
22 to the borrower primarily for competitive reasons.
23 We have found that a higher percentage of our

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
101.
101 Page 102 103
102
1 high risk applicants are being approved because of a
2 combination of credit scores and automated underwriting,
3 because this provides a superior credit, analysis and
4 underwriting decision to manual underwriting.
5 We now look at credit differently than we used
6 to. We used to just consider payment histories. Now,
7 the amount, frequency and recent use of credit are just
8 as important, if not more so. We now have a credit
9 measure that helps us equate credit quality with the cost
10 of default. Credit abusers will be penalized by driving
11 up the cost of housing for themselves. Credit scores
12 will help us develop an automated risk-based pricing
13 model that will allow more credit to become available to
14 high risk borrowers.
15 And finally, the use of credit scores in
16 conjunction with automated underwriting will help lenders
17 become more efficient in making loans, investors more
18 efficient in providing funds for the housing industry,
19 and improve the ability of borrowers to get housing
20 credit.
21 Thanks.
22 MR. MEDINE: Thank you. Peggy?
23
(Applause.)

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
102.
102 Page 103 104
103
1 MR. MEDINE: Peggy Twohig is the Assistant
2 Director for Financial Practices, has a question for
3 Carroll.
4 MS. TWOHIG: Carroll, I just wanted to make sure
5 I understood correctly one thing you said. Did you say
6 you do not use credit scores at the pre-qualification
7 stage at all?
8 MR. JUSTICE: No.
9 MS. TWOHIG: You do not?
10 MR. JUSTICE: No. What we've found, Peggy, is
11 that with a lack of technology to be able to deliver a
12 credit score into the origination process, we found that
13 we're better off using it as part of the total credit
14 report and using it in conjunction with automated
15 underwriting.
16 MS. TWOHIG: So your pre-qualification decisions
17 are judgmental only?
18 MR. JUSTICE: Yes. For the most part they are.
19 MS. TWOHIG: Okay. And do you know if that's
20 typical among lenders?
21 MR. JUSTICE: I'm not sure if it is or not.
22 MS. TWOHIG: Okay. Thanks.
23 MR. MEDINE: Okay. I want to thank Pam, Peter

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
103.
103 Page 104 105
104
1 and Carroll for excellent presentations. We're going to
2 take a 15 minute break and resume with consumers'
3 experiences with credit scoring. Thanks again.
4
(Brief recess.)
5 MS. TWOHIG: As David mentioned, my name is
6 Peggy Twohig. I'm one of the Assistant Directors with
7 the Division of Financial Practices with the Federal
8 Trade Commission.
9 Good morning everyone. I have the pleasure of
10 introducing some panelists who are going to talk about
11 some of these issues from the consumer perspective. We
12 had excellent presentations this morning about what the
13 perspective is on credit scoring from the industry's
14 view, what credit scoring is and how it's used in
15 mortgage lending. And jumping ahead, also we heard some
16 perspectives that address some of the issues we're going
17 to talk about this afternoon. And that was very helpful
18 in laying the ground work.
19 So now we would like to move to credit scoring
20 from the viewpoint of consumers and those that help
21 consumers with mortgage lending. Is anything wrong with
22 this picture? What are the questions? What are the
23 concerns? What are the issues that they see in credit

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
104.
104 Page 105 106
105
1 scoring?
2 And we're going to let our three panelists
3 speak, and then we would like to open it up to you all.
4 We are running a little bit late, but I think we can
5 still accommodate questions and issues raised by the
6 audience. Our purpose here in this session this morning
7 is to raise the issues --ask the questions and get all
8 the issues and questions out on the table. And then this
9 afternoon we'll try to make sure that our panelists
10 address those questions.
11 Our first panelist on this panel is Simi Batra,
12 who is representing himself as a consumer who first
13 encountered credit scoring several months ago when he
14 began thinking about purchasing a home. Simi?
15 MR. BATRA: Good morning. First of all, I want
16 to put a disclaimer here. I speak for myself, not for my
17 employer on this matter. I don't expressly represent his
18 views.
19 I am planning on buying a house in the District.
20 And about a year ago --almost a year ago I ordered my
21 credit reports. Everyone tells me, clean up these things
22 way ahead of time. It takes a while. Look for
23 inaccuracies. And I did that. And it wasn't such a

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
105.
105 Page 106 107
106
1 tough endeavor cleaning up the credit reports.
2 But now after this process --or most of the way
3 through this process, I understand that --I was under a
4 crazy impression that a mortgage banker was going to open
5 up my credit reports, go through my payment history here,
6 and ask me questions about my payment history. And I
7 guess that's not the case.
8 This is going to boil down to a number. I'm not
9 sure who boils down this number, if it's the credit
10 bureau or if it's the mortgage broker himself or herself.
11 But my credit history is going to be boiled down to a
12 number. I have no idea where that number comes from or
13 what it means. It will be rated on a scale. I'm not
14 privy to this number. I'm not privy to the scale. I'm
15 not privy to what creates this number.
16 So I meet with a mortgage broker in two days. I
17 guess I'll find out my number then and see if I was
18 successful over the last year in doing the right things
19 and everything I was supposed to do.
20 Along those lines, I have a few points to make
21 with regard to credit scoring. Some are pretty specific
22 and some are more general. Generally proponents of
23 credit scoring have two opposing things they're talking

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
106.
106 Page 107 108
107
1 about. One, they say we don't need to know the specific
2 factors of what goes into a credit score. If we knew
3 these factors, we might be able to manipulate our credit
4 scores and that would be bad.
5 But on the other hand, if a credit score is
6 indicative of our future repayment, how can it be so
7 easily manipulated? It doesn't really make sense to me.
8 If this is supposed to be a seven year history of your
9 credit repayment, and based on these solid factors and
10 this is how you're going to pay your mortgage in the
11 future, why does it really matter if I know what goes
12 into a credit score then?
13 I shouldn't be able to change these things.
14 It's predictive. It's a matter of a credit card balance
15 or a ratio. I mean, it shouldn't be that arbitrary. So
16 it's kind of --it's an opposing thing you get with the
17 credit score folks.
18 And there are many --one thing I'm kind of
19 concerned about is, I called some mortgage brokers on the
20 phone. I say what's your latest balance --or what's
21 your latest interest rate, and they'll say --they'll ask
22 me the question, what's your credit score? A lot of the
23 mortgage brokers themselves have no idea that we are not

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
107.
107 Page 108 109
108
1 privy to credit scores. Not only are we not privy to it,
2 the law says they don't have to give us our credit scores
3 specifically.
4 And a lot of the mortgage brokers themselves
5 assume that these credit scores are coming from the
6 credit bureaus. I don't think a lot of them realize that
7 they themselves --their company --has ordered these
8 credit scores. They're paying for them. And they kind
9 of think it comes with the consumer and we know what our
10 credit scores are. I mean, there's 40 or 50 credit
11 scoring models out there. I have no idea and I don't
12 know what my credit scores are.
13 So one thing, I think part of this needs to be
14 education of the mortgage brokers and those folks. Let
15 them know that consumers do not know this information.
16 We are not privy to this information.
17 Okay. What goes into a credit score? That's
18 something that I'm concerned about. I'm not really
19 worried about my payment history, or where I've lived, or
20 how long I've been at my job. Things that I can't
21 change. More things like credit card balances, the
22 balance of your debt to your available credit, the number
23 of credit cards you have open.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
108.
108 Page 109 110
109
1 Over the last year in cleaning up my credit
2 reports, what I've tried to do is consolidate my credit
3 card debts. You know, that department store thing I
4 opened ten years ago to get 10 percent off, I forgot I
5 even had it. You know, I ripped it up right away, but I
6 forgot it was still open. So I just wanted to close that
7 out.
8 But in doing so, the economy being so good, I've
9 gotten a hundred credit card offers in the mail with
10 really low interest rates. So what I did, was I --like
11 a lot of people --consolidated all my debt unto one
12 credit card. And as I'm going through this process of
13 consolidating my debt, I'm assuming day by day as I cut
14 up a credit card, or close out one or open one to move
15 it, my credit score is changing, because the ratio of my
16 debt to my available credit changes and the number of
17 open credit cards I have changes.
18 At some point in there, and over the last year,
19 I probably had an ideal credit --an ideal balance of
20 available debt --debt to available income or credit and
21 the perfect amount of open credit cards. I don't know
22 what that perfect balance is. And somehow this is
23 supposed to be predictive of my future repaying my

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
109.
109 Page 110 111
110
1 mortgage.
2 Well, if it's so predictive --it just seems
3 very arbitrary. If I applied for a mortgage half way
4 through this process, maybe I would have had the perfect
5 ratio then. I don't know. Maybe I have it now. I don't
6 know. And the point is, it's not predictive. It's
7 arbitrary. It's entirely arbitrary. It's not really
8 based on any scientific factor.
9 It's something you can change so easily. I can
10 call my credit card company tonight and ask for a greater
11 balance. You know, greater available credit. And maybe
12 they'll give it to me. And if I do, that will change my
13 credit score. I'm not sure that's really predictive of
14 anything, or it's a matter of me asking for the right
15 amount of --it's just a matter of luck.
16 Now, I have a couple real specific points with
17 regard to credit scoring. I mean really specific, and
18 it's part of --I don't know what goes into the charge
19 cards. For example, American Express. These things
20 don't have balance limits on them. Now, what are they
21 using for a balance. I mean, these things are getting
22 plugged into my credit score. I have no doubt, including
23 my payment history, which is easier. It's late payments

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
110.
110 Page 111 112
111
1 or it's not late payments.
2 But in terms of your high credit limit, your
3 available credit and things like that, I'm under the
4 impression from looking at my three credit reports that
5 charge cards --they use the highest amount you've ever
6 charged in a single month and it's essentially viewed as
7 your credit limit. That's kind of what I'm thinking
8 about. Whereas your current balance is what you have on
9 there now, and the difference between the two is your
10 available credit.
11 Now, if that's the case, that's good to know.
12 But maybe I should have known that six months ago, and I
13 would have charged a whole bunch on a charge card, paid
14 it off, and this month charged about 20 percent of that
15 amount, so I have that perfect little ratio that I don't
16 know. What if I were getting married next week and my
17 charge card --this was the month I maxed it out. Maybe
18 I should have done it a month ahead of time and then paid
19 it off. I don't know. It's arbitrary and I just wish I
20 knew that.
21 And one of the credit bureaus does not report
22 American Express correctly. They do not --what they do
23 is, every month they change --they essentially put your

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
111.
111 Page 112 113
112
1 credit --your debt is equal to the highest you've ever
2 owed. And that's --so it almost looks like you have 100
3 percent debt ratio in there. And that's problematic if
4 that is going into my credit score. I don't know,
5 because I don't know what goes into a credit score.
6 And the last specific point I want to make about
7 credit scoring is inquiries. Now, this concept is so
8 ridiculous to me. And this is something that I
9 encountered and it was very problematic. I opened up my
10 credit report and there were like 20 inquiries for auto
11 insurance. I mean, I must have called every 800 number I
12 saw a commercial for on TV to get the best auto insurance
13 rate, not knowing --no one told me, by the way, this
14 phone call will lower your credit score. This phone call
15 will lower your credit score. I wish I had known that.
16 Inquiries are treated as applications for
17 credit, whether or not you were just inquiring in a
18 dictionary sense as to just asking a question, what is
19 the rate you will give me, as opposed to actually I
20 applied for credit. Well, I'm not applying for credit.
21 I just want to know what's the rate you're going to give
22 me.
23 And I wasn't aware that a lot of these auto

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
112.
112 Page 113 114
113
1 insurance companies will look at your credit reports to
2 do that, and they will then show up on the inquiry
3 section in your credit report as an application for
4 credit. And I encountered that problem and had many of
5 those removed, which is another problem with this. It's
6 not easy to remove inquiries from your credit reports.
7 It's very hard to remove.
8 Credit bureaus will not investigate them as they
9 do a normal late payment. Therefore, the obligation is
10 up to the consumer to do that. And to try to call an 800
11 number that really doesn't exist, it's just --they don't
12 exist. I mean, people do not work behind these numbers.
13 And it's unbelievable the amount of time you have to
14 spend to get an incorrect inquiry off your credit report.
15 It takes forever. It's a multi month process.
16 And once you do finally get that letter from
17 that auto insurance saying we did not mean to look at Mr.
18 Batra's credit file, the credit reporting agencies will
19 say, well, we will take up to 90 days to then correct
20 that information. So you spent, you know, six months on
21 this process, and then they'll tell you they have another
22 90 days to correct that information.
23 I find that a bit problematic, and especially

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
113.
113 Page 114 115
114
1 since I don't know how that's going to be affecting my
2 credit score. And 90 days is a long time just to correct
3 information that's already been presented to them.
4 What's really ridiculous about the insurance --
5 the auto insurance thing is, you apply --you do a price
6 quote for insurance, that goes on your credit. But once
7 you get the insurance policy, that's not on your credit
8 report.
9
(Laughter.)
10 MR. BATRA: So it's strange. Asking Geico how
11 much they're going to charge you for insurance shows up
12 as an inquiry, but your insurance policy doesn't. So it
13 just completely makes absolutely no sense to me. And I
14 think that's a glaring example that can be fixed rather
15 easily. All they have to do is code an inquiry versus an
16 application for credit. Did you actually apply for
17 credit, or did you just ask them what the price would be.
18 And there is a big difference between the two.
19 And that would be a simple coding in any formula. It's
20 just these formulas seem to be so rigid and they're
21 probably very much --they're very long. Not a simple
22 thing to do and very expensive, but something that needs
23 to be done. And that's a very simple thing I think they

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
114.
114 Page 115 116
115
1 should do, and rather quickly.
2 But these are a few of the things I've come
3 across in cleaning up my reports. And the concept of
4 credit scoring is foreign to me and to everyone I've
5 talked to at work, my friends, my roommates and my
6 family. No one has heard of it. Everyone is like, I pay
7 my bills on time. I have no credit card debt. Well,
8 there's one more step there that you don't know about.
9 And maybe blissful ignorance is good if you have
10 a good credit score. You never even knew that there was
11 such a thing. I have friends who have bought houses.
12 They never knew about a credit score. That's all right.
13 I mean, everything worked out okay for them. Maybe it
14 was fine. But it just should be a little more brought
15 forth to the public. Let them know this thing exists.
16 It's a tool. It's only a tool, but some
17 mortgage brokers are using it as more than a tool. It's
18 kind of a lazy way out to say here's a number, yes or no.
19 Does it meet our criteria, yes or no. And it should be
20 just a tool and I think it's being abused a bit. And I
21 think some education in the lending sector and the
22 consumer sector is really necessary.
23 Thank you.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
115.
115 Page 116 117
116
1 MS. TWOHIG: Thank you, Simi. I think
2 everybody should know that Simi took time away from his
3 very demanding professional job to come and talk about
4 his experiences, so he did not have the benefit of
5 hearing what Pete McCorkell said this morning about
6 inquiries and how they're counted and not counted. But I
7 think actually this is almost better, because he can say
8 what his impression is from he just knew as a general
9 member of the public, or didn't know or understand, or
10 does not understand about credit scoring.
11 Moving to our next panelist, Caryl Iseman is the
12 owner of Action Mortgage Group in San Diego, California.
13 Caryl has 26 years of experience in real estate and
14 mortgage financing, concentrating on finance in the past
15 nine years. In the process of helping borrowers attain a
16 home loan or refinance their current loans, and in
17 counselling them and helping them with their credit
18 problems, Caryl has encountered several issues with
19 credit scoring that she will identify for us today.
20 MS. ISEMAN: I, too, have to stand up. I spent
21 a lot of time sitting on the phone flying from San Diego.
22 So forgive me. I tend to move around a little bit, too.
23 And I want to say thank you very much to the Federal

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
116.
116 Page 117 118
117
1 Trade Commission for putting on this panel. Because what
2 my hope is, is that at the end of this day with the
3 presentations that are being made, that we will walk out
4 with the benefit of knowing (1) what credit scoring is,
5 but (2) with potentially some resolutions.
6 The system is broke. I am sorry to say. But I
7 don't believe in throwing out the baby with the bath
8 water. We do have to finally decide what we're going to
9 do to make this a fair process so that Mr. Batra and
10 thousands and thousands of consumers like him do not have
11 to go through the process that he has gone through.
12 And by the way, when Mr. McCorkell was talking,
13 I was thinking, gosh --you know, I know that what he was
14 talking about was consumer lending. And I thought, boy,
15 I must have a really high credit score. I'm over 50.
16 I've been in my residence for --I own my residence.
17 I've been there over seven years. I only have a few
18 credit lines. I have low balances. I should have a very
19 high credit score.
20 But unfortunately I don't. And part of my --
21 well, it depends on what day I pull the credit. Because,
22 see, my ex-husband tends to marry women with names very
23 similar to mine. And I was wife three. My name is

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
117.
117 Page 118 119
118
1 Caryl. Wife four's name was Carmen, and Carmen's credit
2 always shows up on my report. And that's good old
3 Experian. I do have to tell you that is not Trans Union
4 and Equifax, but it does happen with Experian.
5 And by the way, I think we need to also
6 understand that there are three scoring models.
7 Everybody has heard FICO. And if I understand right by
8 what I've been told, Fair, Isaac does do the scoring
9 models for all three of these repositories that we're
10 most familiar with. One which is Equifax, the second
11 being Trans Union and the third being Experian, which we
12 all know was TRW.
13 Experian uses FICO. Equifax uses Beacon and
14 Trans Union uses Empirica. So when you get a credit
15 report --and by the way, all the credit reports I have
16 with me and a couple that I will talk about, are with the
17 okay from my clients, who said, go get them, Caryl.
18
(Laughter.)
19 MS. ISEMAN: Clients that are paying higher
20 interest rates than they should because the credit
21 bureaus wouldn't do their jobs. But you'll see that
22 there are three scoring models on each of these reports.
23 And as a mortgage broker, I get to pull the reports. I

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
118.
118 Page 119 120
119
1 get to see the credit scores. And, yes, my clients do,
2 too, because I give them a copy, even though the credit
3 scores are on there. And technically I understand that I
4 can do that.
5 So let me get to the issues of what I want to
6 cover, and I only have 15 minutes so I'm going to have to
7 talk fast. What you heard this morning was great, and I
8 really do believe that Fannie and Freddie are trying to
9 help the home buyers. But you know what? This morning
10 what you heard was the abstract. What you heard Mr.
11 Batra mention is the reality, and what I'm going to talk
12 about is the reality.
13 I'm going to talk about what was mentioned this
14 morning, number one, when they said, you know what? We
15 take anything through Desktop Underwriting. It doesn't
16 matter what the credit score is. You put it through and
17 we're going to try and get it through. Well, that's
18 great. The problem is, I'm a lender --I'm the mortgage
19 broker, and I have to go to the lender to do the Desktop
20 Underwriting.
21 Now, this is one of my lenders. This is the
22 form. This is my cover form that I extend through for
23 Desktop Underwriting. And one of the things that you're

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
119.
119 Page 120 121
120
1 going to see on here, the third box down says, FICO
2 score. Now, what they really mean is not a FICO score.
3 It's the middle of the three scores. And whether that be
4 Empirica or Beacon or the Fair, Isaac --either one.
5 Whichever is the middle of the three. Not the average,
6 but the middle of the three. That's what they want me to
7 put on there.
8 Now, right next to that this says, minimum 620.
9 Now, this was dated February of this year. That means if
10 you don't have a 620 medium score --or middle score --
11 don't send it to me. I'm not interested.
12 Now, here's my next one. This is the same
13 lender. By the way, this lender, I use them a lot for my
14 really good clients, because they have the best rates.
15 And I want my clients to have the best rates, and I
16 really, really would like to have most of my clients go
17 to them.
18 FICO score, the same thing, minimum, 640. So
19 from February to May, they said we don't want 620's any
20 more. You know why? Because this lender has good rates,
21 and they're going to skim off the top all the good loans
22 at the expense of the borrowers that have the 620 and the
23 630 and maybe even a little lesser score.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
120.
120 Page 121 122
121
1 And if you think that that's just one, I'm not
2 going to go through them all. But I have different
3 lender sheets --Chase and some other lenders here. I
4 think this is the best, though. Oh, I can't leave
5 Freddie Mac out, by the way. Freddie Mac has got a great
6 program. A hundred percent affordable goals. You heard
7 Freddie Mac say, oh, we want to look at them all. It's a
8 great loan. This is dedicated to low income borrowers,
9 minimum credit score, 660.
10 Okay. And then --oh, we don't deal with
11 professions or anything. Now, I understand --I spoke to
12 Fannie and Freddie today. I have to find it in here. I
13 know I'll find it. But if you're a doctor, please step
14 forward. First Union. Is there a doctor in the house?
15 if you're a doctor, they came out with a special loan for
16 you. You don't even have to have a 700 credit score.
17 You can do this with a 680. Or you can even be a D. O.
18 But we don't pay attention to any of these
19 things. Now --I mean, what you heard this morning,
20 let's look at everybody. We don't care. We don't pay
21 attention to minority people. We don't pay --we feel
22 everybody is on the same boat.
23 I didn't even realize this was going to be out

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
121.
121 Page 122 123
122
1 there today, but any of you that picked it up, turn to
2 page 27. On page 27 it says, who has riskier scores.
3 Now, what I have underlined here said, borrowers with
4 high risk scores particularly in light of various studies
5 indicating that minority families tend to experience
6 greater credit difficulties.
7 African American borrowers, for example, were
8 about three times as likely to have high risk credit
9 bureau scores, defined as FICO scores, below 620 as were
10 white borrowers, based on Freddie Mac's 1994 mortgage
11 purchases. And Hispanic borrowers were about twice as
12 likely as white borrowers to have high risk scores. But
13 excuse me. Let's ignore this. We're color blind. It
14 doesn't matter.
15 Let's get to talk a little bit about the
16 inaccuracies, because this is really where the problem
17 lies. Where you have mistakes on a credit report and the
18 borrower cannot get it off. And how prevalent are these
19 mistakes? I will tell you that of every ten credit
20 reports I pull, seven of them are incorrect. And the
21 other three we may not pay much attention to, because the
22 borrower had high credit scores on it and so we didn't go
23 through the report.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
122.
122 Page 123 124
123
1 But even when you do --and I have one here, and
2 my borrower's average mid score is like a 780. And even
3 he has two incorrect credit lines that are still showing
4 on there that showed --you know, that were paid off.
5 They're still showing balances. But he has such a high
6 credit score they don't care.
7 This was an article that appeared Friday, March
8 13, 1998. This was done by the Public Interest Research
9 Group. And again, I'm not going to go into the specifics
10 on the numbers here, but they point out all the
11 inaccuracies on the credit report. But there are two,
12 though, that I have to mention.
13 Nineteen percent --and these were the 133
14 consumer credit bureau reports. Nineteen percent
15 contained accounts that could not be identified or did
16 not belong to the consumer. Twenty six percent contained
17 credit accounts that had been closed by the consumer but
18 that were listed as open. And three percent had
19 incorrect Social Security numbers. That was a line that
20 was just total strangers. I know that feeling.
21 The inaccuracies are what effects the credit
22 scores. So that if I have mistakes on someone's credit
23 report that is pulling down their credit score, and they

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
123.
123 Page 124 125
124
1 have found this great house, and they have not 90 days to
2 close this transaction but 30 days to close that
3 transaction, it's only going to be by an act of God that
4 we're going to get it done. And my clients work with me.
5 God bless them. They follow everything I tell them.
6 They send return receipt requested letters.
7 I found phone numbers, by the way, where I
8 actually get a live person. Of course, you can't be sure
9 of what they're going to tell you. Because a gentleman
10 and I was on the phone --he didn't know I was on the
11 phone --from Equifax told my borrower --and this was
12 just a week ago --that Equifax had nothing to do with
13 the credit scores, that each creditor that submitted
14 their credit accounts decided the credit score. Of
15 course I had to hit the mute button at that point in time
16 and kind of let him know that I was listening and that he
17 needed to stop lying to my client.
18 I have to share with you where we get the
19 greatest inaccuracies and where it really hurts people
20 the most: medical, bankruptcies, medical collections,
21 people that have had catastrophic illnesses. Of course,
22 based on the predictability --that is, of course,
23 unpredictable and inconsistent. Based on this

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
124.
124 Page 125 126
125
1 predictability these people, because they've got these
2 low credit scores because they had to file bankruptcy,
3 out the window with them.
4 What do you mean? Because they had a
5 catastrophic illness, they're at 20 percent that their
6 share was maybe of their insurance policy, of maybe 50,
7 60 or a hundred or $200,000, and they could not pay it
8 and they were forced into bankruptcy. People that are
9 hard working people that had never paid a bill late in
10 their life ever. Never had a mortgage payment late if
11 they did own a home. They are so now unable to get the A
12 paper for a finance loan.
13 Now, can I get them a loan? Yes, I can get them
14 a loan at two percent or three percent higher than
15 anybody else. And I can get some of the stuff off when
16 they have filed that bankruptcy. Guess what? It still
17 shows as open as collections. And again, anyone who
18 wants to see this, I will show this.
19 I have one gal here, right in the process right
20 now of her loan. Her bankruptcy, which was partly
21 medical and partly not, was discharged 8/ 16/ 94. We are
22 in just about to be August 1999. She still has three
23 items that are showing up as unpaid collections. Guess

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
125.
125 Page 126 127
126
1 what happened to her credit score? Guess what happened
2 to the loan that we have to close in two weeks? I'm
3 getting her a loan, two and a half percent higher than
4 what she should be getting.
5 So all the savings that Fannie and Freddie spoke
6 to you about, that's great. But do you know at whose
7 expense those savings are? They are at the person who
8 could not get rid of his inquiries. They are at the
9 expense of the people that are going through catastrophic
10 illnesses and had to file medical bankruptcies.
11 And I want to give you another one. You know,
12 supposedly it's not only --are we supposed to look at
13 someone predicting the future and how they're going to
14 pay, because this is what we're looking at. Right
15 Fannie? You want to know you're going to get paid. And
16 the same thing with Freddie Mac. So they look at how
17 somebody paid their previous bills.
18 Well, what about with downsizing. How many
19 people have been put out of work? I know we had a lot of
20 that happening in San Diego. I know that even in this
21 great economy, you have people that get unemployed, and
22 they may have to go for three or four months where they
23 become a little slow paying. But for the 30 years

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
126.
126 Page 127 128
127
1 before, they never had a late payment, ever. So for
2 three or four months, they have a car payment that's
3 maybe 60 days later or whatever. Are these people bad
4 risks if they want to go and buy a house right now? The
5 guy is back working.
6 I have that. I have an airline pilot. And by
7 the way, not in San Diego, in Michigan. He was out of
8 work. He flew for 30 years. They have a great credit
9 record, because as Mr. McCorkell told you, all the good
10 stuff does remain. It stays on there. You look at it.
11 They had a house. I mean everything. Never a late
12 payment. He was out of work for six months. They had a
13 60 day car payment --60 days late on a car payment and
14 two credit cards. Two credit cards that had 30 day
15 lates.
16 And I have to tell you where their credit scores
17 are, because it's recent. Anything recent takes really
18 from that credit, and they have to do a refinance on
19 another property that they own. We're at a very, very
20 low loan to value. But because it's nonowner occupied
21 and they want to do the loan to fix the house up because
22 the tenant just died, I have to tell you, they're a B
23 paper loan. And should they be 54 percent loan to value,

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
127.
127 Page 128 129
128
1 someone who for 30 years never missed a late payment?
2 But the predictability, these people are going
3 to be lousy borrowers. They're not going to pay back,
4 because they have a 610, I think from one of the bureaus,
5 and a 600 --oh, no. They're not going to make the
6 payback. That's why I have difficulty with the setup
7 with the credit scoring.
8 Now, I said before, do we throw the baby out
9 with the bath water? No. Could credit scoring be a tool
10 and a useful tool? Absolutely. Because people who
11 really don't fall into any problems and keep making their
12 payments on time, they have a really, really good payment
13 history. But I don't need a statistic to tell me that.
14 Any underwriter could look at that type of file and say,
15 hey, I can approve it. I don't need someone to give me a
16 credit score.
17 But let's say so they have really good credit
18 scores, and because of that we can move real quickly.
19 Well, that's great, because that person maybe should be
20 compensated and have really a loan move through the
21 system quickly and should have a little better interest
22 rate. I have no problem with it.
23 But the process as it is right now has got to be

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
128.
128 Page 129 130
129
1 changed. The scoring models have to allow for some
2 flexibility. It has to take into account medical
3 bankruptcies or that type of concern, or the short term
4 unemployment. There's got to be some flexibility. Right
5 now there is absolutely none, and that's why I have a
6 problem with the system.
7 These are my suggestions --because I know I
8 have to cut myself short. Number one, everyone in this
9 room that is either representing a regulatory or a credit
10 agency, or represents a lender, or whoever you represent,
11 if you are part of this process, we must work together to
12 get this worked out.
13 Now, there are three bureaus that are going to
14 go kicking and screaming to work this out, and that's
15 Trans Union, Equifax --and I know there's a gentleman
16 here who is representing the three bureaus. I don't know
17 if anyone from those --from the three repositories are
18 here. But you've got to clean up your act. It shouldn't
19 take 30 days for a bureau --for a mistake --an
20 inaccuracy --on a credit report to be done, not in this
21 day and age of technology.
22 And also creditors have to start reporting more
23 promptly and correctly. And when they increase someone's

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
129.
129 Page 130 131
130
1 credit balance so that there will be a greater disparity
2 between what the client owes and what their balance is,
3 they have to report that, and it has to be not three
4 months old. It has to be current, so that we're on par
5 with everything.
6 And inquiries, unless --not only if somebody
7 even checks for an inquiry and, you know, was really
8 sincerely checking on it. But only --only --if an
9 account came as a result of that should that be counted.
10 And why do we deal with this mortgage thing and 30 days
11 of shopping for a car? If someone is shopping for a
12 mortgage, isn't it their entitlement to get the best rate
13 and find the best mortgage? And it might take them 60
14 days and it may take them a hundred calls.
15 And they should have the right to do that and
16 not be penalized, which they are now. If it takes them
17 more than 60 days and it carries over, and they have five
18 or six new inquiries, guess what. Their credit score
19 goes down for trying to do for themselves the best job.
20 And the same thing with credit cards. Don't you
21 want to get the best interest rate on your credit cards?
22 Shouldn't you have the right to shop around. Well, don't
23 do it anywhere near the time that you're thinking of

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
130.
130 Page 131 132
131
1 doing a refinance or buying a home. And you should have
2 the right, and you should have the right to consolidate
3 all your monies. You should have the right to do that,
4 and only if a new account results as an effect of that,
5 that is the only time that inquiry should be counted.
6 And I think we all really need to start paying
7 attention to this. I know we're going to have a question
8 and answer session, and I know that a lot of other people
9 have things to say, so I'm going to thank you all for
10 listening to what I've had to say. I'll be glad to
11 answer any questions when they come up later.
12 Thank you.
13
(Applause.)
14 MS. TWOHIG: Thank you very much, Caryl. Our
15 next panelist is Marcia Griffin, who is founder and
16 President of Home Free USA, and the former Vice President
17 of a mortgage servicing firm. Home Free USA is a HUD
18 approved, nonprofit membership organization that promotes
19 home ownership and assists perspective home buyers with
20 such things as education and counseling on the home
21 buying process, credit counseling and repair, and offers
22 other guidance.
23 And we welcome Marcia to give us some comments

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
131.
131 Page 132 133
132
1 also.
2 MS. GRIFFIN: Thank you. Thank you. Well, I'm
3 going to just give a different perspective here, if I
4 might. I represent hundreds and thousands. Last year we
5 interacted with over 10,000 perspective homeowners. One
6 of the issues that I think that we're --as we discuss
7 credit scoring today, I want you to know that we use
8 credit scoring every day, and we use it in the first
9 meeting that we have with every homeowner --every
10 perspective homeowner.
11 One of the issues that I would like to also
12 bring up is that through education of the credit scoring
13 issue, education of how lenders work, what Freddie and
14 Fannie are looking for, the education is key. Education
15 to the mortgage industry. Education to the nonprofits.
16 And education to the perspective homeowners, which is
17 what we're attempting to do here.
18 We have had good and bad results with the credit
19 score. You know --and I must say, though, and let me
20 just give you a little bit of sort of background. In
21 Home Free we do a lot of education. We have anywhere
22 between eight and 21 classes a month, because people need
23 to know. They need to know their credit score, and they

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
132.
132 Page 133 134
133
1 need to know it early from our advantage point. They
2 need to know what they need to do, how their credit needs
3 to be cleaned up. They need to know the good, the bad
4 and the ugly before they go to Caryl.
5 From our advantage point, if people are
6 prepared, they will get approved. In Home Free we have a
7 100 percent approval rate. Every single person that has
8 been through our process. Now, let me just tell you,
9 some people can fly. Others have to walk, crawl or take
10 a boat or train. I mean, it really varies in terms of
11 the time that we spend with them.
12 But the reality is that from our perspective,
13 and through the perspective of our partners --Freddie
14 Mac has been a wonderful partner. We've learned a lot
15 through credit scoring with them. Chevy Chase Bank has
16 been a great partner. We're in partnership with other
17 lenders.
18 But the reality is that it is important that
19 people understand where a credit score positions them and
20 what they need to do. Now, from the education
21 standpoint, the reality is that a lot of the nonprofits -
22 -and we're in the counseling mode --don't --we're not
23 really in sync on what people need to do to make

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
133.
133 Page 134 135
134
1 themselves better, even if it's going to be, you know, a
2 six month period with the nonprofit. This is where the
3 education of the nonprofits need to occur.
4 But in this class that I conducted --in fact, I
5 certainly open an invitation to any of you. I have this
6 class every Thursday night at 6: 30 at my office here in
7 Washington. It's called Ten Home Buying Secrets Everyone
8 Ought to Know.
9 And the first secret is that the mortgage
10 process has changed, that people should not think about
11 home ownership and just, you know, jump out here, go to a
12 realtor, sign a contract and go to lender. They don't
13 know where they stand financially. You know, they know
14 that their credit situation might not be the best. But
15 there is no preparation at all.
16 So from our advantage point, the process has
17 changed. They need to get prepared. People need to get
18 prepared before they go to a lender so that they can be
19 assured of getting the best rates --the most money,
20 because that --the money issue I know we're not talking
21 about this today. But I can tell you, in Home Free USA,
22 65 percent of our people have credit scores of over 620.
23 We deal only with under served populations: minorities,

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
134.
134 Page 135 136
135
1 African Americans, Latino populations, immigrant
2 populations and also whites.
3 So the credit score issue is one factor.
4 Clearly, the other factor, certainly in our environment,
5 is that we have people with a 700 credit score and no
6 cash. I mean, this is a big issue that we've got to deal
7 with.
8 But let me just get back to my class for a
9 minute. So the first part --the first secret is that
10 the process has changed. And the second secret is to get
11 your credit score. Get your credit report immediately,
12 not from the credit reporting agencies, because when you
13 write, they don't give you your credit score. They need
14 to really get their credit report with a credit score, so
15 that they can understand where they are. And they need
16 to meet with an organization like Home Free, like the
17 NHS's and, you know, other nonprofits around, and perhaps
18 even lenders, to examine that credit report and begin
19 making these changes early on.
20 Clearly, where Fair, Isaac, Freddie and Fannie
21 need to come to the table, we need to clearly understand
22 what the impact is if, you know, a person's credit score
23 --let's say --and I've heard several things from Fair,

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
135.
135 Page 136 137
136
1 Isaac. You can't have, you know, too many credit cards,
2 and they can't be up to the limit.
3 Well, you know, we need to have some clear
4 factors. In fact, I would love to have someone to
5 produce just a little report, something like 30 essential
6 things to know about your credit score. Things that you
7 --I mean, and basically how you can improve it. You
8 know, what's causing the problems. I mean, there seems
9 to be a little bit of question about this across the
10 board.
11 So from our advantage point, clearly education
12 is the key. We use the credit score at the very
13 beginning in what we call our mortgage profile. I mean,
14 that is the first one on one meeting that we have with
15 the person. We pull a three merge credit report, and we
16 see the credit score, and we work with that. And we have
17 seen that, you know, when they go to the lender, that
18 credit score, you know, has improved.
19 The issues, though, are that, you know, what's
20 going to make it improve significantly and what's holding
21 it back. It's a very interesting part that you brought
22 up here, Caryl, about, you know, the inaccuracies. Well,
23 we know the inaccuracies with a credit report. But, you

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
136.
136 Page 137 138
137
1 know, we haven't run across, you know, many issues that
2 sort of relate to how we could tell --I mean, the
3 question about the scoring. I mean, you can't really --
4 no one has a feel for how these scores --how they even
5 come up with the scores. So it's difficult to really
6 assess whether the score is right or wrong or just
7 whatever.
8 But I can tell you that, you know, there are
9 people who certainly when they have a credit score of 500
10 --when we see a credit score of in the 400's, we tell
11 them to just stop. There is nothing more --stop. Hold
12 it. Let us take your hand and walk you through a
13 process. Because while there may be some errors there,
14 we doubt if the credit score is showing up at 490 that
15 it's going to --there is enough error in there to make
16 it a 700 credit score.
17 Where we are finding some issues, however, is
18 right now we have --one of our members has a 710 credit
19 score. Now, this is pretty good, you know. She has a
20 710 credit score, and the lender is having question after
21 question after question about this loan. Clearly --we
22 talked earlier about the Desktop Underwriter and the Loan
23 Prospector. We have asked Freddie Mac to give us the

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
137.
137 Page 138 139
138
1 Loan Prospector, because we have found that when several
2 of our members have gone to lenders and they've begun
3 questioning certain aspects of whether that loan should
4 be approved, then they've gone --we've sent them to
5 another lender and the loan has gotten approved using
6 these technological methods here with Loan Prospector and
7 Desktop Underwriter.
8 I'm sort of betwixt and between, as I said, you
9 know, as I represent under served people, as I represent
10 home buyers --future home buyers --and as far as the
11 credit scoring issue, I think it's a great tool. Quite
12 frankly, I think we need to be better informed and better
13 educated about what's going to make the difference in
14 enhancing that credit score in our environment.
15 At Home Free USA we have on average between 45
16 and 150 people per class. I mean, this is --and we have
17 three classes a week. People have a thirst for
18 knowledge. Under served people in this country are
19 trying to do better. We preach to them, through churches
20 and through employers and through, you know, unions and
21 the like, that look, we have got to go into the next
22 millennium better off than we have been. People want to
23 be informed. They want to get their credit score up.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
138.
138 Page 139 140
139
1 But we just need to have consensus about what
2 it's going to take and what people need to do. And it
3 doesn't need to change every five minutes. But what they
4 need to do and basically how we can all be on the same
5 level so that we can increase home ownership. People
6 want to become homeowners.
7 And, you know, there are a lot of people that
8 are pretty bad off, but even those people --you know,
9 what we tell our people is that despite where you are
10 today, you can be better off tomorrow. A misty morning
11 does not signify a cloudy day. You come on and we're
12 going to work with you with that credit score, with the
13 credit, with these statements and everything else.
14 You know, it is from that advantage point that I
15 sort of represent other nonprofits who work with the
16 people, who if we can get this thing together, by the
17 time the people get to Caryl and all of the mortgage
18 lenders in here, the loans will be right. We won't have
19 to send them to B, C and D paper lenders, which I think
20 is the worse thing of the world.
21 And certainly when they get to Freddie and
22 Fannie through counselling, through education and also
23 pulse purchase counselling --you need to stay with these

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
139.
139 Page 140 141
140
1 people for at least two years --at least at Home Free
2 this is what we do --and keep them with you and teach
3 them what they need to do to better themselves
4 financially. Everybody wants to be better, and this is
5 what the objective of, I think, all of us would be.
6 Thank you very much.
7
(Applause.)
8 MS. TWOHIG: Thank you very much, Marcia. We're
9 running a little late, but as promised we would like to
10 give the audience a chance to raise some questions and
11 issues so that we can be sure to cover them this
12 afternoon.
13 So why don't we take a few minutes and do that
14 now. And there are hand held microphones, so if you
15 raise your hand you can be given the microphone. And
16 those of you in the overflow room, if you have questions
17 at this point, if you come to the main room we can try to
18 get your questions and comments in.
19 MR. PREUDHOMME: Hi, can you hear me?
20 MS. TWOHIG: Yes.
21 MR. PREUDHOMME: Okay.
22 MS. TWOHIG: And please first identify yourself
23 and identify any affiliation you have.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
140.
140 Page 141 142
141
1 MR. PREUDHOMME: Yes. Hi. My name is Bruce
2 Preudhomme. I'm here not only representing myself and my
3 family --okay. Somebody wants the spelling of my name.
4 It's Bruce Preudhomme. That's P-R-E-U-D-H-O-M-M-E.
5 As I said, I'm not only representing myself and
6 my family, I'm also representing two consumer activist
7 organizations I work closely with. One is called the
8 Victims of Chevy Chase Bank. Sorry.
9
(Laughter.)
10 MS. TWOHIG: Okay.
11 MR. PREUDHOMME: And the second is the Economic
12 Justice Alliance. Okay. In both cases what we're
13 dealing with has to do with credit scoring, credit
14 reporting and problems centered around that whole arena.
15 The biggest problem I see --well, there's really two big
16 problems. The data, as has already been mentioned. In
17 my field of computers, we have a term called garbage in,
18 garbage out.
19 Okay. And we've got that going on big time in
20 the credit bureaus. The data they get in is garbage. It
21 comes from the lender. It can actually be used as a tool
22 to attack somebody that has been a vocal person. I for
23 one --after closing an account I paid off, I had a

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
141.
141 Page 142 143
142
1 lender a year after it was closed, report me as 60 days
2 delinquent, which I had to battle and go through the
3 whole 90 day process. I just got my letter the other day
4 saying okay, yeah, you were right. It's not that. And
5 then they put in some other verbiage, okay, which is
6 another pack of lies that I have to battle.
7 The data as far as inquiries is totally bogus,
8 too. Back in 96 I had my credit rate increased on my
9 gold MasterCard from 18 percent to 25.99 percent. Not
10 because I was paying late. Not because I was paying
11 slow. But because of information contained in my credit
12 bureau. And one of the big factors was the number of
13 inquiries.
14 Well, I checked. The number of inquiries that I
15 had during the last year was nine, okay. It happened to
16 be that year I bought two car phones. And because I run
17 an on-line site where I dial up access, I got another
18 four phone lines in the house. Most of those credit
19 inquiries were for phones. Now, I don't know why that
20 would affect me. Why that would justify Chevy Chase Bank
21 raising my interest rate, but they did, okay.
22 I battled this and in --you know, it took I
23 guess two years to go through the process, trying to go

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
142.
142 Page 143 144
143
1 through the system. And, you know, in every step, you
2 know, I contacted the FTC. I contacted the OTS. You
3 know, they didn't want to deal with it. They would
4 rather just say, oh, well, you know, do a better job at
5 paying your bills. I pay them off, you know, so that
6 wasn't the problem. Then they say, okay, well, maybe you
7 should get an attorney. And I have a real problem that I
8 have to get an attorney to get the laws in this country
9 enforced, and to get the laws in this country done right.
10 Now, back to what I was saying, there are two
11 basic problems. Data, as I said, is one. The other is
12 the whole credit system. We've got the cart before the
13 horse. We have lenders funneling garbage data into the
14 credit bureaus. Instead, that data should come to the
15 consumer first. You know, the consumer should have the
16 right to say, hey, wait a minute, this is totally bogus.
17 I should be granted permission to say hey, this
18 never happened before it gets posted. I should get a 60,
19 90 or 120 day period where I can clean it up before it
20 gets posted, rather than waiting until after it's posted
21 and I happen to stumble upon it, you know, and that's the
22 way it is right now.
23 So basically one of the premises that this

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
143.
143 Page 144 145
144
1 country was founded on that you're innocent until proven
2 guilty, the reverse is happening. You are presumed
3 guilty, okay. There's a lot more that I could say, but I
4 want to give other people a chance to talk.
5 MS. TWOHIG: That would be great.
6 MR. PREUDHOMME: Thank you for listening.
7
(Laughter.)
8 MS. TWOHIG: Are there any --no, we appreciate
9 your comments, but it would be good if people could
10 somewhat limit their comments. Does anyone else want to
11 raise questions or make a comment at this point? Again,
12 we're just trying to get issues out on the table so we
13 can maybe get some responses and answers this afternoon.
14 MR. LE FEBVRE: Do you need my name and
15 everything first?
16 MS. TWOHIG: Yes. Please identify yourself.
17 MR. LE FEBVRE: Richard Le Febvre. Affiliation
18 is AAA American Credit Bureau. I'm here for --
19 MS. TWOHIG: Oh, could you spell your name for
20 the record?
21 MR. LE FEBVRE: Le Febvre, L-E-F-E-B-V-R-E.
22 We're a mortgage credit reporting bureau, and we service
23 most of the United States. And in listening to Fannie

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
144.
144 Page 145 146
145
1 and Freddie and Fair, Isaac, we're --and the
2 repositories --we're in the trenches every single day.
3 When a consumer goes to Caryl or anybody else that is a
4 mortgage broker, we're actually pulling the report and
5 actually analyzing it and seeing it on a daily basis.
6 And, of course, the repositories take a lot of
7 the blame. Every consumer group from the legal side to
8 the consumer groups blame them. But a lot of the times,
9 in my own personal experience and in the reports that we
10 do each and every day, it's the creditor that's reporting
11 that data. And one of the things that I feel real
12 strongly about is back in 72 when the Congress enacted
13 the Fair Credit Reporting Act and then amended it again
14 in 97, they knew at that particular time with no blame
15 that the data was wrong. There were creditors
16 misreporting and there are certain things are there.
17 And then DU and LP come along. And my biggest
18 concern is if Congress passed the Fair Credit Reporting
19 Act knowing there were inaccuracies and gave consumers
20 certain rights, why does Fannie and Freddie don't think
21 that they have to give that consumer those same rights as
22 it stands right now.
23 In the old days what would happen if Mr. Batra

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
145.
145 Page 146 147
146
1 had gotten his credit report from us, a credit report is
2 generated. He's sent a consumer copy. He then has the
3 right to dispute that data. Now, because of the 30 day
4 turnaround, the repositories cannot get that data
5 verified that quickly. That's why mortgage credit
6 reporting companies came into effect, because they can
7 represent or protect both consumer and lender to verify
8 that data when Mr. Consumer says it's inaccurate.
9 Now, in the system --Fannie Mae, to get on
10 their system, comes to me and asks for $75,000 if I
11 could, but I can't. Then they ask for transaction fees
12 to come through a system that they don't give the
13 consumer the same right. My proposal, and one of my
14 biggest issues is, if that data is wrong, Fannie and
15 Freddie must then allow third party intervention, either
16 by the repository level or by a mortgage credit reporting
17 company that's out as a third party to verify that data
18 to show it.
19 A lot of credit --now that Fair, Isaac uses a
20 scoring model, a lot of creditors --and if you talk to
21 the repositories, all creditors don't report the same
22 way. High credit to one creditor is different than high
23 credit to another. And if that model is looking at the

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
146.
146 Page 147 148
147
1 balance --the high income ratio, and if you have a
2 Discover Card or an American Express card, every single
3 month you're at your limit, even though you're truly not.
4 And there must be a mechanism in place that both
5 Fannie and Freddie will allow third party intervention
6 and to then correct that file and then send it back
7 through. LP and DU is a great operating system. One, I
8 don't think I need to pay to get on their system because
9 mortgage companies like to use my company or Credco --I
10 know they're represented here --and numerous other
11 companies. But there has to be a way that the consumer
12 and both Fannie and Freddie will then allow third party
13 intervention, and then accept that data, which in turn
14 then will recalculate the score.
15 Right now to talk to Equifax --Equifax right
16 now, out of the three national repositories, has what
17 they call an expedite program. That means, I'm a
18 mortgage credit reporting company, and if a consumer or
19 Caryl or anybody else in this room has inaccurate data
20 that we're able to prove, they then send me a document.
21 We at the mortgage credit reporting company authenticate
22 that document, send it to Equifax, and Equifax is
23 committed to correcting the file within 24 to 48 hours.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
147.
147 Page 148 149
148
1 Which in turn --then we pull the file, which in turn
2 recalculates the score.
3 The other two bureaus have not done it yet. And
4 I think if the repositories working together with
5 mortgage credit reporting companies --and I'm sure there
6 has to be a code of ethics. There has to be certain
7 fraudulently --something in place that they don't
8 receive fraudulently done documents.
9 But my proposal would be, if the document is
10 then faxed from the consumer --if Mr. Batra was able to
11 verify that Discover Card misreported his credit card and
12 received a letter from Discover Card, then it would be
13 sent to my company. We would authenticate that letter,
14 by calling the person --the author of that letter --and
15 then forwarding that on to the repositories, and having a
16 relationship between the repositories and us, so that
17 data they feel comfortable correcting.
18 And that's the system that needs to be in place,
19 because when you talk to everybody, they're upset about
20 the credit scoring system. They're upset with the
21 repositories. And I'll be honest with you, the
22 repositories --all they are is an acceptance of data.
23 It is impossible for them to turn around and look at

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
148.
148 Page 149 150
149
1 every trade line that's reported to them every single
2 month and know if it's right or wrong.
3 The only way the repositories know if it's right
4 or wrong is if the creditor corrects the file, either by
5 magnetic tape or any other means, or if the consumer
6 disputes. So a lot of the pressure is put on the
7 repositories, and they're not the problem.
8 We had a file a couple of weeks ago where we
9 were dealing with BankOne, and BankOne reported a
10 gentleman's perfect BankOne account as a Chapter 7
11 bankruptcy. It decreased his scores by over 70 points.
12 And the only one of the repositories --in dealing with
13 BankOne, he received a dispute letter that said it would
14 be corrected in 30 days. This is now 90 days later and
15 it still hasn't been corrected.
16 At the same time, three of his credit card
17 vendors had notified him that the increase in his credit
18 report went from 9.9 to 22 and a half. And he never had
19 a Chapter 7 bankruptcy at all. He always paid as agreed.
20 So the repositories themselves, I believe, need to lead
21 the front in dealing with third party intervention to get
22 this stuff.
23 And one other quick example, and my biggest

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
149.
149 Page 150 151
150
1 concern, is discrimination towards Hispanic or minority,
2 blacks, whites or whatever it may be. We did a file the
3 other day, and I analyzed it. And a file was pulled by
4 my company on 6/ 7 and then it went through the DU system
5 on 7/ 9, which is one month and two days later.
6 I analyzed the raw credit from all three
7 repositories. Equifax had four trade lines, TU had five
8 and Experian had seven. The only thing this gentleman
9 did was miss a $20 monthly payment to Discover. And of
10 course he disputed it. He said he was not late. But
11 Experian increased his --not Experian. Fair, Isaac
12 showed his score 76 points higher with Experian, 151
13 deduction with TU, and 178 deduction with Equifax. But
14 he paid his mortgage --and he only had one derogatory
15 item on his credit file, and it was a $20 payment. And
16 he's had a mortgage history for three and a half years
17 that was paid on time.
18 So my analysis of it all, since we're in the
19 trenches each and every day, is that something needs to
20 be done, and it needs to be done drastically. Because
21 right now about 80 percent of our incoming phone calls
22 are why is my score this, and how do I increase it, and
23 so on and so forth.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
150.
150 Page 151 152
151
1 Thank you.
2 MS. TWOHIG: Thank you very much for your
3 comments. I think it's important as we think about these
4 issues to think of something --I think Pete McCorkell
5 first suggested this afternoon. That we also have to
6 think of what issues are we talking about that have to do
7 with credit scoring, and what are just the credit
8 processes that would also exist in a judgmental decision.
9 As I understand it, your first comment that you
10 made did have to do with the speed of automation, that
11 makes it more difficult for the consumer, necessarily in
12 the course of that decision making process, maybe at that
13 same time to dispute errors.
14 Are there other questions or comments? Yeah.
15 MR. SMITH: I'm Don Smith. I'm the President of
16 Great Arizona Mortgage Corp. in Scottsdale, Arizona. And
17 this is my first visit to Washington, and what a great
18 town. And I want to thank you for having me at the
19 hearing today.
20 With regards to why I'm here, those two loans
21 that Richard just brought up were loans that were done in
22 my office. The second one was a Hispanic couple that had
23 700 and some change in their FICO scores, and they're now

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
151.
151 Page 152 153
152
1 500 and something in less than 30 days, and there is no
2 difference in the file. Try to figure that out and
3 explain that to a borrower that barely speaks English,
4 why when I told him 30 days ago he would have no problem
5 whatsoever. He's made arrangements for his move. And
6 before I came out here, I had to look him in the eye and
7 say, we may be having a problem.
8 Now, of course, if you listen to Fannie Mae and
9 Freddie Mac, they explained to you that there are ways
10 that we can go about this. We can expedite it. There
11 are no changes on the report. The numbers have shifted.
12 As a lender --I've been doing this 19 years as a
13 mortgage lender. I've been a realtor for 22 years. I
14 have done thousands of FHA loans, VAs and conventionals.
15 And I love what we've done with the automated
16 system. Unfortunately, the premise that this is based
17 upon is based upon a perfect world or in a vacuum. We
18 are not living in a perfect world. And we need to take
19 the time as we talked about --you know, let's talk about
20 what makes up the scoring system with Fair, Isaac. Let's
21 talk about if you have a $1,500 limit on your credit
22 card, that it says you only owe $200 and it's a $1,500
23 limit.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
152.
152 Page 153 154
153
1 But we've got numerous credit reports that I'm
2 sure that everybody in the room has seen over and over,
3 where it says high balance is zero and you owe $1,500.
4 How do they let you do that? I want to know how you do
5 that. How do you borrow $1,500 more than your high
6 limit? It's calculating under scores. And if it's not,
7 I want somebody here to tell me that that's not how it's
8 being done, so I can go back and tell my clients, the
9 ones that I see on a consistent basis, that look, this is
10 how we work it.
11 This is what's important. Instead of saying
12 well, you know, if you pay your bills on time, if you go
13 to church every week and you pay tithe, you do this and
14 you do that, God is going to take good care of you and
15 don't worry about Fair, Isaac. That is not the case.
16 What we need to take care of, is let's kind of get down
17 to it. Let's crunch the nuts and the bolts of this issue
18 and come up with it and say --just like we're talking
19 here with Simi. I don't know. I don't know.
20 You know, it just comes across the paper, and by
21 god, you're either going to get approved or you're not.
22 Unless you take it to an underwriter that knows you and
23 trusts you, and she's willing to step out on a limb and

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
153.
153 Page 154 155
154
1 say, well, I don't give a damn if they're going to buy it
2 or not. You know what? This loan is worth underwriting,
3 and we're going to make the loan.
4 The biggest issue that I've got still is the
5 reporting issues that we have from banks. By the way,
6 that one file, the first one that he mentioned to you,
7 the borrower was buying a $300,000 house in north
8 Scottsdale. It was up to $400,000 before he closed. I
9 had approved him. He had great credit overall going into
10 the transaction.
11 We're getting ready to close on the deal. I
12 said come on back in the office. Let's take a look at
13 it. BankOne --and I'll tell you the name of it, because
14 I've got all the documentation here with me. BankOne
15 rates him on his credit report now as being six times 90
16 days late and bankrupt. His credit score is in the
17 porcelain convenience. The man is livid. His wife had a
18 miscarriage over this period of time. They did not close
19 on the house.
20 I had him come in. I have a letter in the file
21 that I was asked to give you a copy of so you could
22 review it and take a look at it. I have the
23 documentation. In their infinite wisdom, BankOne, when

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
154.
154 Page 155 156
155
1 we called them on the telephone and explained to them
2 that --and we got the automated first. We did the
3 1-800, wait in line and die.
4
(Laughter.)
5 MR. SMITH: We did that and got that
6 information. And you know what it said to me? Never
7 been late. Well, okay, fine. So I hit the O button, the
8 one you're supposed to keep hitting until you get a live
9 person. They got on the telephone. She says, well, Mr.
10 Smith --and I told her who I was and that the client was
11 sitting there. It's a mutual client of ours. And
12 explained to her, your system says that he has paid right
13 on time. She said that's absolutely correct.
14 And I said well, why is it being reported on the
15 credit bureau, and she goes, I have no idea. And I said
16 well, what's your position. Well, I'm the supervisor.
17 How long have you been doing this? Well, what's it to
18 you, is how she came back to me. And I said no, I'm
19 serious. How long have you been doing this? Are you
20 familiar with the Fair Credit Reporting Act? And she
21 says well, yes. And I said are you familiar with it,
22 like have you read it? And she said well, yeah, I'm
23 familiar with it.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
155.
155 Page 156 157
156
1 And I said well, you have an obligation. You
2 need to get me a piece of documentation. We are trying
3 to close on this man's home. I have less than two weeks.
4 This is a surprise. You've already told me that it's not
5 his information. You need to send this information out
6 to me right now. I said, here's his name. You have it.
7 Here's his Social Security number. You need to reference
8 the account. And I need that letter, and you need to fax
9 it to me immediately. I'll give you to Friday. This is
10 Wednesday. It should be easy for you to do. I know you
11 feel bad.
12
(Laughter.)
13 MR. SMITH: I got the letter Friday. And I have
14 a copy of the letter, and I'll be glad to show it to you,
15 because this is what goes on in the real world out there
16 for many of these people out here that are doing these.
17 The letter says: Dear Mr. Smith. Reference the
18 account. Please be apprised we have removed this line
19 item from your credit bureau. For mine. They didn't
20 take it off his. They took it off of mine.
21
(Laughter.)
22 MR. SMITH: She doesn't even know my Social
23 Security number, but I'm sure that all Don Smith's don't

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
156.
156 Page 157 158
157
1 have a BankOne account number. Well, it gets better. I
2 called her back up. I was on that terminal hold thing
3 again. Okay. My client's in the office. He's a big
4 guy. Don't make him mad. Thank God he's a friend. So
5 we're standing there waiting, you know, on hold.
6 We get them back on the line. Noel Simpson, how
7 are you? Who is this? It's Don Smith again. Oh. I
8 noticed the letter that you wrote me. You know, it was a
9 pretty good written letter, but I've got a couple of
10 items I would like to bring up to your attention. You've
11 left out Mr. Acton's name and Social Security number, and
12 you just said that you removed the item.
13 We don't want the item removed. It's his
14 credit. He's had it. He's been paying on it. You've
15 told me that. I want you to take it and state that he's
16 never been late, and he's never been bankrupt, and we are
17 sorry and apologetic. Well, that's going to take a
18 little while. No, Noel, we've been through this. Let's
19 try this again. Get it taken care of.
20 She sent me out another letter. She removed it
21 completely off his credit report. There is no trade
22 line. It doesn't show any payment history. It doesn't
23 show any balance. So we can't figure out if he still

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
157.
157 Page 158 159
158
1 owes the money or not. But guess what's still in the
2 column? He's still showing bankrupt. His credit scores
3 have depreciated and depleted over this period of time.
4 He is now in the mid-500's, when he started --I mean,
5 he's dropped over a hundred points. The man lost over a
6 $100,000 of equity on the house.
7 And I'm going to shut down right now. I see you
8 all looking at the clock. But this is what the rest of
9 us are doing business with out there. This is what we
10 need your help on. Let's put some teeth in the FTC, if
11 you're going to have issues for the credit bureaus with
12 what they have to do, and for the people that provide the
13 information.
14 Thank you very much, and you all have a nice
15 day. A great city.
16
(Laughter.)
17 MS. TWOHIG: You know, when we announced this
18 conference and we started getting feedback, we knew we
19 were going to get some lively comments, but I think we
20 under estimated how lively.
21 Any other quick comments before we break?
22 MR. EVANS: Right here.
23 MS. TWOHIG: Mike?

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
158.
158 Page 159 160
159
1 MR. EVANS: I'm Mike Evans, Fair Lending
2 Specialist with the FDIC. And I noticed this morning
3 that the individuals with Fair, Isaac and some other
4 individuals dealing with credit scoring have said how
5 credit scoring helps fair lending.
6 Well, I want the issue addressed. How does it
7 help fair lending when you look at the data from the Home
8 Mortgage Disclosure Act that is released every year. The
9 denial rates for minorities continue to increase every
10 year. And I also noticed that in the data that will be
11 released in the next couple of days, if you look at
12 closer, denial rates for whites are increasing, also.
13 MR. MEDINE: I think, Mike, that's an excellent
14 question, and we're going to have a panel on fairness
15 after the break, and why don't we have that as the first
16 topic during that panel.
17 MS. TWOHIG: We're going to need to --okay?
18 MS. PRATT: Hi. Sarah Pratt, now with the
19 National Fair Housing Alliance. I'll be very short. It
20 seems to me that there are a couple of really important
21 questions that we need to continue to deal with.
22 One is a synapsis of the issues raised earlier.
23 And that is, what difference --what are the sources of

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
159.
159 Page 160 161
160
1 the various items in the credit scoring technique? How
2 are things weighed? What difference does it make if you
3 have a trade line from Household Finance versus a
4 MasterCard, and why should it make a difference? Where
5 do these scores come from? That's the first item, and we
6 don't know the answers to those questions.
7 The second question it seems to me that ought to
8 be talked about some is the relationship of credit
9 scoring to the availability of loan products, and in
10 particular to referrals to the sub-prime market. In
11 other words, where --what difference does it make if you
12 have a particular credit score as to what loans are
13 actually available and are the most favorable loan
14 products available to you based on your scores? Why not,
15 if not? And at what point do you end up in the sub-prime
16 market because you have a particular score, and what
17 impact does that have on your finances for the future?
18 And even what impact does it have on your future credit
19 scores?
20 Thank you.
21 MS. TWOHIG: Thanks, Sarah. One or two more
22 questions or comments, and I think we'll need to break
23 for lunch.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
160.
160 Page 161 162
161
1 MR. FISHER: It will be quick. Greg Fisher,
2 creditscoring. com. I've listed 25 questions on the web
3 site, creditscoring. com. And that's G-R-E-G,
4 F-I-S-H-E-R. And I have copies of them here, if you
5 would like them. But I would like to ask the first five
6 here.
7 What percentage of loans evaluated by Fannie Mae
8 and Freddie Mac automated mortgage underwriting use
9 scores created by Fair, Isaac? Will the credit reporting
10 agencies release data to anyone who wants to create a new
11 scoring system? If I'm applying for a mortgage, how much
12 does it cost to have a score updated immediately, if I
13 find errors on my report? Do you allow creditors to see
14 scores for the purpose of evaluating existing customers?
15 Do credit card companies access credit scores months or
16 years after issuing the card to determine the risk of a
17 cardholder to raise or lower his rate? And the fifth one
18 out of the 25, how can I tell how many credit cards to
19 have to avoid that factor lowering my score?
20 MS. TWOHIG: Thank you for those questions.
21 There was someone else, I think, who had their hand up.
22 Yes?
23 MR. MANALP: My name is Yuksel Manalp.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
161.
161 Page 162 163
162
1 MALE SPEAKER: Spelling?
2
(Laughter.)
3 MR. MANALP: First name is Y-U-K-S-E-L. Last
4 name is Manalp, M-A-N-A-L-P.
5 MS. TWOHIG: Can everyone hear? Okay.
6 MR. MANALP: Okay. Nobody is talking about the
7 car industry today, so I thought maybe I'm in the wrong
8 place.
9 MS. TWOHIG: No, you're not in the wrong place.
10 We were focussing on credit scoring in the mortgage
11 industry. But we are interested in comments that are
12 broader than that.
13 MR. MANALP: My observation has been --I'm a
14 Finance Manager of a dealership. My observation has been
15 that the credit scores provided by FICO are somewhat less
16 accurate than they would like for it to be. That's
17 obvious from the comments Caryl made there and several
18 other people. Because the basis of the information is
19 the database supplied by credit reporting agencies. And
20 if the information is wrong, of course scores are going
21 to be wrong.
22 So I have two cases that I brought with me that
23 I can just briefly describe. I had a customer who is a

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
162.
162 Page 163 164
163
1 physician. He makes a lot of money. His wife is a
2 physician. They make a lot of money. And they go back
3 to 1983 in the credit bureaus. And I counted three late
4 payments, and they still have 77 percent of their credit
5 line that is still available. There are thousands of
6 payments they made on time. And the three late payments
7 added up to $175 or something like that. And his score
8 was 585.
9 That combined with how the banks are training
10 me, or telling me what to do --they send me a piece of
11 paper at the beginning of each month. If the guy's
12 credit is such and such, this is the rate he gets. If
13 it's less than six points, they don't even look at it,
14 and it's very difficult to change their minds.
15 So I'm just going to take a couple of minutes,
16 and I'm going to read to you what one of the credit
17 reporting agencies had us sign, because we're a
18 subscriber. Okay.
19 Here we signed. The subscriber recognizes that
20 factors other than the Empirica score must be considered
21 in making a credit decision, including the credit report,
22 the individual credit application and economic factors,
23 which is what I also believed. The factors that are

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
163.
163 Page 164 165
164
1 provided by Trans Union as significantly contributing to
2 the score may be disclosed to consumers as the reasons
3 for taking adverse action, as required by Regulation B.
4 However, the score itself is proprietary, may not be used
5 as the reason for adverse action under Regulation B, and
6 accordingly shall not be disclosed to credit applicants.
7 Two things. Number one, the banks are forcing
8 us and themselves --any of us --to use it 100 percent
9 of the time. Some banks are. Number two, we're not
10 allowed to disclose to the customer, so the customer has
11 no idea what his score is, not much less try to figure
12 how to straighten it up.
13 There are errors. The example I'm going to go
14 back to --I have another example.
15 MS. TWOHIG: Would you use the mic, please?
16 MR. MANALP: Yeah.
17 MS. TWOHIG: Thank you.
18 MR. MANALP: I have another example to go with
19 it. I pulled another bureau for a student. He's a full
20 time student and has part time work. The only credit
21 line on there he had is a Cellular One for something like
22 $36 a month for ten months. He had no other credit
23 lines. He had a score of 715.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
164.
164 Page 165 166
165
1
(Laughter.)

2 MR. MANALP: The bank's idea would allow me to
3 spot a $20,000 car to this person because of his score.
4 And the bank's idea would allow me to turn down the
5 physician, unless he paid 50 percent down and he paid 22
6 percent interest.
7 The credit scores are not accurate. And I think
8 the public should be allowed to find out what their score
9 is, and that would allow them to straighten it up.
10 There's only one way.
11 MS. TWOHIG: Thank you very much for your
12 comments.
13
(Applause.)
14 MS. TWOHIG: We're running late, but I think
15 we've had some excellent speakers this morning and some
16 very important issues put on the table. The agenda calls
17 for returning from lunch break at 1: 30, which does not
18 give you much time to eat. So why don't we all try to be
19 back in an hour. I guess it's 12: 50 now, so at 1: 50
20 we'll start up again.
21 Thank you very much.
22
(Lunch recess.)
23

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
165.
165 Page 166 167
166
1
2
3
4
AFTERNOON SESSION
5 -----
6 MR. MEDINE: Thanks very much. If we can have
7 half as lively a discussion this afternoon as we had this
8 morning, I think we'll be doing very well. First, before
9 we get started with the panel, Pam Johnson, who is here
10 from Fannie Mae, will have to leave shortly, and just
11 wanted to make some brief comments based on the morning's
12 discussions.
13 Pam?
14 MS. JOHNSON: Thanks. Is this working?
15 MR. MEDINE: Yes.
16 MS. JOHNSON: Okay, good. I asked David if I
17 could just make a few comments, because unfortunately I
18 am going to have to leave, and I didn't want to leave
19 before having an opportunity to comment on a few of the
20 comments and questions that came up.
21 I guess there are two things that struck me
22 about some of the discussion this morning. One is
23 obviously there is a fairly significant degree or level

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
166.
166 Page 167 168
167
1 of suspicion and doubt about credit scoring. It goes
2 without saying. And that seems to be fueled in part by
3 personal and professional experiences, and also by a lack
4 of information. We don't know they work, and therefore,
5 you know, we're uncomfortable with that. A perfectly
6 good reason to have this forum, and I think it's an
7 excellent opportunity to bring all these issues to bear.
8 The second thing that struck me is that there is
9 a significant level of mis-information about --and I can
10 only speak for Fannie Mae and the secondary market --
11 about the role that we play, and also what our position
12 is relative to credit scoring. And perhaps I didn't do
13 as good a job this morning trying to describe our
14 position, but I would like to take another minute to do
15 that, because I think it's important.
16 Before I do that, I think I would like to say
17 that for those of you who don't know this, and I'm sure
18 most of you do, there is a difference between the
19 secondary and the primary market. Lenders and brokers
20 originate loans in the primary market. They take the
21 loan application. They do the underwriting, now of
22 course using our automated underwriting systems --you
23 know, there is a role for us there --and they sell the

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
167.
167 Page 168 169
168
1 loans to Fannie and Freddie. And we purchase loans
2 directly from lenders and seller servicers to Fannie and
3 Freddie.
4 Now, what does that mean. Well, it means that
5 as the investor, we actually are providing --we do
6 provide guidance around what types of credit quality --
7 credit standards we expect when we purchase a loan, and
8 we provide that guidance in a lot of ways. And one of
9 the ways we provide that guidance is through written
10 communications with lenders.
11 Now, many of the questions or issues that came
12 up today are directly answered in some of those written
13 communications. And I just want to bring that up,
14 because I think it's important to know what our position
15 is. For example, Caryl raised a situation with a
16 borrower who had extenuating circumstances. A very low
17 credit score, but extenuating circumstances, such as a
18 medical --some sort of medical circumstance that created
19 some difficulty with credit. Or a divorce, for example.
20 Now, if you turn to some of our lender letters -
21 -and I'll be happy to send you one, Caryl --they're
22 very specific about that being used as an extenuating
23 circumstance for making a loan. And so we try to provide

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
168.
168 Page 169 170
169
1 the kind of guidance that lenders need to know when to
2 sort of look beyond the score, and look beyond something
3 that's determined to be high risk.
4 As far as the automated underwriting is
5 concerned, the way it works is that we provide a
6 recommendation around whether or not we think the loan
7 should be approved. And for Desktop Underwriter --I
8 can't speak for Loan Prospector. But for Desktop
9 Underwriter that recommendation on average occurs about
10 75 to 80 percent of the time. So Desktop Underwriter
11 users are getting an approved recommendation out of the
12 system about 75 to 80 percent of the time, meaning they
13 do not need to further look beyond that decision and do
14 any more underwriting.
15 Now, the other 25 --20 to 25 percent of the
16 time, the message is very clear. We want experienced
17 underwriters to look at that loan, because there is
18 something about that loan, at least according to the data
19 that the system has taken in, that appears to be high
20 risk. There may be extenuating circumstances for the
21 high risk. There may be problems with the credit report.
22 And again, another specific written guidance
23 that we give back is that if the credit report is wrong

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
169.
169 Page 170 171
170
1 or inaccurate, or there is information in it that's
2 inaccurate, the lender is instructed to disregard the
3 credit score, because it cannot be right.
4 So we do provide a lot of guidance around these
5 issues. I can tell you that based on conversations with
6 Caryl and others that it does not always get translated
7 in the lending community. But I think it's important to
8 know that, you know, we are looking to buy those loans
9 that experienced underwriters and experienced mortgage
10 professionals say are good quality loans. That's the
11 business we're in, and we have every interest in getting
12 those loans from lenders.
13 On the numbers --the examples that Caryl cited,
14 you know, I have no doubt that some lenders put
15 limitations on the loans they will sell. But I will
16 state again unequivocally --did I say that right --that
17 we do not have a minimum credit score in our automated
18 underwriting system. Again, we do not have a minimum
19 credit score in our automated underwriting system.
20 So, you know, do we think that the low score
21 represents on average a high degree of credit risk? It
22 does, but that assumes that the information is correct.
23 Again, a problem we all know. And it does not

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
170.
170 Page 171 172
171
1 necessarily pick up those extenuating circumstances. And
2 we would like and hope that the mortgage professionals
3 who are dealing with those loans will work through those
4 extenuating circumstances and make those loans.
5 It behooves us to buy them as opposed to them
6 having to go to a sub-prime market, for example. We want
7 the loans that perform, and so we have the same goal in
8 mind. And I just wanted to make those comments, because
9 I know it's very easy to assume that there is --you
10 know, we're on different pages. And I know that there
11 are issues here, but I just wanted you to know we are all
12 on the same page about making more loans and as many
13 loans as possible to credit worthy borrowers.
14 MR. MEDINE: Thank you. Pam, just while you're
15 here. The FDIC before the break raised a question, and I
16 just was wondering if you had a perspective on Fannie
17 Mae's point of view. And the question was, if scoring is
18 injecting more fairness in the lending process, why is it
19 that over the years we're seeing approval rates for
20 minority applicants declining rather than increasing on
21 lending?
22 MS. JOHNSON: Well, it's an excellent question.
23 First of all, let me say this. There is no denying that

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
171.
171 Page 172 173
172
1 there is discrimination in the mortgage lending process,
2 like there is discrimination in many industries. And I
3 think this is something that we all have to continue to
4 work to combat.
5 And if you go back to the Boston Fed study that
6 was done in what, 1993?
7 MALE SPEAKER: 1992.
8 MS. JOHNSON: Thank you. You know, it's clear
9 that one of the primary reasons for mortgage
10 discrimination has to do with subjectivity of the loan
11 officer making the decision, so that they are actually
12 weighing into their decision or to their underwriting --
13 or to their decision to make that loan, they're weighing
14 in the fact that this person is African American or
15 Hispanic, etc.
16 And one of the things that we view automated
17 underwriting as helping in this regard is that it takes
18 that subjectivity out of the equation. Now, automated
19 underwriting has not been around for a long, long time,
20 and it doesn't represent a hundred percent of the loans
21 that we purchase, certainly. And so it is not
22 necessarily used as widespread as one might think. We
23 would like it to be used on a broader basis. But I think

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
172.
172 Page 173 174
173
1 over time it is our belief that the use of automated
2 underwriting will in fact have a positive effect on
3 mortgage lending in minority communities.
4 And lastly, I guess I would say that we all know
5 that minority communities are the fastest growing, along
6 with immigrant communities. The fastest growing
7 population in this country. And the fastest growing
8 opportunity for those of us in this business. Besides
9 the moral imperative that discrimination is wrong, there
10 is a business opportunity here that we certainly don't
11 want to ignore, which is, there are plenty of loans out
12 there that should be made that aren't being made today.
13 And we would like to get at that. I mean,
14 that's the business we're in. So ultimately I think that
15 we will see a shift in this issue. That's our hope.
16 That's why we keep working at this. And we're hoping
17 that the subjectivity will, you know, over time be
18 eliminated with the use of tools that don't allow the
19 person inputing the data to make that choice.
20 MR. MEDINE: Okay. Thank you very much for your
21 comments. We're going to turn now to the Is Scoring Fair
22 panel, where we're going to discuss --as we have been to
23 some extent throughout the day --whether the use of

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
173.
173 Page 174 175
174
1 scoring adequately assesses risks for various
2 populations, and how overrides are used and if they're
3 done in a fair way or not.
4 We've had the pleasure of hearing from Peter
5 McCorkell already, so Pete will be around for the
6 question and discussion period. But each of the other
7 panelists will make some brief comments and then we'll
8 engage in discussion among ourselves and with the
9 audience.
10 Bob Cook is the Senior Fair Lending Specialist
11 at the Federal Reserve Board, and has been in that
12 position since 1994. He serves as the in-house
13 consultant on fair lending policy and enforcement issues
14 involving bank exams, complaint investigations, merger,
15 acquisition and expansion applications and regulatory
16 comment and review.
17 Bob?
18 MR. COOK: Good afternoon. It's a pleasure to
19 be here. Is this mic turned on? Can everybody hear?
20 MR. MEDINE: People can hear.
21 MR. COOK: The Federal Reserve's responsibility
22 with respect to credit scoring has to do largely, of
23 course, with examination of institutions that are subject

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
174.
174 Page 175 176
175
1 to its regulatory review. We have recently, along with
2 the other agencies, issued a set of interagency fair
3 lending exam procedures. And within those procedures we
4 have a set of review steps that we use in credit scoring
5 systems.
6 We take the approach in the front end of the
7 exam now that we look for risk factors. And in some
8 instances, a lender that uses credit scoring is going to
9 find a risk factor relative --or several risk factors
10 relevant to the credit scoring system that will cause us
11 to treat the credit scored product as perhaps one of the
12 focal points of the exam.
13 Once we make that review, our primary focus with
14 respect to credit scoring systems is in disparate
15 treatment. Now, that doesn't mean that we look to see if
16 a lender has purchased a system from Pete here, where
17 Pete snuck in females as a factor in the scoring process.
18 We don't think he does that any more. He's promised us
19 he won't.
20
(Laughter.)
21 MR. MCCORKELL: When did I stop beating my wife.
22 MR. COOK: That's right. What we looked at, and
23 primarily from the disparate treatment side, is whether

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
175.
175 Page 176 177
176
1 or not there is a judgmental creep back into the system
2 that uses scoring, such that all of the promised benefits
3 of using scoring that you eliminate disparate treatment
4 because everybody gets treated the same, has kind of
5 turned on its ear. Because after the scoring process is
6 all done, you come in and perhaps ignore --but I think
7 the proper term is that you take into consideration what
8 Pam described as extenuating circumstances, or you
9 override what the score seems to be saying. The score
10 seems to be saying this is a score of 440. It probably
11 shouldn't be made.
12 But there are reasons, many of which we've heard
13 about today, why a lender might want to make that loan.
14 The extenuating circumstances can go from --as both Pete
15 and Peter said, the score doesn't predict how individuals
16 are each going to perform. It predicts how a given
17 segment of the population that made up the scoring
18 development process indicates that a similar group of
19 people would perform. But as to each individual, that
20 score is not going to say whether that person is going to
21 default or going into bankruptcy depending on what the
22 purpose of the score is.
23 So somebody with a very brief band of

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
176.
176 Page 177 178
177
1 delinquency that reflects a short term economic downturn
2 for that person might look the same to the scoring system
3 as a sort of habitual credit delinquent, and overrides or
4 extenuating circumstances provide a means to overcome
5 that. As a consequence, you can see that we kind of
6 favor it. We believe that overrides have a definite
7 place to play.
8 And the only area that we --well, one of the
9 principal areas that we are concerned about is in the
10 fair lending area. And that is, are the overrides being
11 used on a basis that doesn't take a prohibited basis into
12 consideration. It doesn't end up allowing what we call
13 low side overrides, where the score is low and it's a
14 decision to make the loan anyhow, or high side overrides,
15 where the score is 770 and the bank says yeah, but I
16 don't like his looks and denies the loan.
17 Our principal concern, as you can imagine, is
18 with low side overrides. As I said, we look in the exams
19 at risk factors. And one of the risk factors before we
20 select a credit score product would be, doesn't there
21 seem to be a lot of overrides. And that in itself is not
22 necessarily bad. Maybe this is a bank that pays
23 extensive attention to extenuating circumstances. But it

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
177.
177 Page 178 179
178
1 is a red flag for us to see a very high percentage of
2 overrides of an otherwise --of an initially scoring
3 system. So that's going to become a focal point of the
4 exam.
5 The use of the override, to the extent that it
6 seems to be producing disparate outcomes for, let's say,
7 male versus female borrowers, or minority versus white
8 borrowers, is something where primarily we looking for an
9 analysis. The examiners are doing an analysis of the
10 quality of assistance given to those people who they
11 decide to assist in overriding a score, or in taking into
12 consideration extenuating circumstances. Does everybody
13 get kind of the same effort by the lender in attempting
14 to find ways to explain why this person ought to be
15 granted a loan, or are those granted on a disparate
16 basis. So that's sort of a primer on how we are taking a
17 look at credit scoring for the most part.
18 There is --and I know Pete sort of glossed over
19 it a little bit this morning, but promised to get back to
20 it this afternoon, as he knew we would want him to --the
21 issue of disparate impact, where the system itself,
22 although it essentially treats everybody the same,
23 produces different outcomes because of the very nature of

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
178.
178 Page 179 180
179
1 the system as it relates to differences in the population
2 that it scores.
3 So, for instance, one of the factors might be --
4 and I'm not saying that it's yours. One of the factors
5 might be that people have --and I believe somebody
6 mentioned it this morning --finance company trade lines
7 versus bank trade lines, and that people with finance
8 company trade lines receive less of a score or fewer
9 points for that than somebody with a comparable bank
10 trade, perhaps of the same amount, the same duration and
11 otherwise similarly situated. To the extent that
12 minorities or single females, for instance, might be over
13 represented in the population that uses finance companies
14 more often than they use bank trades, you could come up
15 with a disparate impact.
16 So one of the concerns we do have is whether or
17 not disparate impact is an issue. And one of the
18 questions I guess that I'll toss out to Pete to respond
19 to when he gets his turn after this entire table
20 discussion, is to what extent do the Fair, Isaac's of the
21 world --and I suspect you can speak for how some of the
22 other systems operate, but certainly Fair, Isaac --go
23 when they evaluate various possible criteria or factors

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
179.
179 Page 180 181
180
1 to put in the system to kind of weigh the relative impact
2 that one factor may have against another. So that's
3 another issue that I think we do --we are concerned
4 about. I must say that it's not a factor that we have
5 the same concern about as the use of overrides in these
6 scoring systems.
7 That factor brings in another issue, and that is
8 --the disparate impact factor --and that is the
9 lender's liability for disparate outcomes. A lender is
10 responsible under a ECOA and Reg. B for disparate
11 outcomes. And that's kind of a neutral term. That could
12 be disparate treatment or disparate impact, but disparate
13 outcomes from a credit decision process.
14 If scoring --if they follow Pete's book and
15 don't use overrides, and they're rejecting females at a
16 rate of 1.6 times males, then you could say that there
17 seems to be a disparate impact. The lender, however,
18 while it's being held responsible under Reg. B for
19 explaining that disparate impact --indeed, a lender is
20 given the opportunity under sort of the order of proof of
21 disparate impact to offer a business justifiable reason
22 --a business justification for why those outcomes are
23 different.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
180.
180 Page 181 182
181
1 And yet it doesn't know what --it may not know
2 what factors are in the box. It may not know what
3 weights are there. And it may not have the capacity to
4 make an evaluation that regulators are going to believe.
5 You know, the necessary credibility to say that yes, you
6 provided a justification for the difference in outcome
7 between male and female borrowers.
8 So that's an issue which I want Pete to address,
9 because it's been raised by lenders fairly recently as to
10 what do they do. Some of them have asked, how do I know
11 that my program that I've bought from a developer program
12 has been validated. Another obligation that lenders have
13 is to make sure that their program is validated. In many
14 instances, a validation is done by the producer of the
15 system who sold the system to the lender. And we suspect
16 that those validations are done well and wisely. But,
17 again, the lender hasn't got the capacity to show what he
18 did to validate, except to point to Pete and say, well,
19 Pete tells me that my system was properly validated.
20 So there is this kind of disconnect, despite the
21 fact that, you know, as a general matter when you've got,
22 you know, a firm like Fair, Isaac, that is too big to
23 discriminate

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
181.
181 Page 182 183
182
1 MALE SPEAKER: It's like a bank being too big to
2 fail.
3 MR. COOK: Too big to discriminate because there
4 is so much at risk, that we expect that validations are
5 going to be done well and wisely.
6 Another Fed focus --
7 MR. MEDINE: You have about a minute.
8 MR. COOK: Okay. Another Fed focus is on what
9 Carroll Justice spoke about, and that's the use of
10 scoring to support a risk based pricing of loans. And
11 really we're talking in many ways about sub-prime lending
12 there.
13 To the extent that a FICO score can
14 differentiate risk levels, it offers lenders a more
15 precise way to quantify the difference in the cost of
16 that risk, or in the way that that risk ought to be
17 priced for as between, let's say, a B and B-and a C+ and
18 a C loan. And to that extent, scoring serves a very
19 useful purpose in the sub-prime lending area, a risk
20 based pricing area.
21 The danger that we see --one of the dangers --
22 is the extent to which a lender doesn't pay close
23 attention to the risk differential that the credit score

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
182.
182 Page 183 184
183
1 is describing and takes sort of a more pull it out of the
2 blue approach to pricing for risk, adding a little extra
3 just to be sure that they've got it right. And there is
4 a challenge in pricing sub-prime loans, and a lot of
5 lenders --A lenders --who try to get into it, learn
6 that it's a very different task.
7 At any rate, that's something we're concerned
8 about, is that the price differential between one risk
9 level and the next ought to bear some reasonable
10 relationship to the actual difference in risk. The cost
11 of that risk.
12 And finally, from a consumer concern not knowing
13 their score --we heard some talk about that today --I
14 think --I'm not so sure that I share that concern to the
15 same degree that it was expressed today, and I'll tell
16 you why.
17 To an individual --as opposed to a lender
18 trying to see what the band of scores are of this
19 customer for some such analytical purpose, to an
20 individual customer it's probably more valuable to know
21 that he's got a score that's low, that's hurting his
22 capacity to either get a loan or to get an A loan. It's
23 probably more valuable to know what the various reasons

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
183.
183 Page 184 185
184
1 are that are driving that score down, rather than the
2 score itself, because he then has valuable --he or she
3 has valuable information about whether it's something
4 they can do something about in a short term, or whether
5 it's something that's going to have to heal over time.
6 And Reg. B requires lenders to provide the --
7 MR. MEDINE: I'm going to have to ask you to
8 wrap it up.
9 MR. COOK: Okay. Provide the most important
10 reasons why --not only as to judgmental systems, but
11 also as to scored systems, what most adversely affected
12 the score.
13 So with that, I turn it back to you.
14 MR. MEDINE: Thank you. Thank you for
15 highlighting, I think, some of the challenges in this
16 area, where we on the one hand have heard scoring can be
17 used to eliminate some of the bias and judgmental
18 systems, but yet disparate treatment may be inherent in
19 the scoring system itself, and where overrides take into
20 account some of the sympathetic concerns and/ or unusual
21 situations, but again could also be used in a
22 discriminatory manner. So some of these are some of the
23 challenges in using a scoring system.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
184.
184 Page 185 186
185
1 MR. MCCORKELL: I hope you meant to say
2 disparate impact might be inherent. Disparate treatment
3 would not be.
4 MR. MEDINE: Well, hopefully it wouldn't be.
5 Thank you. Turning to Debby Goldberg, Debby is the
6 Acting Director of the Neighborhood Revitalization
7 Project of the Center for Community Change. The project
8 provides assistance to groups that are concerned with the
9 availability and affordability of credit and other
10 banking services in their communities. Debby is also
11 involved with the Center's public policy work on
12 reinvestment and fair lending issues.
13 MS. GOLDBERG: Thanks. I think it's been a very
14 interesting discussion here so far today. The earlier
15 panels really laid out a lot of the issues that are
16 challenges for people who are working in this field, or
17 for consumers who are dealing with these questions. And
18 at the risk of --I'll make an effort not to repeat what
19 people have said before, but maybe to give you a slightly
20 different perspective than some of the people who have
21 spoken before me.
22 One of the things that I see when folks talk
23 about credit scoring and automated underwriting systems

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
185.
185 Page 186 187
186
1 --and I think we see this from a community reinvestment
2 perspective in other ways when you talk about technology
3 and its use in the banking industry and its impact on
4 consumers and their access to banking services, whether
5 those are mortgages or other forms of banking services --
6 is that there is this tremendous promise that technology
7 holds. And I think people want to believe that
8 technology can work to everybody's advantage, and it's
9 clear that it can work to some people's advantage.
10 But I think what you learn if you watch how
11 these systems play out over time, is that you can't
12 assume that the use of technology, whether that's credit
13 scoring, automated underwriting or anything else, will
14 necessarily work to the advantage of everybody in our
15 country. And the folks who are particularly at risk of
16 either not being served well or actually being
17 disadvantaged tend to be low and moderate income people,
18 people of color and other folks who are protected under
19 our fair lending laws.
20 And so I think it's very important not to assume
21 that just because something can be described as being
22 objective, that it will be used in an objective way. And
23 I think our quest to eliminate all subjectivity from the

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
186.
186 Page 187 188
187
1 mortgage lending process as a way of making sure that
2 nobody gets discriminated, is in fact a quest that's
3 going to be vain. And I'm not sure that we ought to be
4 pursuing it in the way that I think some folks are.
5 I think we ought to be looking perhaps at other
6 ways of ensuring that discrimination and unfair treatment
7 doesn't creep into the system, because I don't believe
8 that you can have a mortgage lending process that is
9 completely objective, no matter how you go about it. And
10 I think that that has actually been illustrated by some
11 of the remarks, as I said, of the folks who have gone
12 before me, and perhaps most effectively by this last
13 panel, who made it clear that at every step along the way
14 there is --as you know from the Wizard of Oz, there is a
15 man behind the curtain. You may have this wonderful, you
16 know, image projected on the screen, but there is in fact
17 a person, or in some cases a series of people, who are
18 involved in the process and who have an impact on the way
19 that the process and the system works.
20 I want to just try and kind of highlight for you
21 what I think are some of the points along the way in the
22 use of credit scoring and automated underwriting systems
23 where bias, whether that's illegal or just unfair, may

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
187.
187 Page 188 189
188
1 creep into the system.
2 One of those is at the front end with the data
3 entry. And the gentleman over here talked about, you
4 know, garbage in and garbage out. And I think that's
5 something that we all have to keep in mind, that not only
6 are there problems with the credit reporting system in
7 this country. But as I tried to listen carefully to the
8 speakers who were explaining how credit scoring works, it
9 seems to me that there are elements of the system at the
10 front end that are subject to manipulation, which can be
11 a good thing or it can be a bad thing, depending on
12 whether it's done in a way that helps you or done in a
13 way that hurts you.
14 The Boston Fed study that someone referred to
15 earlier actually identified that as one of the key
16 factors in explaining the disparate levels of mortgage
17 lending to minorities and non-minorities in Boston. And
18 that was how much assistance were people given in being
19 on paper the best qualified applicant that they could.
20 And I think when you talk about this kind of a
21 system, you still have that issue that you have to
22 grabble with. How much help does someone get at the
23 front end to look as best as they can look on paper, so

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
188.
188 Page 189 190
189
1 that the data going into the system is going to serve
2 them in the best possible way, and are people given
3 different levels of treatment.
4 The second question --or the second place where
5 I think unfairness or discrimination can creep in is in
6 the scoring system itself. And I think that, you know,
7 the example that has been cited a couple times now, of
8 how the use of nontraditional sources of credit --the
9 household finances or others of the world --affect
10 someone's score and the way that they get treated by the
11 system, is a very good example of what the potential
12 problems may be.
13 What happens to someone who has mostly paid in
14 cash over the years and doesn't have much or any credit
15 record? You know, the kinds of things that we tend to
16 see more in low and moderate income communities and
17 communities of color, I am troubled are not necessarily
18 treated adequately by credit scoring systems. And I
19 don't think anyone will ever be able to satisfy me, and I
20 think a lot of other people, on the question until there
21 has been somebody who does an impartial and thorough
22 analysis of these systems, with access to all of the
23 details about the way that they work, and can say

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
189.
189 Page 190 191
190
1 definitely at least at some point in time that there is
2 no bias built into the system itself. And that kind of
3 scrubbing of the credit reporting system --or, excuse me
4 --the credit scoring system simply has not happened to
5 date.
6 And, you know, in this same light, I want to
7 make a different point about the use of nontraditional
8 sources of credit. Because I think when you talk about
9 fairness, one of the issues that low income people in
10 particular, and again people of color, have faced is that
11 their access to credit sources in many cases has been
12 limited, and people have been forced to use high cost
13 sources of credit, to pay more for basic credit,
14 regardless of what form that may take.
15 And particularly when you're talking about low
16 and moderate income people, although not exclusively in
17 that case, if you're paying more for credit, then the
18 chances that you're going to have problems in paying on
19 time seem to me to be increased. And I'll come back to
20 this in a minute.
21 But I think one of the issues when we talk about
22 fairness is how do we make sure that people who are good
23 credit risks get access to the best priced, best

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
190.
190 Page 191 192
191
1 structured credit that they ought to be able to have
2 access to. How do we change the dynamics of the system
3 that have led to some of the statistical results that
4 maybe we see in the analysis that's done so far.
5 And then the third point along the way where I
6 think bias and unfairness can creep in is at the back
7 end. And again, people have talked about how that can
8 play out. Once you get a score, or once you get the
9 result of the automated underwriting system, then what
10 happens. Are you --you know, has your file gone back
11 through and underwritten manually? You know, are
12 compensating factors then taken into account? Does that
13 happen on a fair and equal basis or not?
14 In this area, one of the points I want to make
15 sure we're all clear about is that there is a very large
16 part of the mortgage market that is not subject to review
17 by my colleagues who are here on this panel who work for
18 the government --the banking and regulatory agencies.
19 And so even though some lenders have bank examiners who
20 come in and look over their shoulders to try and make
21 sure that these kinds of judgmental aspects of their
22 mortgage lending process doesn't play out in a way that
23 may be biased, there are a lot of other lenders out there

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
191.
191 Page 192 193
192
1 who are never subject to that kind of review. And I
2 think when we look at the impact of credit scoring and
3 automated underwriting, we have to look at the whole
4 market and not just the regulated part of the market.
5 And then the other thing I wanted to mention at
6 this point in terms of the back end is, once you get a
7 score --once you get an outcome, how do you then know
8 that an applicant gets the best possible product that
9 they qualify for. Or to what extent, as, you know, Bob
10 Cook was just saying, does, you know, a little padding
11 get added on to the cost of that loan, whether that's in
12 the form of the interest rate or in the form of other
13 fees that may get tacked on. And who are the people who
14 get the best deals and who are the people who get the
15 worst deals. And how does somebody know what is the best
16 deal that they ought to qualify for.
17 I think the mystic that we see associated with
18 credit scoring systems is particularly problematic in
19 that way, because people don't know what they ought to be
20 able to get. And it's very hard for a consumer to be an
21 effective advocate on their own behalf, given that kind
22 of mysterious nature of the system.
23 I want to just close with two quick points. And

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
192.
192 Page 193 194
193
1 that is --and I think Margot will probably talk about
2 this more a little bit later on. But for me it's really
3 impossible to look at credit scoring and automated
4 underwriting without also then making the linkage to risk
5 based pricing and sub-prime lending. Because in the
6 communities where the groups that the Center for
7 Community Change works, low and moderate income and
8 minority communities across the country, that is a very
9 big issue.
10 The presence of sub-prime lenders and predatory
11 lenders, and the way that they use credit scoring, is a
12 very big problem that groups are really struggling with.
13 And on the one hand we have community groups who have
14 been working for years to develop the kind of mortgage
15 counselling programs that some of the earlier panelists
16 talked about --I guess Marcia Griffin, in particular --
17 as a way of taking people who in the kind of, you know,
18 system that we're talking about, might not qualify for A
19 credit and making them qualify for A credit, and getting
20 them access to A credit.
21 And I think that's a good thing. And I'm very
22 concerned that the widespread use of automated systems
23 and the fear that many --either the fear that some

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
193.
193 Page 194 195
194
1 lenders will have that if they use any kind of judgmental
2 --if they have any kind of judgmental part of the
3 process, that they're going to be, you know, straying
4 into areas of discrimination, so that they'll just kind
5 of shut that down altogether.
6 And they will be --despite all of the letters
7 that Fannie Mae and Freddie Mac may send out, they're
8 just not going to want to deal with applications where
9 somebody's credit score is too low. That and the
10 increased use --or increased activity by sub-prime
11 lenders in low income neighborhoods, I'm afraid may
12 really undermine the success and the continued benefits
13 that housing counselling I think has brought to many
14 communities across the country.
15 And I also worry about the notion that the best
16 way to deal with people who don't on their face --or as
17 a result of this kind of allegedly objective analysis --
18 look like good credit risks, is just to give them credit
19 at a higher price. I think that's a very dangerous
20 notion.
21 And I guess I've been in this business long
22 enough to be mindful of some of the lessons that we've
23 learned from FHA lending over the years, where there were

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
194.
194 Page 195 196
195
1 certain neighborhoods and certain types of people that
2 got labelled --I'm kind of simplifying the process here.
3 But got labelled as FHA communities. And the only kind
4 of credit that was available in those communities was
5 credit from FHA, which in many cases was very poorly
6 underwritten with disastrous consequences for individual
7 homeowners and for the communities in which they lived.
8 And I think if we replace that kind of system
9 with one where the only credit that is available in
10 certain kinds of neighborhoods, or just certain kinds of
11 borrowers, is high priced credit, and in some cases
12 predatory credit, that we run the risk of having exactly
13 the same problems that we suffered with poor use of the
14 FHA program in the past. And I don't think that that
15 works to the benefit of the country as a whole, let alone
16 to the individual consumers who are affected.
17 And let me stop there.
18 MR. MEDINE: All right. Thank you very much,
19 Debby. If I could follow up on one of the points you
20 made earlier, as I understand it, you're suggesting that
21 there might have been a bias in lending that occurred in
22 the past --
23 MS. GOLDBERG: Right.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
195.
195 Page 196 197
196
1 MR. MEDINE: --that might have shifted certain
2 consumers from, say, banks and traditional financial
3 institutions to finance companies at higher rates, and
4 that those consumers, as a result of that bias, were
5 ending up paying higher rates and then more likely to
6 default, and therefore a scoring system would essentially
7 perpetuate --
8 MS. GOLDBERG: Right.
9 MR. MEDINE: --some of that discrimination that
10 occurred in the past. If that's your point --
11 MS. GOLDBERG: Yes.
12 MR. MEDINE: --the question is, how do you
13 today address that concern in evaluating an individual's
14 credit worthiness?
15 MS. GOLDBERG: Well, I think, you know, one of
16 the ways to get at that is through the kinds of
17 counselling programs that were, you know, described and
18 that I talked a little bit about, where you work with
19 someone to both overcome the problems that they may have
20 experienced because they've been forced to use high cost
21 credit, so that they repair their credit record, and then
22 also to make sure that they are really prepared to become
23 a homeowner, if we're talking about mortgage credit.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
196.
196 Page 197 198
197
1 That they know what to expect in terms of future costs
2 associated with home ownership.
3 You know, one of my colleagues has called it the
4 Hechinger's tax. You know that every month you're going
5 to have to go down and spend 25 bucks at Hechinger's or
6 whatever to repair something around your house. And
7 there is not a landlord that you can call to do that, so
8 you've got to built that cost into your future household
9 budget.
10 So that people do become good credit risks, and
11 then you give them access and a shot at a loan that's at
12 the same price as, you know, somebody who has had
13 different advantages in the past --may have had. And I
14 think the experience with the programs that have done
15 that has for the most part been excellent.
16 When I talk to groups around the country that
17 are involved in housing counselling programs, and they
18 talk about the default and delinquency rates associated
19 with the portfolios of loans, you know, where they've
20 done counselling, those default rates are often as low as
21 or lower than the rest of the customers for those
22 particular institutions.
23 And so I think the success is very good, and we

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
197.
197 Page 198 199
198
1 don't want to prevent that from happening in the future.
2 MR. MEDINE: Great. Thank you very much. I'm
3 turning now to Russ Bailey. Russ is the Fair Lending
4 Team Leader in the Community and Consumer Policy Division
5 at the Office of the Comptroller of the Currency.
6 Oh, George Knight? Is George Knight here?
7 Okay. Returning to Russ. He's with the Office of the
8 Comptroller of the Currency, and he has held his position
9 as Fair Lending Team Leader since 1995. He's responsible
10 for developing and implementing the OCC's Fair Lending
11 supervision and Enforcement Program, and for providing
12 advice and counsel on fair lending issues.
13 Russ?
14 MR. BAILEY: Thank you, David. Many of you may
15 not know what the Office of the Comptroller of the
16 Currency is. It is the government agency that regulates
17 national banks and their direct subsidiaries. We are to
18 national banks what the Federal Reserve is to state
19 member banks as far as regularly examining them for,
20 among other things, fair lending matters. We appreciate
21 the invitation to speak to you today and deal with some
22 of these very thorny issues.
23 The OCC issued a bulletin in 1997. It's OCC

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
198.
198 Page 199 200
199
1 Bulletin 97-24. It deals with credit scoring issues.
2 And I put 200 of them on the table outside the door. I
3 would recommend this publication to you if you are
4 interested in credit scoring issues.
5 Let me just explain a little bit about what it
6 contains. About the first 7 and 1/ 2 pages it is fairly
7 arcane. Safety and soundness information on how to
8 develop and validate credit scoring systems and some of
9 the issues that the OCC looks at from a safety and
10 soundness standpoint. And then there are several pages
11 that discuss the OCC's concerns about some of the
12 compliance issues that arise in credit scoring. And if
13 you're more interested in those compliance issues, you
14 might want to skip the first 7 1/ 2 pages with lots of
15 detailed charts and complicated issues.
16 The basic question before this panel is, is
17 credit scoring fair. And to answer that from OCC's
18 perspective in some of the terms that have been used
19 today, I think I would have to say that it is impossible
20 to predict on an individual basis. But we think the odds
21 ratio is that credit scoring systems make an underwriting
22 system likely more fair.
23 Is that okay, Pete?

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
199.
199 Page 200 201
200
1 MR. MCCORKELL: Yes.
2 MR. BAILEY: Is that stated in credit scoring
3 terminology? We would second Bob Cook's remarks that we
4 think that the greater danger with credit scoring systems
5 is that they still can be used in a discriminatory
6 manner. It is not a panacea. Just because somebody has
7 a credit scoring system, that doesn't mean okay, the
8 underwriting system is, therefore, going to be fair.
9 We have cited four of our institutions for fair
10 lending violations involving credit scoring systems and
11 made referrals to the Department of Justice. I would
12 just like to talk about those real quickly. Debby said
13 that there always is a person behind the curtain. Well,
14 sometimes the person behind the curtain does stupid and
15 unlawful things. It doesn't happen very often, but it
16 does sometimes occur.
17 We had one institution that set up their credit
18 scoring system so that if co-applicants were married,
19 they were entered --a little j was entered in a certain
20 field in the credit scoring database. And then the
21 credit scoring system combined their credit histories,
22 eliminated redundancies, and produced one credit score.
23 It also combined their incomes and produced a combined

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
200.
200 Page 201 202
201
1 debt to income ratio.
2 But if the co-applicants were not married to
3 each other, that little j wasn't entered in that one
4 field in the system. And so what was designated as the
5 primary applicant was scored, and only the income of the
6 primary applicant was counted in computing a debt to
7 income ratio. So they set that system up to begin with
8 to treat unmarried co-applicants differently than married
9 co-applicants, and that's discrimination on the basis of
10 marital status, and that's prohibited by the Equal Credit
11 Opportunity Act.
12 We had another situation in which a bank had an
13 age-split scorecard. Now, age is the one prohibited
14 basis that Regulation B permits a creditor to take into
15 account in a credit scoring system that is validated. Or
16 in the language of the reg, empirically derived,
17 demonstrably and statistically sound.
18 But this credit scoring system had not been
19 validated. Therefore, it's use of age in an age-split
20 scorecard was inherently in violation of Regulation B and
21 the Equal Credit Opportunity Act. So they just skipped a
22 fundamental step that they had to take care of in order
23 to use an age-split scorecard.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
201.
201 Page 202 203
202
1 We had another situation in which the same
2 scorecard was used to process all applications for
3 credit. However, if an applicant applied on the Spanish
4 language application, it took a higher score to pass.
5 Let me say that again. If an applicant applied on a
6 Spanish language application, it took a higher score to
7 pass. In the internal memos in the institution, the
8 group of people who applied on Spanish language
9 applications were referred to as "the Hispanics."
10 Whenever you get into credit segmentation, and
11 large credit card operations can have lots of
12 segmentations, you run the risk of segmenting the
13 population into some group that is highly correlated with
14 a prohibited basis.
15 Now, I've heard this --an anecdotal story when
16 I was making a presentation. Someone said to me, you
17 know, several years ago I ran across a situation where a
18 credit card operation had one scorecard. It was an
19 affinity card for subscribers to Town & Country magazine,
20 and then they had an affinity card for subscribers to
21 Ebony magazine. And they used different scorecards to
22 process those applications.
23 Now, wouldn't most of us think, whoa, there

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
202.
202 Page 203 204
203
1 might be a problem there? Okay. And I think many times
2 people don't realize how many segmentations there might
3 be in a credit operation. It's rare to find it in the
4 mortgage market, the home purchase market, that much of
5 the discussion has been about today. But it's very, very
6 common to find it in other kinds of credit products.
7 We also made a referral to the Department of
8 Justice on a case in which credit scoring was used as
9 part of the underwriting process for home improvement
10 loans. Since home improvement loans are residential real
11 estate related transactions, we had the government
12 monitoring information on those decisions, which gave us
13 the race of the applicant. And if you made a passing
14 credit score, your odds were six times greater of being
15 denied anyway if you were black than if you were white.
16 And if you made a failing credit score, your odds were
17 twice as great that you would be approved anyway if you
18 were write than if you were black.
19 We wound up actually going in and pulling a
20 sample --a random sample of files in doing our own logit
21 regression model of the underwriting decisions that were
22 made at this institution, and race turned out to be a
23 very significant variable. And the loan officers were

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
203.
203 Page 204 205
204
1 simply using their override discretion to make a pattern
2 of decisions that we believe was discriminatory. And we
3 referred that case to the Department of Justice as well.
4 We have a couple of other cases in the early
5 stages. Another situation in which it looks like a bank
6 has built into their system treating married co-
7 applicants differently than unmarried co-applicants.
8 That's in the early stages, however. And we also have
9 another case in which it appears that an institution may
10 have a completely different scorecard for people who
11 apply on a Spanish language application. So we have the
12 credit segmentation problem arising again, at least
13 possibly, in a case.
14 So those are some of the kinds of things that we
15 run into from a disparate treatment standpoint. We think
16 that there is much less likelihood of a problem from a
17 disparate impact standpoint, and we believe it's a very
18 difficult issue when you talk about impact in credit
19 scoring.
20 The OCC has decided that it will not cite a bank
21 for a violation of fair lending laws using a disparate
22 impact theory applied to a credit scoring system, until
23 and unless we have developed guidance for the

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
204.
204 Page 205 206
205
1 institutions that we regulate, issued that guidance, and
2 then if they violate the guidance, we would not be happy.
3 One other thing I want to mention. Our
4 Comptroller has spoken out publicly recently about a
5 trend that's developing in the industry of not reporting
6 information to credit reporting services. A number of
7 sub-prime lenders are not reporting performance on those
8 loans, which tends to make it more difficult for someone
9 who is forced into the sub-prime market because of some
10 past credit problems to rehabilitate their credit history
11 and move back into the less expensive market.
12 Some lenders have acknowledged that they have
13 stopped reporting credit limit and balance information.
14 The Comptroller's concern is that this lack of reporting
15 of data may contribute to the inaccuracy of credit
16 scoring, since the credit scoring is based on the
17 information that's reported to the credit bureau. So
18 that's another issue of concern to us.
19 Thank you very much.
20 MR. MEDINE: Thank you.
21 MR. MCCORKELL: I would like to point out that
22 you've all witnessed an historic occasion. This is
23 probably the first time in memory that Russ and I have

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
205.
205 Page 206 207
206
1 ever been on the same panel and haven't disagreed about
2 anything.
3
(Laughter.)
4 MR. BAILEY: Well, I wouldn't say we haven't
5 disagreed.
6
(Laughter.)
7 MR. MCCORKELL: Well, I don't disagree with
8 anything you just said, Russ.
9 MR. MEDINE: Thanks, Russ. I just wanted to
10 point out that the opposite can be true from your
11 experience. We've seen the reverse of a company that
12 used a judgmental system, where they failed to aggregate
13 the income of unmarried co-applicants and got such an off
14 the shelf scoring model that did. And so they
15 essentially backed themselves into complying with the
16 law, because the scoring developer knew what the law was,
17 even though their judgmental system didn't provide for
18 that.
19 So it can work both ways. I'm turning now to
20 Margot Saunders. Margot is the Managing Attorney of the
21 Washington office of the National Consumer Law Center.
22 Her duties include representing low income clients in
23 Congress on financial credit issues. And she has

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
206.
206 Page 207 208
207
1 testified on numerous occasions before congressional
2 committees on the impact of various proposals on low
3 income households, and is the author of numerous articles
4 on consumer and utilities law as they affect low income
5 people in the United States.
6 Margot?
7 MS. SAUNDERS: Thank you. And it's a pleasure
8 to be here today. The National Consumer Law Center is a
9 public interest law firm that has its main offices in
10 Boston. But we represent low income people through the
11 legal services programs all over the country, and I've
12 been doing this for over 20 years.
13 So it's with that perspective that I bring this
14 information to you. Much of what I was going to say has
15 been said. I'll only repeat some of it. My folks, the
16 experts in credit at the National Consumer Law Center,
17 are unequivocally convinced that credit scoring is biased
18 against poor people. And some of the reasons that we
19 expect that to be true are some that have been mentioned
20 already today.
21 For example, as was just mentioned by Russ from
22 the OCC, some major providers of mortgage credit and
23 other credit in the sub-prime market are not reporting

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
207.
207 Page 208 209
208
1 information at all. Those are most likely to be low
2 income people who are not getting the benefit of positive
3 credit scores from this. Much of the credit obtained by
4 my clients from finance companies, rent to own companies
5 or payday loans, either is not reported or is reported in
6 a way that will result in a lower credit score. And the
7 cost of the credit, as Debby mentioned, is very high. If
8 they do get it, it is far more likely that they will
9 default, which will lead to a bad credit rating.
10 Credit scoring is based on statistical
11 relationships, not causal links. I'm a Jewish woman.
12 There is a statistical relationship between Jewish women
13 and a certain kind of breast cancer. That does not mean
14 because I am Jewish that I will get breast cancer. And
15 it's the distinction between a statistical relationship
16 and a causal link that we think has to be made before
17 credit scoring can truly be evaluated.
18 For example, when you look at Fannie Mae's
19 numbers, according to Pam Johnson's presentation, if you
20 have a credit score of less than 620, then there was a
21 3.1 percent chance that you would default, as opposed to
22 if you had a credit score of greater than 720, I think
23 was the number, then it would be less than a one percent

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
208.
208 Page 209 210
209
1 chance that you would default.
2 Well, what about if you've got a credit score
3 less than 620? That means you also have a 96.9 percent
4 chance that you won't default. And has anybody done the
5 analysis between the people in that group to see the
6 distinction in that group and what might be causing those
7 two very different numbers? And guess what's happening
8 to those --that's a hundred people --those 97 people
9 who are not defaulting and yet are paying the price for
10 being somehow lumped in the group where there is an extra
11 two percent chance that they will default.
12 And I want to reemphasize something that Debby
13 said. Credit scoring is intended to provide guidance to
14 multi affiliated lenders to steer riskier borrowers to
15 higher price lenders. But will the reverse also be true?
16 If I walk into Nation's Credit, or a finance company that
17 is affiliated with a bank, and I qualify for the finance
18 company loan, but I also would qualify for the bank loan
19 at a much lower rate, will I be sent down to the bank?
20 No. But if I go into the bank and I don't qualify for
21 anything but the finance company loan, I'm sure that I
22 will be sent over to the finance company. And that is
23 something that I know --I understand the Federal Reserve

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
209.
209 Page 210 211
210
1 Board is concerned about, but let's see some action on
2 it.
3
(Laughter.)
4 MS. SAUNDERS: I'm very popular. The greater --
5 MR. COOK: More than you know.
6 MS. SAUNDERS: We really get most concerned
7 about credit scoring, because it's going to lead to this
8 new thing of risk based pricing. Well, risk based
9 pricing has been in existence for years. When I started
10 practicing 20 plus years ago, we had finance companies in
11 North Carolina that were, and still are, allowed to
12 charge 36 percent. And the stated reason in the laws
13 that created the small loan acts all over the country
14 that purported to allow 36 percent --that did allow 36
15 percent, was that these were riskier loans.
16 Well, the question becomes, at what point does
17 higher risk cause risk. If you --if Peggy and I have
18 the same amount of income, and I get a loan for 100,000
19 at seven percent, and she gets a loan for 100,000 at 12
20 percent, guess who's got a higher chance of defaulting.
21 That's obvious.
22 The other very important point that applies to
23 my clients, is that there is a tremendous difference

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
210.
210 Page 211 212
211
1 between risk of loss and risk of default. When you are
2 dealing with a loan which is more than adequately secured
3 by a home, which sub-prime lending mortgage loans are
4 always --they're very often a very high loan to value
5 ratio. There may be a very high risk of default and
6 rollover --flipping of the loan. In fact, we think --
7 or we are convinced that that is an intent --the intent
8 of some of the lenders. They like the rolling over of
9 the loans, because that way they get to charge more
10 points and closing costs next time and strip more equity
11 out of the home.
12 But there is never a risk of loss to the lender,
13 because there is so much equity in the loan --in the
14 home, excuse me --that even when there is a default and
15 a foreclosure, the loan and all the associated costs with
16 the foreclosure will be covered by the sale of the home.
17 So risk of loss should be what's measured, not risk of
18 default, in creating --in justifying risk based pricing.
19 And as I was happy to hear Robert Cook say, when
20 you do risk based pricing, the risk base --the extra
21 cost of the loan should be closely tied to the extra
22 actual risk imposed by that loan. Well, let me tell me.
23 We've got lots of numbers already which show that riskier

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
211.
211 Page 212 213
212
1 loans --sub-prime loans --are far more profitable to
2 the lending industry than the A loans are. Which means
3 that if they're more profitable, then the amount of
4 interest or the payments on those loans are covering a
5 lot more than just the added risk.
6 We know this by comparing banks' income and
7 profit and loss statements to finance companies' profit
8 and loss statements. For years we have been able to see
9 that finance companies have a return on the dollar that's
10 two or three times what banks have, which is why banks
11 are now moving into the sub-prime market.
12 So I think you already have some jobs cut out,
13 regulators, and we'll be glad to help nudge you along.
14 Thank you.
15 MR. MEDINE: Margot, I wanted to raise a
16 question that you didn't touch on, which is it has been
17 suggested in the past that certain low income people do
18 have spending histories, but they're not with credit
19 cards. They're with utilities. They're with landlords.
20 And those are not reported on the credit reports, which
21 then serve as the basis for risk scores.
22 Do you have a view on how important that is to
23 the fairness of the scoring process, that certain payment

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
212.
212 Page 213 214
213
1 type experiences are not incorporated into the score?
2 MS. SAUNDERS: I think it's unfair. I mean, I
3 was actually struck by the Freddie Mac presentation,
4 where it was said unequivocally that they do not look at
5 the neighborhood, which is a related issue --the
6 neighborhood that the customer lives in or where the
7 house will be purchased --as indicators of whether they
8 will make that loan. But those issues are already
9 probably subsumed within the credit score, which is a
10 factor in their overall mortgage underwriting.
11 So I think there is --and the reason I didn't
12 address the utility question was because I missed the
13 first section this morning, so I wasn't quite sure what
14 he said on that, to tell you the truth.
15 MR. MEDINE: Okay. Well, maybe we'll have
16 people comment on that later on.
17 MS. SAUNDERS: And I also thought, you know,
18 when Pam said that --Pam pointed out that there was very
19 clear numbers that 39 percent of very poor people had
20 credit scores over 740, and 38 percent of very rich
21 people had credit scores over 740. I wonder, though,
22 exactly which poor people those are, because I have
23 trouble --

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
213.
213 Page 214 215
214
1
(Laughter.)

2 MS. SAUNDERS: Trouble assuming that they're
3 many of my clients. Yeah. I think they may be med
4 students and so on.
5 MR. MEDINE: Okay.
6 MS. SAUNDERS: Thank you.
7 MR. MEDINE: Thanks very much, Margot. Our last
8 presentation is from Raj Mehra, who is the Director of
9 Financial Risk Management Consultant Practice of
10 PricewaterhouseCoopers, with responsibility for mortgage
11 and a specialty of finance risk consulting. He clients
12 include mortgage and specialty finance companies,
13 commercial banks and thrifts. He has over 12 years of
14 experience in mortgages in the capital markets. And
15 prior to joining PricewaterhouseCoopers, Raj spent
16 several years on Wall Street as a Senior Mortgage
17 Strategist.
18 Raj?
19 MR. MEHRA: Thanks, David. Glad to be here this
20 afternoon. I just want to talk about a couple of things.
21 Many of my clients are mortgage lenders. A good portion
22 of them are sub-prime mortgage lenders, residential
23 mortgages, as well as manufactured housing that no one

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
214.
214 Page 215 216
215
1 has spoken about out here this afternoon. I just want to
2 talk about a couple of things that some of the folks in
3 the panel made mention of.
4 It was interesting that Bob Cook made mention of
5 overrides, and said what really concerned the Federal
6 Reserve Board more than anything else were the low side
7 overrides. That kind of makes sense when you're talking
8 about safety and soundness of the bank. The sub-prime
9 lenders that we work with do a pretty good job tracking
10 low side overrides, because it affects the bottom line.
11 If you make a loan that turns out bad, well, you know,
12 that's bad, right.
13 The one thing that they do not do as good a job
14 tracking --and we're constantly talking to them about it
15 --is the high side overrides, where the scorecard model
16 says that it's a good credit, accept it, and they don't.
17 That's a high side override, and the lenders don't do
18 quite as good a job tracking the high side overrides.
19 And I guess what I mean by tracking, it often boils down
20 to --if you have a low side override, you can track the
21 performance of the loan with the decision that was made
22 to make that loan.
23 And you ought to be able to do the same thing

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
215.
215 Page 216 217
216
1 with the high side overrides. Loans that you did not
2 make, you ought to be able to track that performance to
3 see if it was a good loan or a bad loan. In fact, a
4 couple of days ago I was speaking with a client of mine,
5 a lender, and I pointed out to the gentleman that the
6 percentage of high side overrides that his firm had made
7 were 30 percent over the last year. And he said, you've
8 got to be kidding. That much? He said that's a lot of
9 business profit we're giving up.
10 So I would argue that for purposes of consumers
11 and discriminatory lending, maybe high side overrides is
12 an area that ought to be talked about more frequently
13 than low side overrides for the purpose of this panel.
14 The second thing --and this really talks to
15 another thing that members on the panel have spoken
16 about, which is the whole issue of is credit scoring a
17 black box or is it not. I believe someone from Freddie
18 Mac spoke this morning and said well, it's not going to
19 be a black box. We've got all the stuff.
20 I can tell you that with the exception of some
21 large, sophisticated lenders primarily in the A market,
22 there are many lenders out there for whom credit scoring
23 models are literally a black box. For sub-prime lenders,

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
216.
216 Page 217 218
217
1 for the most part, once again, unless they're very
2 experienced, large companies and been there for many
3 years, for them to buy an off the shelf system --and for
4 a lot of them, it is a black box.
5 I think Debby made a very good point when she
6 said there are a large number of lenders out there that
7 are not regulated. You've got banks that are regulated
8 by the OCC and the Federal Reserve Board. There are a
9 good portion of lenders out there, sub-prime lenders,
10 monoline independent lenders not affiliated with banks
11 and not affiliated with any other regulatory corporation.
12 They're not regulated at all.
13 And I think it's an issue of --you know, the
14 process needs to be fixed. And that's all I'm going to
15 say for now.
16 MR. MEDINE: Thanks. Maybe you could just give
17 your views as to why as many high side overrides occur as
18 do. It seems somewhat counter intuitive, and I guess the
19 corporate executives had the same reaction. Can you
20 explain some of the factual circumstances that would lead
21 to someone with a high score going to a sub-prime lender
22 and not getting a loan?
23 MR. MEHRA: I think there are a variety of

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
217.
217 Page 218 219
218
1 factors. It could be as simple as --I mean, lenders
2 will never document the information. But it could be as
3 simple as, you know, a judgmental call based upon, you
4 know, some things that are not captured in the scoring
5 model, or that they cannot possibly include in the
6 scoring model. But nonetheless the lending officer --
7 because the whole point of overrides is that it gets
8 reviewed.
9 If someone decides to accept or reject a score,
10 it has to get reviewed by a human underwriter. At that
11 point, judgment enters into the equation. And, you know,
12 if an individual is bias, or an institution is bias, it
13 could enter into that decision. It could be as simple
14 as, I think as Bob mentioned, you don't like the face or
15 didn't like the way credit was performing. Didn't like
16 the way the defaults had been occurring in that
17 particular part of the country.
18 And geographics is a very --studies have shown
19 that geographics statistically are quite important
20 predictors of default. They may not be included in the
21 credit scoring model, but by the same token, if you have
22 an override situation, someone may well just apply that.
23 We would never know.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
218.
218 Page 219 220
219
1 MR. BAILEY: Can I help?
2 MR. MEDINE: Sure.
3 MR. BAILEY: In answer to that question, David,
4 it's our experience that you'll almost always find
5 significantly more high side overrides than low side
6 overrides, because there is almost always other criteria
7 than the credit score. Just to simplify it, let's say
8 it's only a credit score and a debt to income ratio. You
9 must score at least 200 and your debt to income ratio
10 shall not be more than 40 percent.
11 Well, if somebody has a 220 credit score, but
12 their debt to income ratio is 45 percent, and they turn
13 it down for the debt to income ratio, that will be
14 counted as a high side override, because they passed the
15 credit score part of it. And there are typically more
16 than just the one other policy criteria, and so that's
17 how you get so many high side overrides.
18 MR. MCCORKELL: There's also a cover your tush
19 aspect to that. If you make a low side override and the
20 loan goes bad, the boss may come back and say, why the
21 heck did you do that one. That was a stupid decision.
22 If you make a high side override and you turn it down,
23 that loan is not going to come back to haunt you, because

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
219.
219 Page 220 221
220
1 there's no performance there to come back and bite you.
2 MR. COOK: And just a footnote on what Russ
3 said, there are pet reasons. And DTI may, because the
4 GSE's have done studies and shown that DTI is not a
5 particularly --
6 MR. MEDINE: You may want to not use quite as
7 many acronyms here, even though we are in Washington.
8 MR. COOK: Debt to income ratio is not as
9 predictive a factor as, for instance, loan to value, and
10 therefore many scoring systems don't point up. Don't
11 weigh debt to income as heavily as, for instance, loan to
12 value, and therefore a low debt to income ratio might
13 pass the scoring system.
14 But a particular lender, perhaps a lender who is
15 still in the dark ages, may love debt to income. It's
16 his favor ratio. And despite the fact that it's already
17 been taken into consideration for its predictiveness in
18 the scoring system and passed, the lender decides no, I'm
19 not going to give a 44 debt to income ratio alone. And
20 that's a high side override.
21 MR. MEDINE: Pete, we didn't give you an opening
22 statement, but why don't you take a few minutes, if you
23 would like, to respond to some of the questions that have

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
220.
220 Page 221 222
221
1 come in before we open it up for discussion.
2 MR. MCCORKELL: Thanks, David. First of all,
3 this morning I mentioned the LMI and HMA studies we did.
4 They are written up in detail in a booklet. I sent some
5 of those ahead. They apparently got hoovered up by about
6 8: 30 this morning by the early arrivals. If you want one
7 of these and you didn't get a chance to pick it up, give
8 your card either to me or to one of the people on the
9 Commission staff and I'll make sure you get one.
10 Secondly, Margot, I think you actually mis --
11 kind of misquoted what Pam Johnson said. The three to
12 one was a relative ranking. Those were not default
13 percentages. She was saying that the low score folks had
14 a three times high default rate than the ones in the
15 average group that had the one as the reference point.
16 It wasn't one percent and three percent.
17 MS. SAUNDERS: The slide went by so fast, that's
18 very likely. But is the message not still the same?
19 MR. MCCORKELL: Well, yeah.
20 MS. SAUNDERS: It's still not a huge majority of
21 people in that last group?
22 MR. MCCORKELL: Sure. And I brought that out
23 this morning, that for most lenders in order to make

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
221.
221 Page 222 223
222
1 money, you've got to set your cutoff at a point where
2 there would be more goods if you accepted those people
3 than bads. But because the loss from a bad, however
4 that's defined, is going to exceed the profit from a
5 good, you can't wait until you're absolutely sure that
6 everybody you turned down would go delinquent or go into
7 default.
8 First of all, you can't know for sure how
9 individuals are going to perform. And if you try to
10 extend money to everybody, unless you are absolutely sure
11 they will go delinquent on you, you're going to be
12 bankrupt.
13 To address some of the things that Bob really
14 kind of threw my way early on, first of all I very much
15 agree with the concerns about disparate treatment
16 relative to overrides and quality of assistance. But I
17 would argue that those are not credit scoring issues. In
18 fact, if you're not using credit scoring, in a sense all
19 of your decisions are overrides.
20 The overrides are when you are ignoring what the
21 credit scoring system tells you. So I don't consider
22 that a credit scoring issue. It is certainly an issue
23 that remains for somebody using a credit scoring system,

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
222.
222 Page 223 224
223
1 but not using a credit scoring system doesn't make that
2 problem go ahead. It just makes it true for your entire
3 decision set.
4 In terms of a disparate impact analysis and
5 really defense of disparate impact challenges, for custom
6 systems, actually, the credit grantor would get the
7 statistics to sort of answer those questions that Bob
8 posed. For a system like the credit bureau system, in
9 fact, we do have the statistics that say that we put a
10 finance company in there because in the development
11 sample, people with finance company histories were three
12 times more likely to go delinquent on a subsequent
13 obligation than people with no finance companies. And we
14 will provide those statistics to a credit grantor if they
15 are challenged.
16 Finally, on the finance company issue, that's
17 probably the one factor that's in the credit bureau
18 system that people consistently ping on, because they're
19 convinced that it's unfair to minority and low income
20 borrowers. In fact, the research that we did and that is
21 detailed in here indicates that if you use a finance
22 company properly, which is to use it in terms of the
23 number of finance companies, rather than finance

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
223.
223 Page 224 225
224
1 companies as a percentage of total trade lines, there is
2 virtually no difference between high income and low
3 income and between minority and non-minority borrowers in
4 terms of the number of finance companies they have used.
5 So the argument that it even has a different
6 impact, while it certainly has a lot of intuitive appeal,
7 it does not seem to be backed up by the data.
8 MR. MEDINE: Well, I guess two more points that
9 were raised. One is the issue of utility and other
10 nontraditional credit reporting and what impact that
11 would have on low and middle income people. And then the
12 second one was raised by Russ about the trend among some
13 sub-prime lenders of not reporting to credit bureaus.
14 MR. MCCORKELL: Okay. Well, in both cases, you
15 know, Fair, Isaac's position would be that we would love
16 to have the data. And I'm sure that if you march down to
17 the Hill tomorrow and tell Congress that they ought to
18 pass a law requiring utilities to report to credit
19 bureaus, all of the credit bureau folks would be very
20 happy, and so would we. I don't know that that's
21 politically a realistic ambition.
22 There are certainly certain kinds of lenders and
23 others who are extending utility credit that simply don't

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
224.
224 Page 225 226
225
1 routinely report to credit bureaus, and some of that is a
2 cost issue, quite frankly. That data might very well be
3 predictive. The fact is that it's generally not there.
4 All that is there are the negative reports when somebody
5 goes into collection.
6 The same thing is actually true of a lot of
7 mortgage lenders. We know mortgage loans are vastly
8 under reported. Rental obligations are almost never
9 reported unless they are delinquent. We would love to
10 see them report it and collected in such a way that we
11 could explore those data. But right now the data are
12 simply not available in any way that is reasonable to get
13 at.
14 The same thing with the sub-prime folks not
15 reporting, apparently to try to protect the low risk sub-
16 prime customers from being picked off by non-sub-prime
17 lenders. I agree that's a problem, but I don't know what
18 to do about it. Again, if somebody thinks you can get
19 mandatory credit reporting through Congress, I'll testify
20 in favor of it.
21 MR. MEDINE: I guess the question that turns
22 around, though, is have you been able to assess what
23 impact it has on scores and their predictiveness to not

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
225.
225 Page 226 227
226
1 have the sub-prime lender reporting, and to not have the
2 utility and other information in the database.
3 MR. MCCORKELL: Well, it's hard to assess the
4 impact of information that you don't have and have never
5 had. The one --the other sort of non-reporting issues
6 that came up recently is some credit card issuers not
7 reporting limits. There for a long time we did have that
8 information. We did do an assess there, and it turns out
9 not to have a drastic impact on our credit scores to
10 date.
11 And I sort of need to qualify that a little bit.
12 Credit scoring systems, if they're well designed, are
13 designed to be fault tolerant in a sense, that if some
14 information is missing, the scoring system will still do
15 a good job of giving you an accurate risk assessment. So
16 if somebody takes a piece of information out of the
17 system, like balances on some credit card accounts, the
18 scoring system is designed to still continue to provide a
19 pretty decent risk indicator. And the indications are
20 that it's doing the job.
21 Now, it does --the loss of that information has
22 impacted the system slightly. No clear indication yet
23 that it's either favoring or not favoring consumers in

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
226.
226 Page 227 228
227
1 general, or favoring or not favoring particular classes
2 of consumers. It is making the system a little bit less
3 predictive. But I have to stress little.
4 Our big concern there is that that doesn't
5 become just the first loop of a death spiral, where we
6 get into this kind of tit for tat game among credit
7 issuers, were the first guy says, well, I'm not going to
8 report limits any more. And then somebody says, well,
9 I'm not going to report limits, nor am I going to report
10 high balance. And the next guy comes along and says,
11 well, I'm not going to report limits, high balance or
12 current balance. And all of a sudden not very much is
13 getting reported.
14 And certainly we are concerned about that. I
15 know the credit bureaus and ACB are concerned about that.
16 A lot of the regulatory agencies are concerned about
17 that. Again, the only foolproof solution may be
18 mandatory credit reporting. I'm all for it If you think
19 you can sell it down the street, let me know. I'll go
20 with you.
21 MR. MEDINE: Okay, thanks. Why don't we take
22 this chance to open up the discussion to the audience. I
23 know people from earlier --people have a lot of strong

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
227.
227 Page 228 229
228
1 views in this area. And if people would like to focus in
2 particular on the fairness issue with regard to scoring,
3 that would be great.
4 Why don't we want until you get a microphone,
5 and then once again, identify yourself, and please for
6 the reporter, spell your name.
7 MS. ISEMAN: It's Caryl Iseman. And, Peter,
8 this question is to you. The thing that puzzles me when
9 I look at reports --and this is a report I have right in
10 my hand. Under the scoring models that I'm looking at
11 for Trans Union and TRW and Equifax, one of the things
12 that this borrower is being dinged on is the number of
13 recent inquiries, and he has one. One. I have another
14 one that has none on it.
15 And I guess what I have, it always says --
16 MR. MCCORKELL: Has no inquiries and that's
17 still cited as a reason?
18 MS. ISEMAN: Yes. He has one inquiry.
19 MR. MCCORKELL: No, I'm sorry.
20 MS. ISEMAN: Thank you.
21 MR. MCCORKELL: One inquiry in fact is a little
22 bit riskier than no inquires. That's probably not the
23 first reason cited. It's probably the fourth, and they

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
228.
228 Page 229 230
229
1 are cited in order of importance. And the one inquiry
2 probably made very little difference in the score, but
3 the system is set up to provide the top four reasons,
4 whatever they are.
5 The third and fourth reasons may have very
6 little contribution to why the score is less than
7 wonderful, if in fact that's somebody that's got a less
8 than wonderful score. If it's somebody that's got a
9 bankruptcy, a whole string of delinquencies and one
10 inquiry, the one inquiry is not their problem.
11 MS. ISEMAN: Well, then I guess my question to
12 you would be, if one is too little and maybe ten is too
13 much, what's perfect?
14 MR. MCCORKELL: Well, unfortunately that all
15 depends. We talked about the segmented scorecards. One
16 inquiry for somebody that's got one trade line and has
17 only been in the file for six months may be more than
18 ideal. And again, understand that we're talking the
19 points off of maximum here. We're not talking about
20 what's average.
21 One inquiry in fact is average for the entire
22 credit bureau population. So the one inquiry there, they
23 may have gotten fewer --they would have gotten more

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
229.
229 Page 230 231
230
1 points if they had no inquiries instead of one. They
2 might have gotten three more points.
3 MR. SMITH: How do you know?
4 Well, I don't know, unless I have a chance to
5 review that credit report in detail and see everything on
6 that credit report and compare that against the
7 scorecards.
8 MS. ISEMAN: Okay.
9 FEMALE SPEAKER: Can I --
10 MR. MEDINE: Wait, one person at a time. We
11 only have one mic out.
12 MS. ISEMAN: Just going along with this, too,
13 because --and, yes, sometimes the three points does mean
14 something. You have --again I'm back to a required 640
15 credit score, and I have to use the mid score. And I've
16 got a 639 and a 637 and a 625, and the mid score might be
17 that 639. But I can't use it, and because of the 639 and
18 because this guy had one inquiry, he lost three points.
19 That's where my problem is, too. One of the
20 other things --and by the way, this is a good --this
21 particular person has over 700, but I'm just pointing
22 this out.
23 MR. MCCORKELL: Okay. They had --you know, 700

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
230.
230 Page 231 232
231
1 is not a perfect score. Almost nobody gets a perfect
2 score. Somebody could have an 880 and they would still
3 get four reasons back.
4 MS. ISEMAN: I understand that. I'm just saying
5 that I see this on --I just happen to have this report
6 in front of me. The other thing here that we have is
7 insufficient length of credit history. They have
8 accounts on here since 1994, and most of their accounts
9 are 1994 and 1995.
10 My question is --and they have quite a few
11 accounts that are paid off, etc. What is pp
12 MR. MCCORKELL: It would be better if they had
13 10 or 20 years of credit history instead of four or five.
14 MS. ISEMAN: So you're telling me that six years
15 --five years of credit history is insufficient?
16 MR. MCCORKELL: I'm telling you it's not ideal.
17 I'm telling you that that's why they got a 770 instead of
18 an 870. I'm not telling you that they are a bad credit
19 risk. There's no lender in this room who would tell you
20 that 770 is a bad risk, and you know that.
21 MR. MEDINE: Okay. Why don't we just get some
22 comments.
23 MS. ISEMAN: I guess the problem I have is the

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
231.
231 Page 232 233
232
1 inconsistency. Because I can pull an entirely different
2 report that also shows up where the people have credit
3 maybe two years, and I would not see a statement on there
4 that says anything about, you know, lack of credit
5 history.
6 MR. MCCORKELL: Well, in this case --in this
7 case you probably don't have any serious delinquencies,
8 because you're dealing with somebody in the 700's. And
9 so we're talking about why they're not perfect. Not why
10 they're not a good credit risk. They are a good credit
11 risk. We're talking about why aren't they --why didn't
12 they get a perfect score.
13 Somebody who has some delinquencies, some public
14 record items or some collections, the length of credit
15 history and the number of inquiries may be essentially --
16 they may be down --they may be the seventh and eighth
17 reasons on that report, not one of the first four.
18 MR. MEDINE: Well, let's get --
19 MS. ISEMAN: What I'm trying to point out --
20 MR. MEDINE: You know what --
21 MS. ISEMAN: It's not the perfection. It's the
22 scoring models don't make sense. And when you tell me --
23 and by just what you said, okay --that one is worse than

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
232.
232 Page 233 234
233
1 none, I have a problem with that. So we do have a
2 problem with the scoring models.
3 And can I ask --
4 MR. MEDINE: No. No, you can't. There are a
5 lot of other people who would like to chat. And if we
6 have some time, we'll certainly be happy to come to you.
7 If there is anyone who has not had a chance to speak
8 today who would like to make a comment? All the way in
9 the back. And just a reminder. If there are folks in
10 the overflow rooms who would like to come down and ask a
11 question, they're very welcome to do so.
12 If you could just identify yourself and spell
13 your last name.
14 MS. HUMPHRIES: Kim Humphries,
15 H-U-M-P-H-R-I-E-S. There are about three points I would
16 like to make. And, Mr. McCorkell, if you addressed this
17 in your comments on the finance company, I apologize. I
18 missed it. But I believe you were stating that there was
19 not really any statistical difference between the amount
20 of usage of finance companies?
21 MR. MCCORKELL: Number of finance companies.
22 MS. HUMPHRIES: The number.
23 MR. MCCORKELL: Which is the way it's in our

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
233.
233 Page 234 235
234
1 scorecard.
2 MS. HUMPHRIES: Okay.
3 MR. MCCORKELL: Now, if you had a characteristic
4 that was percentage of finance companies, or percentage
5 of balances on finance companies versus other kinds of
6 lenders, then there would be a difference. But that's
7 not how we use it in our bureau scoring system.
8 MS. HUMPHRIES: Well, I question the negative
9 viewpoint of that at all, given, I guess, you know,
10 particularly when a lot of those offer 60 days or six
11 months same as cash and, you know, the convention wisdom
12 of financial planners is using other people's money. If
13 you can have that free cash for six months, why not.
14 MR. MCCORKELL: Well, again, it's in the scoring
15 systems, because the historical data shows that as a
16 group people that use finance companies are a little bit
17 riskier than people who don't use finance companies. If
18 you remember from the presentation this morning, I listed
19 the five categories of credit bureau information that we
20 look at. Type of credit was the last one. It's the
21 least important of those categories. And so even where
22 it comes up, it's not a dominant feature in the credit
23 scoring systems.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
234.
234 Page 235 236
235
1 But statistically there is still evidence that
2 the use of finance companies indicates a higher level of
3 risk with respect to future credit performance. Now, I
4 can tell you that because of the 60 days same as cash or
5 the six months same as cash kind of programs that we're
6 seeing from a lot of retailers, where in fact if you sign
7 up for that you wind up with a finance company account,
8 that the predictive value of finance companies is going
9 down over time. But it hasn't gone away entirely.
10 MS. HUMPHRIES: Okay. And what is the lowest
11 credit score? Is there a minimum?
12 MR. MCCORKELL: Well, there are many different
13 credit scoring systems. In the Fair, Isaac credit bureau
14 systems, which is probably what you've got in mind, the
15 range is from somewhere in the low 300's to something
16 close to 900. Again, you don't see many 900's. You
17 don't see many 325's. That's kind of where the rubber
18 meets the road in terms of decisions. In most cases it's
19 somewhere in the 600's.
20 For most kinds of credit, 700 or maybe a little
21 bit up in the 700's. Anything above that is considered
22 golden for most kinds of credit. Anything below about
23 550 is considered awful. And so the range where

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
235.
235 Page 236 237
236
1 decisions are really being made typically is in that high
2 500's to low 700's range.
3 MS. HUMPHRIES: Okay. I think --I was looking
4 through some of the brochures I picked up outside this
5 morning, and several of them --and these are consumer
6 directed literature. Several of them state that, you
7 know, consumers have the right to everything in their
8 credit report, and that's not true. I think it's
9 deliberately misleading them. I mean, the credit score
10 is a part of the credit report, and they don't have the
11 right to that information.
12 And I guess what I haven't heard is a good
13 reason as to why not. Certainly it is helpful to have
14 the supporting information, sort of the narrative
15 explanation of what your credit report is and what on
16 your credit report is being a problem with you getting a
17 loan. But it's also helpful for the consumer to have the
18 actual score, so that they can evaluate for themselves
19 where they stand in relation to everyone else.
20 I really would like to hear, I guess, some
21 rationale as to why not. Why don't they have the right
22 to that?
23 MR. MEDINE: I think that's an excellent

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
236.
236 Page 237 238
237
1 question that will actually be discussed in the next
2 panel. The short answer is that Congress in 1996 changed
3 the Fair Credit Reporting Act to tell credit bureaus they
4 did not have to provide that information. But we will be
5 discussing the pros and cons of that in the next panel.
6 If there are any --did any questions come up
7 that any panelist wants to weigh in on?
8 MS. GOLDBERG: Yeah. I just wanted to kind of
9 speak back to this last question about the impact of
10 things like finance company usage and kind of take away
11 from you the statistical analysis and the defensive, is
12 credit scoring inherently good or inherently bad, or is
13 the FICO system or any other inherently good or
14 inherently bad.
15 Just kind of think of it from a bigger consumer
16 protection and fairness perspective. And I think --to
17 me it seems like it is very important for us to realize
18 that we are not starting from square one in this country.
19 We do not have a perfect economic system. We have a
20 system that has a long history of discrimination, and of
21 restricting the access of some people in this country to
22 mainstream credit and limiting them to what some people
23 have termed the fringe banking system with all that goes

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
237.
237 Page 238 239
238
1 along with that.
2 And I think one of the questions that we have to
3 answer in trying to figure out is this kind of approach
4 to allocating credit not --don't get nervous now,
5 regulators. You know, to making credit decisions. Let
6 me put it that way. A good one or a bad one, from kind
7 of a larger public policy question, is --you know, it's
8 important to understand what it means to have been
9 someone who lives in a community, or who for other
10 reasons has only been able to get high cost credit, who,
11 you know, maybe they're good and their credit record
12 doesn't show up in a credit bureau, because it is cash
13 payments, or because it's been sub-prime lenders who
14 aren't reporting that back to the credit bureau.
15 And do people get stuck in the fringes of the
16 banking system, where they are going to be paying more,
17 even though they are good at paying, and even though if
18 they could have credit on terms that were more
19 affordable, they would be even better at paying.
20 And to me, that's one of the big questions that
21 kind of remains unanswered. We tend to focus on the
22 statistical question of, you know, is this a good
23 predictor or is this not a good predictor. And not so

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
238.
238 Page 239 240
239
1 much on what is the impact on individuals' lives, and how
2 do we make it possible for everybody in this country to
3 get access to the best possible credit when they need it.
4 Best possible meaning the lowest cost and the safest
5 terms.
6 And I think that's something that we really need
7 to focus on more.
8 MR. MCCORKELL: Okay. I actually agree with you
9 very much there, Debby. You know, I may have a suit on,
10 but I'm actually a Berkeley grad, a Kennedy Democrat --
11 and I mean John Kennedy, Sr., not any of the current
12 crop. Well, maybe some of the current crop. Not all of
13 the current crop.
14 And clearly we're not starting from a level
15 playing field, and clearly I believe, personally, that
16 there is still subsidies and other forms of special
17 assistance that are going to be required to get people
18 out of that kind of vicious cycle. And unfortunately
19 subsidies or affirmative action have become a dirty word
20 in this town and across the country, and especially in
21 California.
22 So, you know, you've got to live with that
23 political reality, as much as it may gaul you. But

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
239.
239 Page 240 241
240
1 credit scoring is not the answer to those bigger social
2 issues. Credit scoring is intended to provide the most
3 accurate possible risk assessment that we can without
4 regard to race or gender or any other prohibited basis.
5 And all of the evidence is that it does that.
6 Having said that, there may be people that need
7 some help to get out of that vicious cycle that has still
8 large segments of the population as second class citizens
9 when it comes to economic factors.
10 MR. MEDINE: Any other first time commentors?
11 MS. MS. WELSH: Hi. Kristy Welsh, K-R-I-S-T-Y,
12 W-E-L-S-H, and I'm with creditinforcenter. com. And I had
13 a question about what Mr. Cook said, that the Federal
14 Reserve Board regulates or somehow evaluates the FICO
15 scoring.
16 How is this done? How is it tested? Are we
17 going through a statistical analysis? Do we take case A,
18 B, C and B and run them through the scoring? Hey, we
19 know for sure that it matches what it should match. And
20 if so --I mean, you said that this was done. Is this
21 published anywhere? Can we get a copy of it?
22 MR. COOK: Well, I may have mis-spoke. I don't
23 think I said that we evaluate credit scoring systems per

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
240.
240 Page 241 242
241
1 se. I will say that HUD has currently embarked on a
2 study of the scoring systems of the GSE's --Fannie and
3 Freddie.
4 No. I do --I can tell you that in 1996 --and,
5 Bob can correct me if I'm wrong, if he's still here --
6 the Fed did undertake an analysis of credit scoring
7 systems in the sense of trying to determine whether or
8 not there were solid correlations between low scores and
9 poor performance. And generally they found that there
10 were. That credit scoring systems do tend to be
11 predictive of behavior.
12 But that leaves a lot of room, I think, for
13 issues today.
14 MS. WELSH: Right. So they haven't been
15 evaluated for bias or discrimination or any of these
16 kinds of things?
17 MR. COOK: No. This was not an exhaustive study
18 of each factor in a given scoring system and a
19 relationship. I guess one of the questions that I didn't
20 quite follow up on with Pete, you stated that the
21 statistics are available for your clients to understand
22 and appreciate the various risks associated with each of
23 the factors, so that if they're called upon to justify

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
241.
241 Page 242 243
242
1 the inclusion of a particular factor, it's predictiveness
2 can be quantified for them based on the statistics that
3 you've done.
4 But let me ask you this, sort of a further
5 question. Recognizing --sort of taking off on Debby's
6 approach --that we're talking perhaps more than just the
7 legal requirements here. Recognizing that probably you
8 don't have a legal obligation to do the rest of the
9 analysis, to do the third stage of a --what did we call
10 it --disparate impact analysis, that while there may be
11 a justification for using something like finance
12 companies --a number of finance companies --is that the
13 least --does it produce the lowest impact, or is there
14 another set of factors that you could have used that
15 would produce a lesser impact?
16 MR. MCCORKELL: Well, of course the lesser
17 impact is only half of the third leg of a disparate
18 impact case. Is there something that is equally
19 effective but has a less discriminatory impact. And with
20 the dearth of data on protective classes, especially race
21 and ethic origin, accept in mortgage lending it's very
22 difficult to do that. Is there a less discriminatory
23 alternative analysis?

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
242.
242 Page 243 244
243
1 We have tried to do that using the zip code
2 surrogate that I talked about this morning in the HMA
3 study and, indeed, in a sense our choice of number of
4 finance companies rather than some other way that we
5 might have used finance companies, reflects that kind of
6 decision. We said if we use number of, it won't have a
7 disparate impact, whereas if we use percentage of trade
8 lines that are finance companies, that, at least
9 arguably, looks more like it has a disparate impact.
10 So where we have any evidence of that sort, we
11 certainly can take advantage of it to try to minimize
12 that problem for ourselves and our client. But, in fact,
13 probably in most cases, the leg that we would expect to
14 stand on is that if we've done the best job of selecting
15 the most predictive set of characteristics.
16 And the weighing for that set of
17 characteristics, you're going to be hard pressed as a
18 regulator or as a private plaintiff to come along and say
19 here's an alternative that's less discriminatory but
20 equally effective, because we think we've kind of
21 foreclosed the equally effective part of the argument by
22 building this system that's most predictive.
23 MR. MEDINE: We have time for maybe one or two

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
243.
243 Page 244 245
244
1 more comments. I see Frank Torres from Consumers Union
2 in the back. Welcome.
3 MR. TORRES: Oh, good. That means I don't have
4 to say my name and spell it then?
5 MR. MEDINE: T-O-R-E-E-S.
6 MR. TORRES: I just have three questions, and
7 they're all interrelated, to the whole idea behind
8 garbage in and garbage out. And the first goes towards
9 an issue that I don't think has been adequately addressed
10 here, and maybe we need to have a whole other workshop on
11 it.
12 And that's the ability for consumers to easily
13 fix problems that they see in their credit report. And I
14 know Congress has done some things, and the FTC has been
15 good on that. But I think that will help ease some of
16 the pain of consumers, who when they go and apply for
17 credit and they're subjected to higher rates and higher
18 fees because of their credit scores --the credit report
19 coming back and there are some problems with it, and then
20 you seem to be at the mercy of the credit issuers for
21 reporting things in --to try to get it fixed. And I
22 think there is still a lot of confusion and concern about
23 that.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
244.
244 Page 245 246
245
1 The second point is, will it make any difference
2 if your good payment history is reported if the factor
3 that lowers your score is the fact that you've gone to a
4 finance company? Because if that's the case, then we're
5 talking about perhaps a bigger difficulty in the system
6 that doesn't seem to be quite logical, at least to me.
7 If I'm pulling myself up from the bootstraps, I
8 may not have a credit history to begin with, so that's
9 almost a strike against me, that I haven't gone out there
10 and gotten a lot of credit, because I may not trust
11 credit card companies.
12 MR. MCCORKELL: You're better off to have a good
13 history with a finance company than no history at all.
14 MR. TORRES: Yeah. But there are still problems
15 if I have a history with a finance company rather than a
16 bank, though, right? Is that right?
17 MR. MCCORKELL: Well, you would be better if you
18 had a good history with a bank than if you had a good
19 history with a finance company. But, again, I'm speaking
20 in terms of --you know, in relative terms. A good
21 history with a finance company and everything else looks
22 goods, you may be just fine.
23 MR. TORRES: And my final question is --and I

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
245.
245 Page 246 247
246
1 just, you know, was in some discussions over lunch, and I
2 saw in some literature that actually if you are a
3 consumer and you want to try to do kind of the
4 responsible thing in some cases --and that is, you know,
5 perhaps consolidate some of your credit cards because
6 you've gotten a lower --you've gotten one of the --
7 you've taken advantage of a teaser rate or something,
8 let's say, and you think you can kind of work the system
9 so that you are lowering the cost of your credit. And
10 you actually consolidate maybe three or four cards into
11 one.
12 And that actually can hurt you on your credit
13 score, is that right?
14 MR. MCCORKELL: Could it? The answer is yes.
15 Would it always? No. Again, I would have to see the
16 entire credit report, and I could tell you if this person
17 does this and nothing else changes, then here is the
18 effect that consolidating those three accounts into one
19 would have. But as I mentioned this morning, something
20 else will change, and I can't tell you what will be,
21 because of the passage of time.
22 MR. MEDINE: Let's put it another way. It is
23 partly the relationship between credit balances and

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
246.
246 Page 247 248
247
1 credit limits?
2 MR. MCCORKELL: Yeah.
3 MR. MEDINE: Consolidation could actually give
4 you a much closer tie to those two if you pull all your
5 debt into one account?
6 MR. MCCORKELL: Yes. Yeah.
7 MR. MEDINE: Okay. We have --Frank, did you
8 have another question?
9 MR. TORRES: No, that's it.
10 MR. MEDINE: Okay.
11 MR. TORRES: I think it just goes to the point
12 that especially if there is going to be a movement
13 towards more risk based pricing that there is still a lot
14 of seemingly unanswered questions or kind of quirky areas
15 that need to be resolved if consumers are to be helped
16 and not hurt by this system.
17 MS. SAUNDERS: David?
18 MR. MEDINE: Sure, Margot?
19 MS. SAUNDERS: I would like to respond or follow
20 up a little bit on what Frank was saying. I think it
21 disturbed a lot of us to hear Peter say that you're
22 better off having a good history with a bank than you are
23 having a good history with a finance company.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
247.
247 Page 248 249
248
1 There are some --obviously, I would guess, that
2 you have found that the statistics are that people who
3 have a good history with a bank will have a higher
4 probability of paying on time than those who have a good
5 history with a finance company?
6 MR. MCCORKELL: Yea.
7 MS. SAUNDERS: So you have numbers to back up
8 that statement. But intuitively it feels wrong, and
9 that's --let me give you some reasons why it feels
10 wrong. I think some of it has been stated already. One
11 is that there are no banks in low income neighborhoods,
12 so the people who have no choice but to go to finance
13 companies are getting dinged because that choice is not
14 wrong --that choice is not available.
15 Number two --and number two is, in some ways a
16 good history with a finance company borrowing the same
17 amount of money should give you a better credit score
18 than the same --than a good history with a bank --
19
(Applause.)
20 MS. SAUNDERS: Because --
21 MR. MCCORKELL: Well --okay.
22 MS. SAUNDERS: Because you're paying a higher
23 rate, so that repaying the same amount of money is that

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
248.
248 Page 249 250
249
1 much more difficult. Then when you add to that the
2 factor that you're probably poorer than the people who
3 are borrowing the same amount of money from the bank.
4 This is why it feels wrong.
5 MR. MCCORKELL: I understand why it feels wrong.
6 All I can tell you is that the data don't support the
7 argument in terms of predicting future credit
8 performance.
9 MR. MEDINE: Why don't we --I'm going to --
10 right. We're going to leave it --I think Margot's
11 question and Pete's response demonstrate that the issue
12 of fairness --that this is just the beginning of a
13 discussion and not the end. But it's the end of this
14 session.
15 We're going to take a 15 minute break, and we'll
16 come back again. Thank you.
17
(Brief recess.)
18 MS. TWOHIG: Thanks. We're running a little bit
19 late, but I think what we'll do is, we'll try to give
20 this discussion --give it its full due. And I think
21 we'll just probably make closing remarks pretty brief.
22 So I think we'll still be able to cover the subject
23 adequately, which, of course, has already been touched

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
249.
249 Page 250 251
250
1 on.
2 Our last panel is going to talk about what
3 information should consumers receive about credit
4 scoring, and who should provide it to them. What should
5 they know and who should tell them about it. Some very
6 strong feelings have been expressed so far. Why
7 shouldn't consumers know their scores. Why shouldn't
8 consumers learn the factors that go into that. And those
9 are some of the questions I would like these panelists to
10 address.
11 In addition, one question I have is, how do
12 whatever consumers know or don't know about their scores
13 --how does that relate to shopping and whether they can
14 get the best price, or know what the best price is they
15 could get. So those are some of the questions we would
16 like this panel to address. And then as we did before,
17 we'll open it up for questions from the audience.
18 One thing I would like to say just as background,
19 just so we all understand the legal situation, as David
20 mentioned earlier, the Fair Credit Reporting Act
21 currently states that consumers have a right to all
22 information in their credit file, except their credit
23 scores.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
250.
250 Page 251 252
251
1 But something that has been eluded to by some of
2 the speakers, I want to make sure everyone also
3 understands that there is nothing in the Fair Credit
4 Reporting Act that says that consumers --that the credit
5 bureaus cannot voluntarily give consumers their scores,
6 nor is there anything that says that lenders could not
7 voluntarily choose to give consumers their scores.
8 So just because the law says consumers don't
9 have a legal right to it doesn't mean that that could not
10 be done as a matter of voluntary choice.
11 With that background, we would like to start
12 with Virginia Ferguson. Ginny is Vice President of the
13 National Association of Mortgage Brokers. She has over
14 18 years experience in the real estate industry, and is a
15 loan originator and currently is the co-owner and broker
16 of Heritage Valley Mortgage in California.
17 She has acted as co-chair for the National
18 Association of Mortgage Brokers Credit Scoring Task Force
19 for the past three years, and is an industry spokesperson
20 and trainer on credit scoring.
21 Ginny?
22 MS. FERGUSON: Thank you very much, and thank
23 you for having us share and participate in this forum

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
251.
251 Page 252 253
252
1 today. I think if we've learned nothing else today, it's
2 that credit scoring is a very large mystery to everybody
3 in the industry. The lack of education out there is
4 incredible on both sides of the fence, whether you are a
5 consumer or whether you are a member of our industry.
6 To that end, this panel was asked to address who
7 should be educating the consumer. And whether that is a
8 consumer counselling group or the front line mortgage
9 originator that is meeting with that consumer, that
10 individual has to make sure that they are extremely well
11 schooled in what makes --goes into credit scoring and
12 what impacts the score. Because until you can explain
13 that thoroughly to a consumer, there is no point in even
14 reviewing a credit report with them. All you're going to
15 do is confuse them.
16 And unfortunately today they are being bombarded
17 with bad information from all venues, and the media being
18 one of them. Many of the consumers out there are
19 arbitrarily being told to go out and consolidate credit
20 before anybody reviews what their credit picture looks
21 like. They are being told to close credit card accounts
22 without someone reviewing what that is going to do to the
23 overall picture and how that may or may not skew the

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
252.
252 Page 253 254
253
1 overall perception of their credit utilization.
2 And, in fact, consolidation is not a bad thing.
3 But if it artificially skews what your credit utilization
4 looks like to a system and causes a negative impact to
5 the consumer, they've done something with the best of
6 intentions, but because they have not been well educated
7 as to what they should be doing, they're hurting
8 themselves.
9 And there are many instances today that because
10 of lack of education on the part of any or all of us in
11 this industry, when credit scoring came to the
12 marketplace three years ago, we have allowed the
13 consumers to do a lot of things that have not necessarily
14 been in their best interest.
15 So to that end, we first need to, in our
16 industry, to be thoroughly educated about the ends and
17 outs of credit scoring. Our Association has put out a
18 piece that is geared specifically to explaining credit
19 scoring in simple terms to the mortgage professional. We
20 have another piece that is being done to explain credit
21 scoring and the ends and outs of credit scoring to the
22 consumer in lay person terms, that tells them what it's
23 about, what goes into it, how it's weighted and the

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
253.
253 Page 254 255
254
1 things that they can do to help themselves or hurt
2 themselves. It's their choice.
3 One of the things that consumers have to realize
4 out there is that they have to be proactive in what they
5 do with their credit. But they have to know what they're
6 doing first before they can do that.
7 And so that front line person needs to be able
8 to sit down with a credit report and review that credit
9 report for all aspects of that credit report, whether
10 it's good, bad or indifferent. I heard someone earlier
11 today talking about the fact that many times, I think,
12 many originators do not worry about something being on a
13 credit report that really is inaccurate if the score is
14 fine and they can get the loan done.
15 We should be reviewing good credit as well as
16 bad credit, because while it's good today, that doesn't
17 mean that just because it's good today it isn't going to
18 necessarily be bad six months from now. So that consumer
19 needs to review their credit report in full and make sure
20 that all of the information that is reported there should
21 be reported there.
22 The next issue that we have to deal with is
23 explaining the other components of credit scoring to the

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
254.
254 Page 255 256
255
1 consumers, so that they aren't doing things like taking
2 five charge cards and consolidating them into one charge
3 card and closing out the rest of the charge cards,
4 thereby looking as though they have maxed out the only
5 card in their life. And the next time they have an
6 inquiry, it looks like they're looking to live on credit.
7 That skews the perception of somebody's credit risk.
8 Credit scoring in and of itself is a very fine
9 tool. And I think one of the things that we all have to
10 remember is that underwriting didn't start with credit
11 scoring. We have been underwriting for years, and we've
12 had all sorts of benchmarks to try and measure people's
13 capability of repaying a loan against. For years in the
14 industry we used a very huge measure, which was called
15 debt to income ratios --percentages of gross income --
16 that someone was earning to identify whether the payment
17 that they were asking for on an application was good, bad
18 or indifferent.
19 Now, for many years I butted heads against many
20 underwriting walls, because I had a consumer that was
21 already making the payment that they were looking to
22 acquire as a rent payment with no tax write off ability,
23 only to be told that because it was outside of guidelines

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
255.
255 Page 256 257
256
1 that they either had to come in with more money or accept
2 a lesser quality loan. That really didn't seem too
3 terribly beneficial to me, and it made no common sense
4 whatsoever.
5 But because of credit scoring today, we no
6 longer have to deal with that kind of nonsense. Ratios
7 are not as much of an issue as they have been in the
8 past. So we do have positives to this scenario as well
9 as negatives, and I think we need to remember that as we
10 move through this.
11 The person that needs to explain this is the
12 person that understands underwriting guidelines as well
13 as what is available out there. How can the consumer
14 know whether they're getting the best pricing and what
15 they should be getting. They're going to have to shop,
16 and they're going to have to ask questions. And that
17 goes along with the actual education process about what
18 loan products are available to them and what their
19 choices are.
20 And, again, if they're dealing with a good, well
21 educated, competent loan originator or counselor up
22 front, that person's job is to educate the consumer to
23 all of their options and what the repercussions are of

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
256.
256 Page 257 258
257
1 those choices that they may make, not to twist their arm
2 to do something bad to them. And that may sound very
3 Polyanne-ish, but that is the way I believe probably 95
4 percent of the people in the lending industry function.
5 Somebody asked me if the repositories should be
6 the ones explaining credit scoring. The tough thing for
7 a repository to explain as far as credit scoring goes is
8 they don't have the personnel that are schooled in
9 underwriting criteria, so that's a very major reason not
10 to have them explaining credit scoring. But to explain
11 an exact score from the repository level, they have no
12 benchmarks to put it against. It has to be explained in
13 tandem with the underwriting criteria.
14 A score taken out of context is a snapshot that
15 unless it has a reference point means nothing. So when
16 you as a loan originator sit down with a consumer and
17 take that number, you must take that number in reference
18 to some type of loan product. You must, in addition,
19 rather than just saying your score is good, bad or
20 indifferent, utilize those reason codes that are there to
21 assist you in why the score was not higher than it
22 actually could have been.
23 Now, the thing about reason codes, scores, etc.,

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
257.
257 Page 258 259
258
1 the higher the score, the less impact the reason code
2 has. So if we're dealing with somebody that has an 800
3 credit score, the reason codes may be very, very minute.
4 They're simply telling you why that person didn't score
5 900. Does it make a difference? Heck no. But at 620 a
6 reason code is going to have a substantial impact, and
7 it's going to give that loan originator a reason to be
8 able to point to and say, Mr. Consumer or Mis. Consumer,
9 this is what you need to chance in your credit profile to
10 be able to be on track to start to improve your credit
11 picture and the availability of the best credit out there
12 for you.
13 Again, it has to be the consumer's choice. If a
14 consumer chooses to make their payments pass the due date
15 on a regular basis, with the impression that because they
16 have paid the late charge it's no big deal, they have to
17 understand that they are going to represent to a lender,
18 and the people that are putting up the money for that
19 lender to make that loan with, a higher risk. Because
20 somebody is going to have to chase them every month, and
21 that means they're going to have to pay a salary to get
22 that somebody to remind them every month to make the
23 payment.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
258.
258 Page 259 260
259
1 If they have someone who has consistently paid
2 their payments on time, whether that individual makes
3 $1,800 a month or $7,500 a month, they made an agreement.
4 They made a pact with someone when they signed a contract
5 for credit, and they have followed through on that. They
6 have not just decided that it's okay to be late if I pay
7 somebody an extra ten bucks a month. That's being
8 proactive and consumers need to understand that.
9 So front line people have to be there. It's not
10 something that a repository can do. Why do consumers not
11 get a score on the report that they get from the
12 repository? Because it has no point of reference. They
13 can get a number. What does that mean to the average
14 person on the street? They don't know whether it's good,
15 bad or indifferent, and they don't know what to do with
16 it once they have it. It has to come on a full report,
17 through a lender, so that the lender has the opportunity
18 to review it with them, otherwise they lose the benefit.
19 As far as correcting errors, that is also that
20 counselor's job to instruct them on how to go about
21 correcting those errors. And the Fair Credit Reporting
22 Act that went into place in October of 97 gives a very
23 defined time line on what is supposed to happen.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
259.
259 Page 260 261
260
1 Now, we all know that in the real world what is
2 supposed to happen happens most of the time, and what's
3 supposed to happen does not happen all the time. But you
4 can't say that it's all bad simply because it doesn't
5 always work. We have to continue to look at those
6 issues, take those infractions and work toward making the
7 system better.
8 If in fact --as an originator I know for a fact
9 that I get credit reports that have errors on them, just
10 like anybody else in this industry. That is not the
11 repository's fault. It's a trade line reporting to the
12 repository. However, the repositories do have a certain
13 amount of input in their relationship with those trade
14 lines.
15 MS. TWOHIG: Ginny, if you could wrap up?
16 MS. FERGUSON: Will do. And they can in fact
17 start to create those kinds of expectations from their
18 trade lines.
19 So I think we have a place to go with this. It
20 has to start at the front line and work its way through.
21 It has to be a partnership in our industry. But I think
22 that education for everybody in the industry is where we
23 have to start.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
260.
260 Page 261 262
261
1 MS. TWOHIG: Thanks, Ginny. Just one point of
2 clarification. You said you think it should start at the
3 front line and that the repositories are not the best
4 places for the consumers to learn their scores from or
5 talk about their scores.
6 Do you think lenders and counselors and brokers
7 should tell consumers their scores?
8 MS. FERGUSON: I think that they need to tell
9 them the score in relationship to the reason code. You
10 can't just hand somebody a number.
11 MS. TWOHIG: Okay.
12 MS. FERGUSON: Because as we've all heard today,
13 there are various types of scores, and every investor out
14 there uses a different scale to benchmark.
15 MS. TWOHIG: And just for a point of
16 clarification also, because not everyone might know the
17 terminology. The reason codes that you are referring to
18 are the reasons for a denial under Regulation B?
19 MS. FERGUSON: Absolutely not. On a credit
20 report --when a score is issued from each repository on
21 the credit report, it is followed by four reason codes or
22 four reason explanations. And those are listed in the
23 order of significance of weight as to why a score did not

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
261.
261 Page 262 263
262
1 come in at the maximum. It is not a denial.
2 MS. TWOHIG: Okay. So that's available even if
3 the loan is going to be granted?
4 MS. FERGUSON: Oh, heavens, yes.
5 MS. TWOHIG: Because, as some of you know from
6 the consumer standpoint, the law does not entitle the
7 consumer to know those reasons unless the loan is being
8 denied. If there is a counteroffer situation and the
9 consumer accepts whatever counteroffer is given, there is
10 no legal right under the Equal Credit Opportunity Act to
11 know those reason codes.
12 But, again, like the Fair Credit Reporting Act,
13 that does not mean that it can't be told to the consumer
14 in those situations.
15 MS. FERGUSON: And I think that's just common
16 good business practice. If we are in the industry to
17 work with consumers, our job as the professionals is to
18 educate. And we can't do that if we don't share that
19 information with them.
20 MS. TWOHIG: Thanks, Ginny. Our next panelist,
21 Ray Crescenzo is going to speak. Ray has been with the
22 Associated Credit Bureaus since 1981, when he joined the
23 staff as Director of Education. Today as Vice President

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
262.
262 Page 263 264
263
1 of ACB, Ray is responsible for industry relations. He
2 serves as staff liaison to ACB standards Committees and
3 task forces, the credit reporting, mortgage reporting and
4 specialized reporting divisions, the federal insuring and
5 lending agencies, and various credit grantor and mortgage
6 lending interested groups.
7 And we are trying to focus in this panel on what
8 information should be given to consumers about scoring,
9 but I've also told Ray that since there has been quite a
10 few comments so far today about credit bureaus, he should
11 feel free to respond to those comments at this time.
12 MR. CRESCENZO: And what did I say --
13 MS. TWOHIG: You said --
14 MR. CRESCENZO: --to all those questions?
15 We'll see what I can do. First, let me start by saying
16 that we've been attempting to listen to what's been going
17 on in the mortgage lending arena for the last ten years,
18 it seems. This is an internal process that you all have
19 adopted and turned into a public exercise in frustration.
20 Internal credit scores are the risk management
21 tools of thousands of major credit grantors around the
22 country. Pete won't tell you this, but they start with
23 FICO scores and then they figure that they can have their

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
263.
263 Page 264 265
264
1 own internal expertise hired, and they begin to develop
2 their own scores.
3 So one of the challenges that the Federal Trade
4 Commission had earlier in this decade was whether a risk
5 score should be included --a credit score should be
6 included in a credit report. That was a can of worms.
7 It was difficult to deal with because of the myriad
8 numbers of score models that were out in the marketplace,
9 and how can one person attentionally explain in the
10 details that you require and you need in a manner that
11 would allow us as credit reporting agencies to do that
12 for the consumers that are in the marketplace. That
13 exercise resulted in credit scores not being included
14 with the new law in 1997, the new Fair Credit Reporting
15 Act. Credit scoring is not included as part of the
16 credit report.
17 So that in and of itself is the challenge. It's
18 an internal risk management took that now has become very
19 public. We have been trying to stay ahead of the
20 evolution. Stay up with the evolution of its
21 implementation in the mortgage community. In the course
22 of doing that, we've been working with organizations such
23 as the American Homeowners Education and Counselling

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
264.
264 Page 265 266
265
1 Institute, Fannie Mae, Freddie Mac, the National
2 Association of Realtors, the National Association of
3 Mortgage Brokers, the Mortgage Bankers Association. All
4 of the industry players that are being impacted by this
5 new technology. This new information --this new
6 knowledge based risk management exercise requires
7 communication and education.
8 We've been trying to facilitate that
9 communication and education process. Not getting in the
10 way of it, but making sure that the people that are
11 trying to develop these messages so the consumers can
12 understand them have consistent information coming out of
13 FICO or Trans Union or Equifax or Experian or ACB.
14 Consistent messages so that consumers aren't getting told
15 one here and another thing there. Consistent messages so
16 that the education process can begin.
17 I won't steal Forrest's thunder. You know, he
18 might think that the realtor should be on-line and
19 talking to the consumer about this process and about the
20 scores that they have. There are --this is not a lack
21 of interest or willingness on the part of all the
22 participants who take a part in this. So it's not a
23 single contact point solution, I don't think, and we've

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
265.
265 Page 266 267
266
1 been trying to work with all participants in trying to
2 enhance the communications that takes place.
3 One of the questions that came up earlier had to
4 do with credit data --credit grantor information that's
5 coming into the credit bureaus. This is a three legged
6 stool. I've been in this industry for over 20 years, and
7 I've been saying this for 20 years. It's a three legged
8 stool.
9 The credit granter has a responsibility to
10 report complete and accurately. The consumer has a
11 responsibility, especially if they're going to be credit
12 active, to check out their credit report. See what's
13 inside of it and make sure that it's accurate on a
14 regular basis, on an annual basis or if you --you know,
15 obviously with all that you've learned today, you would
16 probably want to do that prior to getting involved in a
17 home purchase.
18 But if there is something that goes beyond the
19 home purchase, you want to check your credit to make sure
20 that it is accurate and ready for when you need it. You
21 deserve to use your good credit when you need it, and in
22 order to protect it, you have to look after it.
23 The other leg of the stool is the credit

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
266.
266 Page 267 268
267
1 reporting agency. We have a responsibility to make sure
2 we're doing our jobs also. There hasn't been a year at
3 ACB in the 18 years that I've been there that we haven't
4 had complete and accurate reporting as not only a goal,
5 but have had programs to facilitate complete and accurate
6 reporting. As recently as March and April, we have
7 programs for major credit grantors around the country.
8 Two years ago we had programs similar to this to educate
9 credit grantors and how to report completely and
10 accurately. We're putting it into the budget for the
11 year 2000. It's a continual process.
12 We have tried to facilitate also an automated
13 system to facilitate consumer disputes being handled in
14 an automated basis. There are over 20,000 data
15 furnishers. There are over 20,000 credit reporters --
16 excuse me --credit grantors reporting to the bureaus.
17 They are --we have over 1,000 mail boxes on the system
18 right now, and that doesn't --that's just over 50
19 percent of the disputes that come into the credit
20 bureaus.
21 So we are challenged now. And we've just moved
22 forward with a RFP today. It's going out next week to
23 enhance that system. We want to remove barriers to its

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
267.
267 Page 268 269
268
1 usage by the credit granting community. We don't want
2 them to have to use its software, and we don't want them
3 to have to sign contracts with the network. All the
4 barriers that we have identified to them getting on the
5 system are trying --we will be trying to remove those in
6 our business solution that will be coming out within the
7 next 12 to 18 months. By the end of next year --by the
8 end of year 2000 it will be implemented.
9 So it's --I said it's an evolving process. We
10 are trying. We are listening. You told us that
11 inquiries were a problem relative to too many when a
12 person is shopping for a mortgage and that's going
13 against the consumer. We worked with the national
14 systems and the mortgage brokers and mortgage bankers to
15 get FICO to understand that issue, and we changed that
16 and within six months they were blocked.
17 We're hearing now that some other things need to
18 be considered, and we're listening again. So that's it,
19 Peggy.
20 MS. TWOHIG: Thanks, Ray. I have one question
21 about something Peter Mahoney said this morning. When he
22 was arguing that the black box was a myth, which, I
23 think, a lot of folks in the room since disagreed with,

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
268.
268 Page 269 270
269
1 he said that you can contact each of the three
2 repositories and they will give you what he called a
3 feedback list. I'm not sure if that's the same as a
4 factor list.
5 And I was just wondering if you know anything
6 about that?
7 MR. CRESCENZO: The list of factors that are
8 part of --that could be construed as a full list?
9 MS. TWOHIG: I assume that's what he was
10 referring to. Yeah. That's what he was referring to.
11 MR. CRESCENZO: Well, I have a full list right
12 here.
13 MS. TWOHIG: So that is the case, that those
14 factors that are used by the --
15 MR. CRESCENZO: The whole list is available.
16 MS. TWOHIG: Okay. I just wanted to clarify
17 that. And if anybody called and asked for that, would
18 you mail it out to them?
19 MR. CRESCENZO: I would.
20
(Laughter.)
21 MS. TWOHIG: Call Ray.
22 MR. CRESCENZO: We have been also --one of our
23 other staff members has been working with the scoring

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
269.
269 Page 270 271
270
1 development folks to try to expand the comment factor
2 codes --the explanation codes --so that there is an
3 educational message besides just the statement of lack of
4 payment. So it's, you know, more of the evolution
5 process that continues.
6 MR. MEDINE: Ray, do you want to make that part
7 of the record of this, so that people can have access to
8 it and have a better understanding of how the process
9 works?
10 MR. CRESCENZO: The list?
11 MR. MEDINE: Yes.
12 MR. CRESCENZO: Yeah, sure.
13 MR. MEDINE: Okay.
14 MS. TWOHIG: And Ginny already stated that she
15 doesn't think it would be helpful for consumers to obtain
16 their scores from the repositories, that it needs to be
17 from a person who understands the score. And I take it
18 you would agree with that?
19 MR. CRESCENZO: Yes, at the delivery point.
20 MS. TWOHIG: Right, at the delivery point. But
21 what would be the problem, though? Would it be so
22 horrible for the consumers to see that score when they
23 got their credit report? I think somebody posed that

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
270.
270 Page 271 272
271
1 question earlier.
2 MR. CRESCENZO: When they got their credit
3 report at the re-seller location or at the --
4 MS. TWOHIG: No. If they just got their credit
5 report. They order their credit report. Would it be so
6 terrible if they see what scores had been calculated on
7 them? Or what are the problems with that?
8 MR. CRESCENZO: Well, again, what scores? Was
9 it Providian out in San Francisco that believes they have
10 the best analysts and statisticians in the country and,
11 you know, produce scores for their own lending decisions,
12 or are we talking about Chase Manhattan's scores, or are
13 we talking about a FICO score?
14 MS. TWOHIG: FICO, I think people wanted to
15 know. If it was just a FICO score?
16 MR. CRESCENZO: It's an issue I'm not going to
17 be able to address today.
18 MS. TWOHIG: Okay. We'll continue with our
19 panelists. Our next panelist is Forrest Pafenberg, who
20 is Director of the Real Estate Finance Research in the
21 Research Division of the National Association of
22 Realtors. In that position he has responsibility for
23 conducting research on primary and secondary market

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
271.
271 Page 272 273
272
1 mortgage finance, conducting quarterly surveys of current
2 financing techniques for both residential and commercial
3 real estate, and preparing materials for members of
4 Congress and their staffs concerning finance issues.
5 Forrest?
6 MR. PAFENBERG: Hello. Thank you, Peggy. I
7 would like to thank the FTC --or commend the FTC for
8 holding this hearing. I think it's very important. I'm
9 going to skip a lot of things I was going to say, and
10 talk just --address a couple of issues.
11 First, I would like to pose the question of why
12 disclose mortgage credit scores. And I would say home
13 buyers need to know their credit scores with meaningful
14 explanatory data to better shop for mortgages and to
15 basically evaluate lender products.
16 I would also say that credit score disclosure
17 legislation is pending in at least California state
18 legislature, so this is not a trivial issue. It is, I
19 think, in a harbinger of a new direction that automated
20 underwriting is taking as it becomes more completely
21 integrated into mortgage lending transactions.
22 Another reason why I think mortgage scores
23 should be --mortgage credit scores should be disclosed

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
272.
272 Page 273 274
273
1 is effectively some lenders are disclosing the credit
2 scores, and I think it's a fairness issue as much as
3 anything else.
4 The National Association of Realtors supports
5 disclosure of the credit score if it would lead the
6 mortgage lender to fund or not fund the mortgage
7 application. And it raises a question of who discloses
8 what to whom and when. Ideally the consumer should be
9 the whom, and the lender should be the who. I'll speak
10 about the what in a minute. And the when should be
11 generally when the borrower is shopping for a mortgage.
12 The details of the disclosure clearly should be
13 worked out and agreed to by the credit reporting industry
14 and lenders. However, the benefits to the consumer
15 should override any other issues in there. Consumers
16 need to be fully informed as they make the decision to
17 accept a mortgage that is being offered by the lender.
18 And I want to just clarify what I mean by
19 disclosure. Disclosure is not an explanation of how a
20 credit scoring model works, but what the effects the
21 score will have on the terms and conditions of the loan
22 that would be offered to the consumer.
23 And let me just shorten up a lot and just talk

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
273.
273 Page 274 275
274
1 about what should be disclosed, because I think this is a
2 very, very complicated issue. Now, clearly the score
3 should be disclosed. But along with that score, you need
4 to know what is the scale of the model that's being used.
5 If you're using a scoring model, the type of the model,
6 i. e., is this a credit scoring model, is this a mortgage
7 scoring model, and I'm surprised not to have heard it
8 mentioned, a debt scoring model, which is effectively
9 looking at your credit. Excuse me, your credit card
10 behavior, which is a model that's been developed and is
11 in use for those with what are called the thin file,
12 i. e., little credit history.
13 In addition, you need to know who is the vendor,
14 i. e., whose model is it. Is this Desktop Underwriter or
15 Loan Prospector? Is it developed by an MI? Is it
16 developed by a lender? You also need to know what data
17 was used in the scoring. Now, again, it hasn't been
18 discussed in detail, but you can generate using the same
19 information multiple score values from the different
20 repositories, primarily because (a) the models are
21 slightly different, if it were even FICO model, and (b)
22 the data that is in the repositories does vary by
23 geographic area.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
274.
274 Page 275 276
275
1 So that not all of the information that's in one
2 repository is in the other. So it's extremely important
3 to know what data is being run on what model and what
4 type of model it is that's being used. In addition, as
5 it relates to the data, is this a merged score or a
6 single source. So that these are all elements that come
7 in play.
8 Additionally you need to have reason codes.
9 This is really to educate the consumer. Again, I think
10 the discussion has been pretty full on the reason codes.
11 I won't go into it. And clearly you need to know the
12 date. Now, clearly the consumer knows when he applied,
13 but he needs to know the date when the credit granting
14 was done. We've already heard how dynamic the score is.
15 So you need to have a point in time to know when that
16 score was granted.
17 And just to comment on the list that Ray was
18 mentioning, these are a list of variables that
19 potentially may be used in any model. So that it is
20 information that would be present within the repository,
21 but it doesn't mean any particular model uses those
22 variables or that all models use any one. So that
23 knowing what the list is, is not going to tell you a

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
275.
275 Page 276 277
276
1 whole lot of anything.
2 And I think I will just end it there, because
3 there are more to speak.
4 MS. TWOHIG: Thanks very much, Forrest. And I
5 think that's a good segue to our first lender on the
6 panel. Dusty Lashbrook joined mortgage. doc, which is
7 based in Plantation, Florida, last fall as Executive Vice
8 President and Chief Operating Officer of its Consumer
9 Direct Division. Mortgage. com is a private label,
10 technology based, outsourcing company and a national
11 residential mortgage lender.
12 And Dusty is a nationally known speaker --I
13 know I've heard him speak before --on the subjects of
14 risk management and secondary market, and a guest author
15 for Mortgage Banking and Secondary Market Executive.
16 Dusty?
17 MR. LASHBROOK: Thank you, Peggy. It's a
18 pleasure to be here today. One of the good things about
19 going late in the day, besides the fact that the crowd
20 thins out a little bit, you do get a chance to listen to
21 what everybody has said so far and maybe alter what
22 you're going to say so you're not going to antagonize any
23 of the group out there.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
276.
276 Page 277 278
277
1 Anyway, to follow up and just give you a couple
2 more seconds of background on where my perspective is
3 coming from, because I think that's important, I'm going
4 to try to hit really what we think the consumer really
5 wants to see today. And this comes from a lot of
6 surveying that we as a company are doing.
7 We are primarily a consumer direct Internet
8 based lender. We are the outsource partner for Intuit on
9 the Quicken mortgage site, and we act as the fulfillment
10 side behind six of the lenders on the Quicken site, ad
11 well as serving as the lender on the site ourselves. In
12 addition, we have a substantial origination operation
13 through our own web site, mortgage. com.
14 From that, because on-line lending is so new --
15 the first on-line lending really started in about 1993 --
16 became our Consumer Direct part of our company. It was
17 American Finance and Investment based here in Washington,
18 DC, which we purchased. But it's so new that we're still
19 trying to learn what motivates the consumer on-line and
20 what the consumer wants to see.
21 The Internet gives you a great opportunity to
22 apply technology. It gives you a great opportunity to
23 change what is an extremely ugly process. Today what

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
277.
277 Page 278 279
278
1 you're seeing on the Internet is a lot of --a web facade
2 in front of the same ugly process out there. And what we
3 believe is that there is a great opportunity to change
4 the whole process, some of which is what we're talking
5 about here today. And to make it an experience for the
6 consumer that is truly different than what they
7 experience today. It's worse than a root canal to get a
8 mortgage.
9 But anyway, also we think the Internet gives you
10 an excellent opportunity to provide education, and that
11 is what the consumer wants. The consumer wants to know
12 more about this process. They want to know how to get a
13 mortgage loan. They want to know why they don't get the
14 same rate as someone else that applies. And so we feel
15 that there is a great opportunity to provide that
16 information over the Internet.
17 Through our surveying process --and we survey
18 every applicant at three points in the process through an
19 outbound survey and through a survey they complete at
20 closing. And besides learning about our service levels,
21 we're learning about what does a consumer want to see and
22 how should we change the process to make it a better
23 experience for them.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
278.
278 Page 279 280
279
1 This is what the consumer tells us this is what
2 they want to see. Number one, they want to be treated
3 fairly, and they want to be treated fairly in relation to
4 another applicant, and they want to be priced fairly on
5 that. It's not that they want to have the absolute best
6 price. They want to be priced fairly on that.
7 Number two, they want to understand the process.
8 They want to know what this mystic called Underwriting
9 is. You know, what is that process. What goes into it.
10 They want also to be able to --before they purchase or
11 go out to look for a home, they want to know exactly or
12 roughly what they can qualify for. So they want to be
13 really pre-qualified. They want to have the ability to
14 do that. They know that credit is going to come into
15 play in regard to that.
16 And today it's extremely important that they're
17 pre-approved. As I think Forrest could probably tell you
18 today, it's a very hot real estate market out there
19 today. And the ability for the consumer to be approved
20 ahead of time, and to know what they can buy, and to
21 basically go out there knowing that they can make an
22 offer on a property, as fast as properties are being
23 snatched up in some of the hot areas today, that's

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
279.
279 Page 280 281
280
1 important.
2 They also want to know, though, if in fact --
3 and they want to have a quick timely decision made so
4 they can move onto another lender, if they have to, or
5 they can actually go ahead and put behind them this
6 decision process and get ready for the move into the
7 property.
8 And the last thing is, if in fact they are not
9 approved, or if in fact they have to settle for a lesser
10 --I wouldn't call it a lesser quality --a higher risk
11 mortgage which they're going to pay a higher rate for
12 that, they want to understand how to change that
13 situation. How over time can I become a borrower that
14 gets as good as a rate as anyone else out there.
15 And that is what the consumer tells us that they
16 want. And I'll make one other note, because I think this
17 is important, too. We offer full credit spectrum
18 lending, all the way from A all the way down to D paper,
19 and we do it on a fully automated basis with six
20 underwriting engines employed. And the sub-prime engine
21 is one that actually does both credit analysis and also
22 risk based pricing.
23 We also don't allow our people to charge a

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
280.
280 Page 281 282
281
1 higher rate or a higher point structure and make overage
2 on that. So it's a very fair process that we put them
3 through, and the consumer likes that today. They love
4 being able to get an answer on a sub-prime loan within 30
5 seconds on that.
6 What is interesting is that in our site and the
7 Quicken site, we allow the consumer to self assess
8 themselves. And if you've looked at the site, or if you
9 go look at it, you can see they get to answer questions
10 about themselves, and it puts them down a channel to get
11 a quicker answer on their loan. Forty percent of the
12 people that self assess themselves as sub-prime and put
13 themselves into what we call a custom credit solution or
14 A paper borrowers.
15 Now, I know that's a surprising statistic, but
16 that's an accurate statistic today. So the consumer
17 doesn't know enough about what goes into making a credit
18 decision that puts them into a high risk category to be
19 able to properly assess themselves. And that's something
20 we're addressing right now.
21 Now, we in turn take those borrowers and move
22 them back over into the A paper channel, and they're the
23 happiest people in the world, I will tell you, because

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
281.
281 Page 282 283
282
1 they thought they're going to get a high interest rate
2 loan. Not every lender does that, as you all well know.
3 Many lenders look at that as a profit opportunity there.
4 What we think that should be provided to the
5 consumer --this is kind of a list. They need a full
6 explanation of what the mortgage process is all about.
7 They need a full explanation of credit scoring and how
8 credit scoring goes into the approval process. And that
9 approval process may be different for different lenders,
10 if you're using an automated engine, or if you're using
11 an automated engine, which we do, with also a live body
12 underwriters on the loans that are referred out of there.
13 They need to understand how credit scoring factors into
14 that. They need to understand what goes into the credit
15 score and how can I approve my credit score.
16 They also, we think, need the ability to have,
17 in our case, on-line credit counselling. So if you're
18 not able to --were not able to get a loan at a rate
19 today that makes sense for you, how do I get there. What
20 do I need to do. Do I need to work with a local
21 organization that specializes in this. How can I make my
22 credit better.
23 And we also think that as risk based pricing

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
282.
282 Page 283 284
283
1 comes on --and we obviously use risk based pricing in
2 what we call our custom credit solutions or the sub-prime
3 side. But they need to understand what goes into credit
4 grading and how those grades equate to what interest rate
5 I'm going to end up paying on that, and that it is based
6 on the risk out there associated with the loan, and what
7 are the factors that go into making a riskier loan.
8 So in our estimation, these are the things the
9 consumer needs. We feel that the lender is the one that
10 needs to provide that information on that. The lender is
11 the one that knows the whole process, or should know the
12 whole process, and should be the one that should be able
13 to communicate that to the consumer.
14 Now, there is obviously other partners in this.
15 Our real estate partners out here have a good opportunity
16 to help the consumer very early on in the process on
17 that. We are seeing more and more consumers, though,
18 that are actually --and maybe this is just due to the
19 adoption to the Internet --that are actually coming in
20 before they ever find a realtor and looking to pre-
21 qualify themselves. Then they go find the realtor. Then
22 they go find the property after that.
23 So we're seeing a big change that is occurring

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
283.
283 Page 284 285
284
1 out there today. But I will say in just kind of wrapping
2 up, the consumer wants to know more. The consumer wants
3 to be taken out of this mystic, taken off of the lending
4 process, so they can feel more comfortable with it and
5 know how they can control their own destiny.
6 MS. TWOHIG: Thank you, Dusty. Well, I think
7 we're pretty readily debunking --or pretty much
8 establishing that there is a black box still that
9 consumers --that everybody seems to be in agreement,
10 except maybe Peter Mahoney, that there is a need for a
11 lot more information from customers. And there seems to
12 be a consensus forming, among this panel in any event,
13 that lenders are the key communicator of what is going
14 on.
15 Dusty, you raised a very important issue, I
16 think, in terms of sub-prime borrowers or consumers who
17 think they might be --need to go to the sub-prime
18 market. If they weren't working with mortgage. com, they
19 very likely might call up, phone or otherwise contact a
20 sub-prime lender, just assuming that same thing, that
21 some credit problems mean they have to go to that. And I
22 suspect --in fact we've heard --that if that sub-prime
23 lender then pulls their credit score and it is very good,

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
284.
284 Page 285 286
285
1 that nevertheless the consumer won't likely be told about
2 that.
3 Is that your impression of what happens in --of
4 course, you know, not your company, but other companies?
5 And if so, what could be done about that?
6 MR. LASHBROOK: Well, first of all, I think that
7 it happens less, maybe, than you think, because the
8 consumer doesn't typically go to one person to look for a
9 mortgage loan. They're shopping, and it's a competitive
10 market out there today. And if I'm able to put a
11 borrower in a mortgage at seven percent, and someone else
12 is trying to sell them a nine and a half percent sub-
13 prime, then it's going to be very difficult for them to
14 get that business.
15 Now, I will say that we are seeing more and more
16 of a move on the part of the larger, more established
17 sub-prime lenders to where they are looking to take their
18 A paper borrowers, who come into them really for a sub-
19 prime loan that are definitely A paper, and handing them
20 off to an A paper lender. We're working with
21 relationships right now where we're doing that, where
22 we're taking the A paper business from some of the major
23 sub-prime lenders, and actually going ahead and

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
285.
285 Page 286 287
286
1 processing that into an A paper loan.
2 So I think there is a move a foot to try to get
3 away from what you're talking about. And I don't think
4 it occurs as much as you think. But my point really was
5 trying to say that the consumer doesn't know what they
6 are today. They need to have more information about how
7 do they fit in from a risk standpoint. A lot of them
8 think that if you're over 50 --you know, if you get a
9 late notice that that's looked at pretty derogatory as it
10 relates to a mortgage loan. Unless it's 30 days late, it
11 doesn't show up as a late on the credit report.
12 There are things like that that the consumer
13 needs to know more about.
14 MS. TWOHIG: Right. I guess I was asking, if
15 the consumer, though, was given its credit score and then
16 had some sense that that was a pretty good score, don't
17 you think that would help in the instances where it did
18 happen?
19 MR. LASHBROOK: Absolutely. And you would be
20 surprised at how many e-mail responses we get when people
21 come to our site where they know their credit score. I
22 don't know where they're getting it. They know their
23 score, and they are telling us what kind of rate will I

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
286.
286 Page 287 288
287
1 get if I --you know, I've got a 740 FICO score. And, I
2 mean, they know exactly what it is, and so they are
3 bargaining with that.
4 MS. TWOHIG: Very interesting. Moving to our
5 next panelist, Elizabeth Prentice. Beth is the New
6 York/ Puerto Rico District Director of the Neighborhood
7 Reinvestment Corporation. The District Office provides
8 technical assistance, training and grants to the 27
9 affiliated Neighbor Works Organizations in New York and
10 Puerto Rico. Beth is a founding member of the Neighbor
11 Works campaign for home ownership, and serves as a chair
12 of its National Loan Product Committee.
13 Beth?
14 MS. PRENTICE: Thank you. And also the chair of
15 the Predatory Lending Committee. So this is an issue
16 that is near and dear to my heart, and something that
17 we're very concerned about, obviously.
18 Just in terms of background, our organization,
19 the Neighborhood Reinvestment Corporation, comes at this
20 probably from the same perspective, to some extent, as
21 Marcia Griffin --are you still here --does, in that the
22 local organizations with which we work are working very
23 hard to educate generally first time home buyers and

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
287.
287 Page 288 289
288
1 prepare them for the mortgage buying process, and to the
2 extent possible, prepare them to move into the A market
3 rather than the B and C market.
4 We don't have, I don't think, as many borrowers
5 as mortgage. com does, who come in with a credit report in
6 hand knowing their credit score, nor is there typically
7 in the under served population a population that feels
8 all that comfortable shopping around for the best rates.
9 It's often a population that has not done well in terms
10 of being accepted by mainstream lenders, and so is not
11 that comfortable shopping around.
12 So our perspective comes from that of a group
13 that is trying to get fairly priced, regularly available,
14 mainstream capital back into under served parts of the
15 country.
16 That said, in answer to the two questions that
17 were put before us for this panel --what information
18 should consumers receive about credit scoring and who
19 should provide it --I think I --and probably Marcia and
20 other people who are on the direct lines working with the
21 under served population, first time home buyers --would
22 say that the more information and the earlier, the better
23 everybody is.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
288.
288 Page 289 290
289
1 To the extent that we're talking about credit
2 information, I do think there is a role very early in the
3 process for a mortgage counselor or a loan officer,
4 whoever the consumer is talking to first, to look at the
5 credit report with them. And I think they should have a
6 credit score as part of that early information.
7 Now, the question is, as someone said earlier,
8 which credit score? What I would propose for the early
9 part of the process, before someone has gone in to apply
10 for a loan, is that probably the most readily available
11 credit score to pull at that point and disclosure would
12 be the FICO score, because that is the most widely used
13 sort of generic score in this industry.
14 But, again, a score does not stand by itself and
15 needs to have someone explaining to the consumer what
16 part of the mortgage application or credit approval
17 process will be dependent upon this score and how they
18 fit within the scheme.
19 So in our world, we're very strongly encouraging
20 people to pull their credit report early, understand what
21 they're getting into, understand whether they would be
22 looked at as a good credit profile or a bad credit
23 profile, and work very closely to move them over --I

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
289.
289 Page 290 291
290
1 don't mean just to clean up your credit situation. You
2 know, you've taken the Kaplan test and now instead of
3 scoring your 580 on the SAT, you're going to score the
4 670, but your brain is no different.
5 We're looking to work with people so that in
6 fact they do understand their credit and budget and
7 financial behavior much more thoroughly, so that by the
8 time they get to a mortgage application, they are indeed
9 credit worthy and that credit worthy shows on their FICO
10 score.
11 Secondly, I would say that the other point at
12 which the credit score and credit information should be
13 disclosed is at the point of underwriting. Whatever
14 credit information is being used at that point in time, I
15 think it's important that the consumer get that one. So
16 let's say you came into a counselor to begin with. You
17 are using a FICO score, but you ended up going to maybe
18 Chase, which is using a proprietary mortgage scoring
19 process. I do think that if in fact you're going to be
20 turned down by Chase, you need to be given the credit
21 score and what that means within their system, along with
22 the reason codes.
23 So that's my answer to the two questions.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
290.
290 Page 291 292
291
1 MS. TWOHIG: Thank you, Beth. And what about if
2 you're not turned down, but you're just given a
3 counteroffer?
4 MS. PRENTICE? Yes.
5 MS. TWOHIG: But a higher price?
6 MS. PRENTICE: I think that's the other --thank
7 you for asking that. To me, being turned down and being
8 given a counteroffer are kind of the same thing in some
9 regards, in that most of us would hope to get the best
10 priced credit available. And if you're turned down for
11 that best priced credit and you're then given an
12 alternative, then I would yes, in that case these things
13 should also be disclosed to you. And, if you are being
14 given a different offer, you need to understand the
15 difference in the terms, because people don't.
16 And I think until people have more information
17 about this, they're not going to be able to scream in an
18 informed manner as they should if they're given an
19 unfair, quote/ unquote, risk based price. You know, if
20 you've gone from a seven percent, 30 year loan with two
21 points, and then because you didn't make the lender's 640
22 or whatever it is, you then are thrown into a seven and a
23 half percent, three points, you should understand that

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
291.
291 Page 292 293
292
1 and be free to shop around some more.
2 MS. TWOHIG: Thank you very much, Beth. And our
3 final panelist is Kirk Willison. Kirk is Senior Vice
4 President and Director of Government Affairs for
5 Countrywide Home Loans, which I'm sure everyone knows.
6 Countrywide --and since 1992 Kirk has been responsible
7 for corporate policy related to legislation and
8 regulation at the federal and state levels.
9 And for the past two years, Kirk is very well
10 known to many of us in this room for his efforts at
11 facilitating and directly and cajoling the mortgage
12 reform working group, which is an ad hoc organization of
13 industry and consumer groups that are --were --are
14 trying to reach consensus on modernizing laws related to
15 mortgage originations.
16 Kirk?
17 MR. WILLISON: Thank you. And thank everyone
18 for staying. Can you imagine the size of the straw that
19 I drew to be the last panelist?
20
(Laughter.)
21 MR. WILLISON: On the last panel of the day? So
22 you're to be commended for staying around. Listening to
23 the day's activities and the frustrations that I hear

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
292.
292 Page 293 294
293
1 from consumer groups and lenders and even some of the
2 credit reporting agencies, I'm reminded of that wonderful
3 line in the movie, Cool Hand Luke: what we have here is
4 a failure to communicate.
5 You look at a situation where, let's say, some
6 credit provides bad information to the credit reporting
7 agency. The credit bureau passes on that information to
8 the consumer's potential lender. And then the lender
9 turns around and likely denies credit on the basis of the
10 bad information to begin with. And if you're in the case
11 of someone like Countrywide, our agreement with Fair,
12 Isaac says we're not allowed to disclose to the consumer
13 what their score is.
14 Any of you who have lived in the Washington area
15 for a while will remember the ad slogan for a men's --
16 and I guess women's --clothing store, Sy Sims. He would
17 say that an educated consumer is our best customer. And
18 at Countrywide, we really do feel that way. And after
19 listening to a lot of discussion today, I've become
20 convinced that Congress was wrong, in that Congress
21 should not have said that consumers really aren't
22 entitled to their scores.
23 I think consumers need as much information as

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
293.
293 Page 294 295
294
1 they possibly can, particularly as we get now to the
2 slicing and dicing of risk based pricing. They need to
3 know how close might they be to that A loan versus a
4 premier loan or a A minus loan or a B loan. Maybe you're
5 never are going to see inside that black box and you can
6 take a look at this large list of all the characteristics
7 that make up a credit score. But it's not going to tell
8 you looking at that what the most important
9 characteristic on your experience --his credit
10 experience --has kicked you out of, perhaps, getting
11 that A loan.
12 We need to have an opportunity --Countrywide,
13 like I think probably a lot of different organizations,
14 provides a booklet, Credit Scoring and You, for consumers
15 as they come in and get the application. And any of you
16 who want a copy of that --it's written in very lay terms
17 and easily understandable --leave me your business card.
18 I'll make sure that you get sent a copy of that.
19 But we really do feel that we ought to at least
20 have the opportunity to do a better job explaining to
21 consumers what their score is, and then the steps that
22 they need to take. Countrywide also runs a counselling
23 center. And if we could get the opportunity to talk to

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
294.
294 Page 295 296
295
1 the consumers well beforehand, we could save them a lot
2 of frustration.
3 A few years back --and this is no affront at
4 the realtors here for us --the mortgage banking industry
5 --and Countrywide was pretty active in it --began a
6 campaign to see your lender first instead of seeing your
7 realtor first. I think I would abridge that to say see
8 your credit report first. Because you look at your
9 credit report first, see what dings you have in it, then
10 go in to either a counseling organization or your lender
11 and say, how can we work together to get me the absolute
12 best loan possible.
13 As a lender --and I think I want to make this
14 really clear to, I think, a lot of consumer groups who
15 may have the assumption that we're eager to turn down
16 loans. We're in the business of making loans. We're in
17 the business of evaluating credit, but we're in the
18 business of making loans.
19 Countrywide was very reluctant to be pushed in a
20 situation of having to rely on credit scores --FICO
21 scores --to the degree that we have been. And my CEO
22 made it very clear when we finally made the decision that
23 that would be part of the underwriting process that he

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
295.
295 Page 296 297
296
1 would dismiss any loan officer who used credit scoring
2 alone as a decision to turn down a loan. That's not
3 permitted at Countrywide.
4 And we've got to push everyone to realize that
5 there's got to be other factors, because if we get a
6 borderline score, we will do everything we can to push
7 with the GSE's to accept that loan. And given our size,
8 perhaps we're in a better situation than a couple others.
9 Finally, let me just wrap up here, because I
10 know that there has been some discussions about how sub-
11 prime lenders work. And not to make this a commercial,
12 but others have had a chance to get a few comments in.
13 Countrywide's sub-prime lender does move its --recommend
14 its borrowers up to our prime lending office if we
15 realize that the buyer is qualified. And we'll actually
16 do one better. If you have a Countrywide sub-prime loan
17 and you have a good credit payment history, we will,
18 after a certain period of time, automatically move that
19 up to a better interest rate.
20 So we are trying to do our steps to make sure
21 that consumers are getting the best deal possible.
22 Thanks.
23 MS. TWOHIG: Thank you very much, Kirk.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
296.
296 Page 297 298
297
1
(Applause.)

2 MS. TWOHIG: Kirk, I just want to make sure I
3 understood one thing. Is it the case, though, that you
4 don't give consumers their actual score when you explain
5 to them the reasons why they might be denied?
6 MR. WILLISON: Again, the agreement that we have
7 with Fair, Isaac, is that we are not to release that
8 score.
9 MS. TWOHIG: Okay.
10 MR. WILLISON: Our hands are tied.
11 MS. TWOHIG: Pete, do you want to say something
12 about that?
13 MR. MCCORKELL: Thanks. Actually this is a
14 question that you or David, I think, threw at Ray, and
15 Ray kind of threw back at you. You know, what would be
16 so terrible about giving consumers their scores? And my
17 answer to that is nothing. And I don't have a problem
18 with consumers getting their scores and the reason codes.
19 What I do have a problem with is then getting
20 into the discussion of well, my score is 680 but I need a
21 685. What can I do to change it? Because while Beth and
22 Marcia made it clear that their aim in their programs is
23 to try to change the credit behavior of their clients so

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
297.
297 Page 298 299
298
1 that in fact they are different credit risks,
2 unfortunately there are an awful lot of people when they
3 hear that score and they are some number of points away
4 from a cutoff or from a lower rate, they want to know
5 what can I do to change it.
6 And in fact, I'm sorry to tell you this, but
7 Fair, Isaac's job is not to tell you how to get a better
8 score. Our job is to produce a score that is the best
9 possible predictor of your credit performance. And so we
10 want to score your behavior. We don't want you altering
11 your behavior to change the score, what I think Beth
12 described as okay, I took the prep course for the SATs,
13 and so I don't know anything more now, but I'm just a
14 better test taker.
15 We don't want to have consumers trying to alter
16 their behavior in short term ways that will --regularly
17 has nothing to do with their long term credit risk. And
18 we don't want to get into that discussion with consumers.
19 And unfortunately, we have an awful lot of experience,
20 that when a consumer hears a score, especially when they
21 have it in the context of, but you needed five points
22 more or ten points more to qualify for this loan, then
23 they want to have that discussion about well, what can I

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
298.
298 Page 299 300
299
1 do this afternoon so that my score tomorrow will be
2 higher.
3 MS. TWOHIG: Thanks, Pete. Any reactions to
4 that from the panel? Joe? Identify yourself, please.
5 MR. GOLDBERG: Joe Goldberg from the
6 Pennsylvania Attorney General's office and the Bureau of
7 Consumer Protection. I have a very serious concern with
8 what you said, and I also have a very serious concern
9 with the credit bureaus. Nobody wants to take
10 responsibility for the information that you are
11 transferring.
12 What Fair, Isaac is doing and what the other
13 systems do, is making a statement about the character --
14 the financial character --of a consumer, and you are
15 conveying that to a third party. It's not purely
16 internal. And there is a reliance on that statement for
17 better or for worse.
18 What you do need to do is, if you're going to
19 take that step, I think you have the responsibility to
20 tell the consumer the basis for the statement about the
21 character, and the effect that changes in their behavior
22 will have on that statement. I disagree that you don't
23 tell them that. You are making a statement and conveying

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
299.
299 Page 300 301
300
1 that to a third party. You have an obligation to allow
2 this consumer to act. And by just giving a list of four
3 factors, if that consumer adds to the consumer's
4 detriment, in my opinion that's an unfair or a deceptive
5 trade practice.
6
(Applause.)
7 MS. TWOHIG: I think we've now officially opened
8 it up to questions and comments.
9
(Laughter.)
10 MS. TWOHIG: We do need to try to wrap up. So
11 anybody who hasn't spoken here today? No? All the hands
12 are of people who have already spoken. So let's take
13 them in turn. Caryl? And again, it's getting late in
14 the day, so if you could just try to keep your points
15 brief.
16 MS. ISEMAN: Well, this is again to what Peter
17 has said, and it's also to something that Virginia said,
18 that the credit scoring is altering the way people do
19 business. Virginia made a statement that she would not
20 allow or would suggest that a customer not go out and
21 consolidate their debts. No. It's better that we keep
22 them paying higher interest rates on their credit cards.
23 I mean, if they get an invite to accept eight percent on

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
300.
300 Page 301 302
301
1 their credit card --no. She did.
2 MS. FERGUSON: That's not what I said.
3 MS. ISEMAN: Well --
4 MS. FERGUSON: No.
5 MS. TWOHIG: Wait, one at a time. Caryl, why
6 don't you finish up your point, and then Ginny can
7 respond.
8 MS. ISEMAN: Well, what I'm saying is, is that
9 one of the counseling things that we do do when a
10 consumer comes to us is to take a look at their credit
11 cards and the balances that they have outstanding, and
12 find out maybe what the interest rates are, and do
13 suggest that they do go ahead and consolidate or to lower
14 their payments, or to close out some of the accounts.
15 And not only pay off the balances, but actually close the
16 account. And, yes, in many cases or in most cases that
17 will help their credit score.
18 I think that that is good principal. It's a
19 good ethical thing to do. It is a good economical thing
20 for the borrowers to do. And in no way should that --is
21 that cheating or should that be held against them.
22 MS. TWOHIG: Okay. Ginny, do you want to say
23 something:

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
301.
301 Page 302 303
302
1 MS. FERGUSON: Okay. I'm not arguing the point
2 What I said is, they should not arbitrarily go out and
3 consolidate debt without first having their overall
4 picture reviewed by an expert, so that they aren't doing
5 something to damage themselves inadvertently. I firmly
6 believe that people don't need 16 pages of credit cards.
7 I firmly believe that there are instances when certain
8 accounts need to be closed, but they need to do that with
9 the benefit of someone counselling them up front.
10 And unfortunately the media has put out there
11 that people need to close accounts and consolidate debt.
12 But they are doing that without the benefit of any
13 guidance, and that can in fact hurt someone.
14 MS. TWOHIG: Okay. Russ?
15 MR. BAILEY: Just a question for the lenders who
16 are engaged in risk based pricing. Could I ask if you
17 use the same scorecard to price the loan that you do to
18 underwrite it, and if you do not, does the pricing
19 scorecard include information other than the likelihood
20 of repayment like a pre-payment risk or the likelihood of
21 selling ancillary services?
22 MR. LASHBROOK: I'll be glad to address it from
23 our standpoint. On our risk based pricing, we are using

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
302.
302 Page 303 304
303
1 a combined engine that does both risk based pricing and
2 also the actual credit grading of the loan. The two are
3 tied together. So the same --you know, if you want to
4 look at it from a scorecard standpoint, the same
5 scorecard is actually used. It's the same engine that's
6 actually doing it.
7 And I agree with what Russ is saying. If you do
8 the two separate from each other, then you've got some
9 issues or got some potential for problems there as it
10 relates to maybe the effects that it might have on the
11 price, or you might not be getting the desired effect
12 overall. But we believe strongly in doing that.
13 Now I think the last part of your question was?
14 MR. BAILEY: Well, if it's the same one, you
15 probably don't have the problem that I was asking about.
16 But I think some institutions that use a separate pricing
17 model include factors such as the likelihood of
18 repayment, or the likelihood of selling the consumer
19 ancillary services. So what they call a risk based model
20 is actually a profitability based model, and it doesn't
21 really consider just the likelihood of the person being
22 able to repay the loan.
23 MR. LASHBROOK: Ours doesn't do that, but you're

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
303.
303 Page 304 305
304
1 absolutely right. There are people out there that in
2 their pricing formula --whether it's on an A paper also,
3 as opposed to it's also --I've seen it where those
4 factors actually do go in. But they do also contribute
5 to the overall value that the company obtains from that
6 loan, whether they're going to retain the servicing or
7 whether they're going to sell the servicing.
8 And all of us in the mortgage industry basically
9 set pricing off of a margin off of the ultimate value of
10 that loan as we perceived it. So those factors do go
11 into most companies' calculations of how they're going to
12 price a loan to the consumer.
13 MR. WILLISON: And that's exactly the same way
14 we are, and it's clearly the same in the A when you're
15 talking about the opportunity to sell, just like any
16 other bank product is done as it relates to the risk
17 based pricing. And mortgage. com would probably be pretty
18 much in the same situation. As a true mortgage bank, we
19 don't retain --we don't portfolio any of those loans.
20 So it also depends on what the market is like
21 for being able to sell those loans in the secondary
22 market, and we have to price off of that as well. So
23 it's not just a pure --I mean, you can't say someone who

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
304.
304 Page 305 306
305
1 comes in with a 619 is going to get a better, you know,
2 rate than someone who comes in at 620, if that's the only
3 factor that you're considering, because there are so many
4 other factors out there in the market. And it all
5 depends also on the type of --you know, you may be
6 trying to pull a certain type of loan, and those type of
7 things are certainly affected in a price.
8 MS. FERGUSON: Peggy?
9 MR. MR. BAILEY: I just --I think most
10 consumers when they hear risk based pricing think one
11 thing, when the person saying it may mean something
12 completely different. It's sometimes good to define the
13 term.
14 MS. TWOHIG: Okay. Ginny, you wanted to add
15 something?
16 MS. FERGUSON: Yeah. I think that there is a
17 little bit of a misconception. When people start talking
18 about risk based pricing, it doesn't always mean that
19 somebody is getting a worse interest rate. Today we see
20 many people in the high 700's that are getting lower
21 interest rates because their credit scores are good. And
22 I think a lot of people fail to take that into
23 consideration. And that's in the A paper arena.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
305.
305 Page 306 307
306
1 MS. TWOHIG: Okay. A few more questions or
2 comments.
3 MR. PAFENBERG: Peggy, could I make a comment?
4 MS. TWOHIG: Sure.
5 MR. PAFENBERG: Just to follow up on her
6 comment. Risk based pricing is actually present within
7 the system now, you know. I mean, you've just got to
8 consider that it's really there. But when people talk
9 about risk based pricing in the future, they're trying to
10 price basically four elements of risk: collateral risk,
11 credit risk, business risk and pre-payment risk. These
12 will all be elements of a future risk based pricing.
13 It's going to be a tough decision when you start
14 disclosing to consumers that I think you're going to
15 prepay your mortgage in two years, so that I'm going to
16 charge you a higher price. But that's the reality of
17 risk based pricing. The investor community is demanding
18 of the lender that this be priced. I mean, they're
19 demanding a premium.
20 MS. TWOHIG: We have questions here. The
21 microphone is closer to there. Why don't you go first,
22 sir, and then we'll take you two.
23 MR. PREUDHOMME: Hi. Thanks for the opportunity

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
306.
306 Page 307 308
307
1 to speak again. I'll try to keep it brief. I guess I
2 came here today with a lot of different expectations and
3 coming out of here with not too many answers.
4 But I see some big problems, one of which we're
5 sort of all agreeing on that the consumer needs to get
6 more information. I think that's a good thing. I think
7 the lady over there brought up the fact that a very big
8 problem we have is that the credit scoring systems have
9 not been evaluated for bias and fairness and so forth.
10 That needs to be done.
11 Something else --and I might be inaccurate on
12 this. But it seems like we almost have a monopoly. I
13 keep hearing Fair, Isaac, Fair, Isaac, and everybody is
14 like this to Fair, Isaac. And I think something is
15 terribly wrong if we've got a monopoly really going there
16 where the three main repositories go to them for their
17 scoring information. That's wrong. And I thought that
18 FTC was all about breaking up monopolies and everything
19 else.
20
(Laughter.)
21 MR. PREUDHOMME: You know, this is like --
22 MS. TWOHIG: Different bureau.
23 MR. PREUDHOMME: --a bad situation.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
307.
307 Page 308 309
308
1
(Laughter.)

2 MR. PREUDHOMME: And that to me is a big
3 problem. I mean, it really is.
4 MR. MEDINE: Mr. Preudhomme, I think we're only
5 against breaking up ill gotten monopolies, not monopolies
6 per se.
7 MR. PREUDHOMME: Well, I don't know. I mean, a
8 monopoly has a lot of inherent problems. If they don't
9 have to answer to anyone else, they don't have
10 competition. I think on your web page you talk about
11 competition, and there doesn't seem to be. And I hear a
12 great deal of arrogance from a lot of these type of
13 people.
14 But anyway, just let me hit five other topics
15 that are sort of related that I'm only going to --
16 they're just little one sentence things that I think need
17 to be thought about.
18 Okay. We talked about education. Another type
19 of education --just a word that need to get back to
20 Congress --is education in school doesn't occur. People
21 don't even know how to balance a checkbook. They don't
22 know what a loan to value ratio is. They don't know any
23 of that kind of stuff. That needs to be addressed. I

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
308.
308 Page 309 310
309
1 don't know that this group can do it.
2 There is a lot of privacy issues. Just
3 yesterday the Commissioner was speaking on H. R. 10 to
4 Congress about concerns they have about privacy, okay.
5 And I think there are a lot of privacy issues involved in
6 this data that gets so eagerly thrown around, shared,
7 scored, rated and gyrated.
8
(Laughter.)
9 MR. PREUDHOMME: I mean, that needs to be --you
10 know, somebody has got to look at that, okay. A real big
11 problem that I see that's going on related to what I
12 talked about earlier with Chevy Chase Bank. Okay. More
13 and more banks --and I heard Citibank came out and said
14 if you have with any company one late payment, we're
15 going to raise your interest rate because you're now a
16 risk. And then I heard GM has come out and said that.
17 And one bank after another after another. They're using
18 this risk based mentality.
19 And I think this is very scary, okay, that
20 they're going to --based on not your payment history or
21 what you're doing with them. Based on what they get on
22 this credit bureau report that we know is filled with
23 holes. It's bogus. It's not checked. I mean, this data

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
309.
309 Page 310 311
310
1 is garbage.
2 MS. TWOHIG: You promised one sentence.
3 MR. PREUDHOMME: Okay. Sorry. All right. The
4 targeting of youth is another big serious problem. There
5 have been articles, several of them --L. A. Times has
6 published some on that. That issue needs to be looked
7 at. Targeting of youth as well as elderly and the naive.
8 There is a lot of naivety out there as far as all this
9 stuff. Just like I said before, people don't understand
10 what it means to only pay your minimum payment.
11 Okay. There needs to be a national user limit.
12 Okay. We've got to have a ceiling somewhere. And I
13 don't know --
14 MS. TWOHIG: I think we're going beyond credit
15 scoring issues at this point.
16 MR. PREUDHOMME: Okay.
17 MS. TWOHIG: Why don't we give some other people
18 a chance to say a few words, and then we really do have
19 to wrap up.
20 MR. MANALP: I'll keep this very short. I
21 brought you up the FICO information scoring sheets.
22 You're more than welcome to pass out copies to everyone
23 in here that would like to see how the scores are

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
310.
310 Page 311 312
311
1 figured, if you don't know how already.
2 One of the things that you --yeah, those are
3 presented by FICO, right. That is their information that
4 they have passed out to me as the mortgage broker. I
5 think you need to keep in mind that you're talking about
6 the educational process that the mortgage lenders are
7 going to be responsible for to pass on to the consumer.
8 I think that if you really haven't had the
9 opportunity to be out there, you ought to go out one of
10 these days and go talk to a mortgage broker. Mortgage
11 brokers, I just found out this last month --it was in
12 one of the periodicals that we receive --are doing 75
13 percent of the business in the marketplace right now.
14 If you went out and talked to your average
15 mortgage broker or loan officer that's out there on the
16 front line, he was flipping burgers at Arby's about eight
17 or ten weeks ago. He does not have an education. You
18 are relying upon somebody with an I. Q. slightly hotter
19 than that hamburger that's been sitting on the counter
20 shelf out there to pass quality information on with
21 regards to his financial future.
22 So we've either got to start looking at the
23 industry itself and how to take care of ourself and get

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
311.
311 Page 312 313
312
1 some education to the people in our industry, so they
2 know this information. And let's get the black box
3 opened up a little bit to people that they can take it
4 and read it. And the black box that I look at is the
5 perpetual motion machine. We're told that it works, and
6 that it works on a regular basis. We just don't know
7 how, but we're supposed to believe in it. It's that
8 curtain that the wizard is behind.
9 Thank you.
10 MS. TWOHIG: Okay.
11 MR. WILLISON: Now, you know why Countrywide
12 gets go much business, right? They don't want to have to
13 deal with those flippers.
14 MS. TWOHIG: Thanks very much. One more comment
15 and then we're going to need to close.
16 MR. LE FEBVRE: I have a --is Pete still in the
17 room?
18 MALE SPEAKER: No.
19 MR. LE FEBVRE: Okay.
20 MS. TWOHIG: You drove him away.
21 MR. LE FEBVRE: Oh, no. Then I'll just make a
22 comment with regards to Fair, Isaac, and then I want to
23 address Ray that my association is.

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
312.
312 Page 313 314
313
1 When Congress amended the Fair Credit Reporting
2 Act, it made all repositories to come up with an 800
3 number that both --with a human being that answers the
4 phone. So my question is, as far as credit scoring is
5 concerned, since there is so much ambiguity out there,
6 why do we not force Fair, Isaac to handle 80 percent of
7 my phone calls that I can't answer and call forwarded
8 right to their office? Because right now, we can't
9 answer the questions.
10 Then my question to Ray is, when the FTC --and
11 this is kind of to David and Peggy at the same time.
12 When you guys made your ruling against Credco, you were
13 real clear that you do not desiferate (phonetic) --
14 whatever that word goes --between repository and credit
15 reporting companies. When I read the statute it says I
16 must, when a consumer disputes, to report that data and
17 get it accurate.
18 Now, the only repository, Ray, that is allowing
19 mortgage credit reporting companies to do that is
20 Equifax. Right now I have the ability to take that
21 verified data, have it then sent to Equifax within 48
22 hours, get the file corrected and recalculate the score.
23 Why can we not push Trans Union and Experian to

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
313.
313 Page 314 315
314
1 follow that same length, and this whole credit scoring
2 thing that we're all talking about here, will be then
3 corrected, at least to a point. So I want to kind of
4 hear your comments.
5 MR. CRESCENZO: A marketplace event. It means
6 each one of them makes their own decisions about how
7 they're going to address market needs. Equifax decided
8 that they would do a test, and they have and it's
9 working. How long has it taken anyone of the three
10 national systems to decide that a good idea across the
11 hall is a good idea to install in their environment.
12 You know, it's a migrating marketplace. If
13 somebody's got a good idea, it doesn't stay a secret
14 long. The marketplace is going to allow to work itself
15 out.
16 MS. TWOHIG: Thanks very much. It's getting
17 late. I know we've kept all of you way past the time we
18 had anticipated that we would close. But I think it's
19 just a testament to the interest in this subject, and
20 we've had lots of good presentations and comments.
21 And David Medine is going to close for us.
22 MR. MEDINE: Kirk thought he drew the short
23 straw, but I drew a shorter straw than Kirk, and so I'm

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
314.
314 Page 315 316
315
1 going to be extraordinarily brief.
2 I wanted to thank Kellie Cosgrove, Cynthia Lamb
3 and all the staff here for putting this together. It may
4 look seamless, but it wasn't.
5
(Applause.)
6 MR. MEDINE: I also want to thank all of today's
7 panelists for really engaging in a very good faith
8 dialogue about the subject. And I also want to thank
9 Pete especially for taking a lot of heat and being
10 willing to stand up and answer the questions. I don't
11 think it's an easy job. Of course we get that sometimes
12 ourselves, so we can sympathize with that.
13 I think that the theme that we heard from this
14 last session is, I think, really the theme for the day,
15 which is communication. And I think we started off the
16 day by hearing how credit scores were created. We moved
17 on to how they were used, what the benefits are. Then
18 raised some of the concerns about their use and their
19 fairness. And then wound up with how to best communicate
20 with consumers what the credit scoring process is all
21 about.
22 We hope today's discussion really is the
23 beginning of a dialogue among all the players here on how

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
315.
315 Page 316 317
316
1 we can make scoring fairer, easier to understand and
2 really to test the marketplace. If it works, if it gets
3 people better loans, if it's effective, that's great. If
4 there are problems with it and if it has kinks in it,
5 let's work on those. Let's work on better disclosures
6 and better understanding.
7 So thank you all for coming.
8
(Whereupon, the meeting was concluded at 5: 16
9 p. m.)
10
11
12
13
14
15
16
17
18
19
20
21
22
23

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
316.
316 Page 317 318
317
1
2
3
4
5
C E R T I F I C A T I O N O F R E P O R T E R
6
7 DOCKET/ FILE NUMBER: P994810
8 CASE TITLE: Credit Scoring Workshop
9 HEARING DATE: July 22, 1999
10
11 I HEREBY CERTIFY that the transcript contained
12 herein is a full and accurate transcript of the notes
13 taken by me at the hearing on the above cause before the
14 FEDERAL TRADE COMMISSION to the best of my knowledge and
15 belief.
16
17 DATED:
18
19
20 JOEL ROSENTHAL
21
22
C E R T I F I C A T I O N O F P R O O F R E A D E R
23

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
317.
317 Page 318
318
1 I HEREBY CERTIFY that I proofread the transcript for
2 accuracy in spelling, hyphenation, punctuation and
3 format.
4
5
6 JACKY KARNCHANAPHATI

For The Record, Inc. Waldorf, Maryland
(301) 870-8025
318.

Page Navigation Panel

1 2 3 4 5 6 7 8 9
10 11 12 13 14 15 16 17 18 19
20 21 22 23 24 25 26 27 28 29
30 31 32 33 34 35 36 37 38 39
40 41 42 43 44 45 46 47 48 49
50 51 52 53 54 55 56 57 58 59
60 61 62 63 64 65 66 67 68 69
70 71 72 73 74 75 76 77 78 79
80 81 82 83 84 85 86 87 88 89
90 91 92 93 94 95 96 97 98 99
100 101 102 103 104 105 106 107 108 109
110 111 112 113 114 115 116 117 118 119
120 121 122 123 124 125 126 127 128 129
130 131 132 133 134 135 136 137 138 139
140 141 142 143 144 145 146 147 148 149
150 151 152 153 154 155 156 157 158 159
160 161 162 163 164 165 166 167 168 169
170 171 172 173 174 175 176 177 178 179
180 181 182 183 184 185 186 187 188 189
190 191 192 193 194 195 196 197 198 199
200 201 202 203 204 205 206 207 208 209
210 211 212 213 214 215 216 217 218 219
220 221 222 223 224 225 226 227 228 229
230 231 232 233 234 235 236 237 238 239
240 241 242 243 244 245 246 247 248 249
250 251 252 253 254 255 256 257 258 259
260 261 262 263 264 265 266 267 268 269
270 271 272 273 274 275 276 277 278 279
280 281 282 283 284 285 286 287 288 289
290 291 292 293 294 295 296 297 298 299
300 301 302 303 304 305 306 307 308 309
310 311 312 313 314 315 316 317 318
email
contact
Discussions
creditnet
About.com
credit-repair
Credit Info Center
epinions
planetfeedback
Bayhouse
National Mortgage News
MSN MoneyCentral