2998 1 UNITED STATES OF AMERICA Before the 2 OFFICE OF THRIFT SUPERVISION DEPARTMENT OF THE TREASURY 3 In the Matter of: ) 4 ) UNITED SAVING ASSOCIATION OF ) 5 TEXAS, Houston, Texas, and ) ) 6 UNITED FINANCIAL GROUP, INC., ) Houston, Texas, a Savings ) 7 and Loan Holding Company ) ) OTS Order 8 MAXXAM, INC., Houston, Texas, ) No. AP 95-40 a Diversified Savings and ) Date: 9 Loan Holding Company ) Dec. 26, 1995 ) 10 FEDERATED DEVELOPMENT CO., ) a New York Business Trust, ) 11 ) CHARLES E. HURWITZ, ) 12 Institution-Affiliated Party ) and Present and Former Director ) 13 of United Savings Association ) of Texas, United Financial Group,) 14 and/or MAXXAM, Inc.; and ) ) 15 BARRY A. MUNITZ, JENARD M. GROSS,) ARTHUR S. BERNER, RONALD HUEBSCH,) 16 and MICHAEL CROW, Present and ) Former Directors and/or Officers ) 17 of United Savings Association of ) Texas, United Financial Group, ) 18 and/or MAXXAM, Inc., ) ) 19 Respondents. ) 20 TRIAL PROCEEDINGS FOR 10-15-97 21 22 2999 1 A-P-P-E-A-R-A-N-C-E-S 2 ON BEHALF OF THE AGENCY: 3 KENNETH J. GUIDO, Esquire Special Enforcement Counsel 4 BRUCE RINALDI, Esquire RICHARD STEARNS, Esquire 5 and BRYAN VEIS, Esquire of: Office of Thrift Supervision 6 Department of the Treasury 1700 G Street, N.W. 7 Washington, D.C. 20552 (202) 906-7395 8 ON BEHALF OF RESPONDENT MAXXAM, INC.: 9 FRANK J. EISENHART, Esquire 10 of: Dechert, Price & Rhoads 1500 K Street, N.W. 11 Washington, D.C. 20005-1208 (202) 626-3306 16 12 ON BEHALF OF RESPONDENT FEDERATED DEVELOPMENT CO. AND 13 CHARLES HURWITZ: 14 RICHARD P. KEETON, Esquire of: Mayor, Day, Caldwell & Keeton 15 1900 NationsBank Center, 700 Louisiana Houston, Texas 77002 16 (713) 225-7013 3 17 ON BEHALF OF RESPONDENT FEDERATED DEVELOPMENT CO., CHARLES HURWITZ, AND MAXXAM, INC.: 18 JACKS C. NICKENS, Esquire 19 of: Clements, O'Neill, Pierce & Nickens 1000 Louisiana Street, Suite 1800 20 Houston, Texas 77002 (713) 654-7608 21 22 3000 1 ON BEHALF OF JENARD M. GROSS: 2 PAUL BLANKENSTEIN, Esquire MARK A. PERRY, Esquire 3 of: Gibson, Dunn & Crutcher 1050 Connecticut Avenue, N.W. 4 Washington, D.C. 20036-5303 (202) 955-8500 5 ON BEHALF OF BERNER, CROW, MUNITZ AND HUEBSCH: 6 JOHN K. VILLA, Esquire 7 MARY CLARK, Esquire PAUL DUEFFERT, Esquire 8 of: Williams & Connolly 725 Twelfth Street, N.W. 9 Washington, D.C. 20005 (202) 434-5000 10 OTS COURT: 11 HONORABLE ARTHUR L. SHIPE 12 Administrative Law Judge Office of Financial Institutions Adjudication 13 1700 G Street, N.W., 6th Floor Washington, D.C. 20552 14 Jerry Langdon, Judge Shipe's Clerk 15 REPORTED BY: 16 Ms. Marcy Clark, CSR Ms. Shauna Foreman, CSR 17 18 19 20 21 22 3001 1 P-R-O-C-E-E-D-I-N-G-S 2 (9:00 a.m.) 3 THE COURT: Be seated, please. The 4 hearing will come to order. Mr. Guido, you have a 5 problem? 6 MR. GUIDO: No. Your Honor, we're 7 ready to call the next witness. 8 THE COURT: All right. 9 MR. GUIDO: I'd like to call 10 Dr. Darrell Duffie at this time, Your Honor. 11 12 DR. DARRELL DUFFIE, 13 14 was called as a witness and having been first duly 15 sworn, testified as follows: 16 17 THE COURT: Be seated, please. 18 THE WITNESS: Thank you. 19 MR. NICKENS: Your Honor, before we 20 begin Mr. Duffie's examination, Mr. Andrew Carron 21 who is our expert witness on mortgage-backed 22 securities is present in the courtroom. At some 3002 1 point, Mr. Rinaldi had invoked the Rule, but I 2 believe the Rule is universally interpreted not to 3 apply to expert witnesses. But I nevertheless 4 wanted for the record to point out that Mr. Carron 5 is here. 6 THE COURT: All right. That is the 7 customary practice. Do you take exception? 8 MR. GUIDO: We have no exception, Your 9 Honor. 10 11 EXAMINATION 12 13 Q. (BY MR. GUIDO) Would you please state 14 your full name for the record? 15 A. James Darrell Duffie. 16 Q. And what is your profession? 17 A. I am a professor of finance at Stanford 18 University. 19 Q. And how long have you been in that 20 position? 21 A. Since 1984. 22 Q. Can you describe for the Court your 3003 1 educational background post-high school years? 2 A. Certainly. I have a bachelor's degree 3 in engineering from a Canadian university, a 4 master's degree in economic statistics from an 5 Australian university, and a Ph.D. in engineering 6 economic systems from Stanford University. 7 Q. After you received your Ph.D., what did 8 you end up doing after that? 9 A. I went immediately to the finance 10 faculty at the Graduate School of Business at 11 Stanford University and have been on that faculty 12 since then. 13 Q. And what was your first position at the 14 Stanford Graduate School of Business? 15 A. Assistant professor. 16 Q. And for how long were you an assistant 17 professor? 18 A. Would have been about three and a half 19 years. 20 Q. Did you do any -- did you take any 21 sabbaticals from your position at Stanford 22 University as an assistant professor? 3004 1 A. No. 2 Q. Did you take any leaves? 3 A. Yes. I took a six-month leave at 4 Berkeley. 5 Q. And what did you do at Berkeley? 6 A. I did a post-doc in economics for a 7 six-month period funded by the National Science 8 Foundation. 9 Q. And who did you work with at Berkeley? 10 A. A number of prominent economists -- 11 Gerard Debreu, Kenneth Arrow, Andreu Mas-Colell, 12 perhaps 30 or so economists from around the world. 13 Q. And how long did you work on that 14 research project? 15 A. Six months. 16 Q. Six months? And was there a 17 publication that came out of that? 18 A. There were several. 19 Q. Okay. And what were they? Do you 20 recall? 21 A. They dealt with security pricing when 22 there are insufficient securities to hedge every 3005 1 possible contingency. 2 Q. And what did you do after you finished 3 your research at Berkeley? 4 A. I returned to Stanford to continue my 5 research and teaching there. 6 Q. And what position did you hold at 7 Stanford at that time? 8 A. As of the time that I -- immediately 9 after returning from Berkeley, I was still an 10 assistant professor. 11 Q. And did you at some point in time 12 receive a different title at Stanford? 13 A. Yes. I was reviewed and granted 14 promotion to associate professor in roughly 1987 15 or '88. 16 Q. And are you a tenured professor? 17 A. Yes. 18 Q. And are you still an associate 19 professor at Stanford? 20 A. No. I'm a full professor now. 21 Q. When were you made a full professor? 22 A. That would have been approximately 3006 1 1992. I can't give you the exact date. 2 Q. Now, can you describe the research that 3 you've done as a professor at Stanford with regard 4 to finance matters? 5 A. Yes. I've provided a resume that has a 6 list of my publications which deal with various 7 areas in finance, including the measurement of the 8 values of different types of securities including 9 interest rate securities, the management of risks 10 of those types of securities, the design of those 11 types of securities, corporate strategies for 12 managing the risks of those securities, and 13 several other areas in finance. 14 Q. Have you published any books 15 regarding -- as an outgrowth of the research that 16 you've conducted as a professor at Stanford? 17 A. Yes, I have. 18 Q. And what were those? 19 A. There are three books, one of which is 20 in its second edition. I'll just list them off. 21 The first one is titled "Security Markets: 22 Stochastic Models," published by Academic Press. 3007 1 The second one is titled "Futures Markets" 2 published by Prentice-Hall, and the third one 3 which is now in its second edition is titled 4 "Dynamic Asset Pricing Theory." 5 Q. Can you define for us briefly hopefully 6 what you mean by the word "stochastic models"? 7 A. Yes. That means subject to 8 uncertainty. So, models in which there are 9 uncertain changes in economic conditions. 10 Q. And what are the models designed to do? 11 A. Well, for example, the Black-Scholes 12 option pricing formula is designed to estimate the 13 market value of an option or to hedge the risks 14 associated with such an option. These models can 15 also be used to estimate the market value of bonds 16 or of derivatives on bonds such as options. They 17 can be used to measure the market value of 18 currency options or equity options. Other forms 19 of derivative securities -- you asked for a brief 20 answer so maybe I should stop right there. 21 Q. Now, the second publication you said 22 dealt with futures markets? 3008 1 A. That's correct. 2 Q. In what aspect of futures markets? 3 A. Both the valuation and hedging of 4 positions with futures, options, and other 5 derivative securities. 6 Q. And the third, it says "Dynamic Asset 7 Pricing Theory," I think is what you said? 8 A. That's correct. 9 Q. What does that refer to? 10 A. That refers to the valuation of 11 securities in a model where there are 12 possibilities for trading over time. So, 13 retrading as market conditions change. And the 14 book deals with the valuation and management of 15 risk of those types of securities. 16 Q. Now, have you published papers in 17 addition to the three books that you've just 18 mentioned? 19 A. Yes, I have. 20 Q. Can you briefly describe the subject 21 matters of those papers that -- in a general 22 sense, what is the subject matter of those papers? 3009 1 A. Again, they are in my general area of 2 research which is the valuation of financial 3 securities, the management of risks of those 4 securities, the design of securities, and related 5 matters. Also, corporate hedging strategies. 6 Q. Does that include valuation of 7 securities with options embedded in them? 8 A. Indeed, it does. 9 Q. Now, have you received any academic 10 honors? 11 A. Well, I've been granted research grants 12 by the National Science Foundation. I was awarded 13 the Batterymarch Fellowship which is awarded 14 annually to top researchers in finance. I was 15 made a fellow of the Econometrica Society which is 16 an international society of economists, and 17 various other research awards. 18 Q. Now, have you ever provided any 19 consulting advice to any entities with regard to 20 the management of portfolios that have financial 21 instruments in them? 22 A. Yes, I have and I continue to do so. 3010 1 Q. Can you tell us what you have done and 2 for whom? 3 A. Okay. Let's start with for whom. That 4 might remind me case by case what I've been 5 working on. I've been working with Wall Street 6 investment banks such as Morgan Stanley, Lehman 7 Brothers, with global banks such as Nikko 8 Securities of Japan, Royal Bank of Canada, Bank of 9 America, J.P. Morgan, Enron. That's not a bank. 10 That's a corporation here in Houston that deals 11 with trading in energy derivatives. Mobil 12 Corporation, Shell, Merck & Company, various other 13 corporations. 14 In the case of the corporations, the 15 work entailed risk management; that is, dealing 16 with risks associated with changing interest rates 17 or changing currency prices or changing energy 18 prices. 19 In the case of the financial 20 institutions that I mentioned, the work entailed 21 the design, valuation, and hedging of different 22 types of derivative securities and also risk 3011 1 management of the institutions; that is, 2 controlling the risks associated with changing 3 interest rates and principal for the banks and 4 also changing prices of other financial 5 instruments, as well as credit risk. 6 Q. Now, have you ever testified as an 7 expert witness in any trials? 8 A. Yes, I have. 9 Q. And which cases were those? 10 A. There are two cases that involve 11 trading in mortgage-related securities. One is 12 the Franklin Savings case. Franklin Savings is an 13 institution or was an institution in Kansas, and 14 there was litigation in that case involving 15 roughly a 10 billion-dollar portfolio and 16 mortgage-related securities. And the other case 17 is Far West Federal Savings which is an Oregon 18 thrift and again involved a large portfolio of 19 mortgage-backed securities. 20 Q. What was the issue involved in the 21 Franklin case? 22 A. Well, the government placed Franklin in 3012 1 conservatorship in light of its risks and Franklin 2 or at least its owners sued the government in 3 order to reestablish control of the institution 4 claiming that the government's decision to place 5 the institution in conservatorship was arbitrary 6 and capricious. 7 Q. And what was the question that you 8 addressed as an expert witness? 9 A. I addressed whether the risks, the 10 various risks associated with Franklin's 11 mortgage-backed securities portfolio were large, 12 whether they were well-managed, and whether they 13 were excessive in light of Franklin's capital. 14 Q. And in Far West, what was the issue in 15 that matter? 16 A. To be quite honest, I'm not familiar 17 with the overall litigation strategy associated or 18 the issues, the main issues associated with the 19 litigation. I testified on the specific issue of 20 the risks associated with their mortgage-backed 21 securities portfolio in light of the capital that 22 they had to sustain those risks. 3013 1 Q. And have you testified in any other 2 matters? 3 A. Yes, I have. 4 Q. And what was that? 5 A. The only other matter in which I've 6 testified in a trial is a Commonwealth Savings 7 institution matter that involved a Florida thrift. 8 This was of quite a different character. The 9 issue in that case was whether or not a certain 10 trading strategy constituted hedging; that is, 11 risk reduction or not. 12 Q. What was the outcome of the Franklin 13 case? 14 A. Well, the Court ruled in favor of the 15 plaintiffs; that is, that the government should -- 16 in common language, the government should not have 17 placed the institution in conservatorship. 18 Subsequently to that, there was an appeal and the 19 original decision was overturned. And subsequent 20 to that, there was an attempt to appeal again to 21 the Supreme Court. But my understanding -- and 22 I'm not familiar with the details -- is that the 3014 1 Supreme Court denied the opportunity to hear that 2 case. 3 Q. And what was the outcome in the Far 4 West case? 5 A. To be quite frank, I didn't actually 6 follow the outcome in that trial so I couldn't 7 tell you. 8 Q. Do you know whether or not there were 9 any findings with regard to your testimony in that 10 case? 11 A. Again, I couldn't tell you. I wasn't 12 told the outcome of any findings in relationship 13 to my testimony. 14 Q. And do you know the outcome in the 15 third case, the Commonwealth case? 16 A. I believe that the government's -- I 17 think it was the Resolution Trust Corporation or 18 the FDIC's case against Commonwealth was viewed as 19 meriting damages, but I couldn't tell you exactly 20 the -- what the outcome was. Again, I didn't 21 follow it. 22 Q. I'd like to offer into evidence as 3015 1 Exhibit A1116, Your Honor, Darrell -- Dr. Darrell 2 Duffie's resume, if I may. 3 MR. NICKENS: No objection. 4 THE COURT: Received. 5 Q. (BY MR. GUIDO) You were asked to 6 answer certain questions as part of your report in 7 this case, were you not? 8 A. Yes, I was. 9 MR. GUIDO: And, Your Honor, at this 10 time, I'd like to offer as A1112, Your Honor, the 11 report that was prepared by Dr. Duffie at the 12 request of the Office of Thrift Supervision. 13 MR. NICKENS: May I confer with counsel 14 for just a second, Your Honor? 15 THE COURT: Yes, you may. 16 17 (Discussion off the record.) 18 19 MR. NICKENS: We have no objection to 20 the report, Your Honor. It does differ in one 21 slight respect that I would note in that at the 22 top of this page, it indicates attorney work 3016 1 product on the documents produced to us, there is 2 no such designation. I will check it to see if 3 there are any other changes. But based upon that 4 reservation, we have no objection. 5 THE COURT: Received. 6 Q. (BY MR. GUIDO) Now, can you tell the 7 Court what questions you were asked to prepare an 8 expert report on? 9 A. Yes, Mr. Guido. I was asked to address 10 a list of questions. The first was what were the 11 types and what were the degrees of the risks 12 associated with the mortgage-backed securities 13 investments adopted by USAT. 14 The second question was whether or not 15 the assumptions that USAT used in its planning 16 documents for that MBS investment strategy were 17 sound. 18 The third question was whether there 19 was a likelihood that the goals stated for that 20 investment strategy, the mortgage-backed security 21 investment strategy, were likely to be attained. 22 The fourth question that I was asked to 3017 1 address was whether or not the accounting reports 2 that were provided by USAT fairly reflected the 3 actual economic performance of that 4 mortgage-backed securities portfolio. 5 And then the last question that I was 6 specifically asked to address was whether or not 7 the goals that were stated for this MBS investment 8 strategy were actually attained. 9 MR. GUIDO: I'd like to offer as 10 Exhibit A1115, Your Honor, the list of documents 11 that were reviewed and relied upon for the -- 12 Dr. Duffie's report. It's Exhibit No. 3 from his 13 deposition. 14 MR. NICKENS: No objection. 15 THE COURT: Received. 16 Q. (BY MR. GUIDO) I've handed you 17 Exhibit 1115, Dr. Duffie. Is that the list of 18 documents that you reviewed and relied upon in 19 preparation of your report? 20 A. Yes, it is. 21 Q. Can you explain to the Court what some 22 of those documents are starting with the FHLBB 3018 1 financial reports? 2 A. Yes. The FHLBB refers to the Federal 3 Home Loan Bank Board, and these financial reports 4 are reports that are provided regularly by thrift 5 institutions to the Federal Home Loan Bank Board 6 at the time covered by the dates indicated. And 7 the specific reports that are listed here are the 8 reports provided by USAT, semi-annual, quarterly, 9 and a thrift financial report for various dates 10 indicated. These show the assets and liabilities 11 by category, among other information. 12 Q. Now, the second makes a reference to 13 10Ks. Do you see that? 14 A. Yes, I do. 15 Q. Is that the 10Ks that were filed by 16 United Financial Group pursuant to the Securities 17 and Exchange Act as a publicly-traded company? 18 A. Yes. As with any publicly-traded 19 corporation, UFGI was required to provide the SEC 20 with annual reports, among other reports, showing 21 its financial status and other information about 22 the risks associated with investments in UFGI and 3019 1 that's the reports that I'm referring to here. 2 Q. Now, the next says "UFGI gap position 3 reports." What does that refer to? 4 A. "Gap" refers to the exposure of an 5 institution in terms of how much assets there are 6 at a given maturity relative to the liabilities at 7 that maturity. So, gap position reports would 8 indicate at various times how much exposure UFGI 9 had to changes in interest rates by maturity 10 bucket. Bucket is a term that's used a lot in the 11 finance industry. It refers to specific date 12 intervals. For example, from one year to two 13 years, from two years to five years, and so on. 14 Q. What are the USAT regulatory compliance 15 reports that are referred to in this list of 16 documents that you reviewed? 17 A. Yeah. Thrift institutions are required 18 to report to regulators whether or not they are 19 complying with minimum capital requirements, among 20 other regulatory requirements. And these indicate 21 the dates at which USAT's reports to regulators 22 were supplied. 3020 1 Q. Now, these are reports prepared by the 2 institution itself? 3 A. They are, indeed. 4 Q. Okay. Then the next one refers to 5 Sheshunoff reports on USAT. What are Sheshunoff 6 reports? 7 A. Well, I'm going to give you my best 8 sense of this. I understand that Sheshunoff is a 9 service that -- a commercially-available service 10 that provides information on financial 11 institutions regarding, for example, assets and 12 liabilities. And these -- this is two series of 13 reports in big thick binders that I received from 14 the OTC. 15 Q. Now, the next makes reference to Smith 16 Breeden Associates. Who is Smith Breeden 17 Associates? 18 A. Smith Breeden is and remains a 19 consulting and investment management firm. The 20 Breeden refers to professor Douglas Breeden of 21 Duke University. I haven't met Mr. Smith. And 22 this firm would provide during the mid-1980s and 3021 1 beyond that assistance or consulting services or 2 investment advisory services to institutions, 3 including thrift institutions and including 4 specifically USAT. 5 Q. Now, the next says "FHLB reports of 6 examination responses." What does that refer to? 7 A. Well, the Federal Home Loan Bank which 8 was the regulatory authority at this time 9 periodically conducted examinations. They 10 provided reports of those examinations and in some 11 cases, those reports requested responses by USAT, 12 for example, in light of variations from 13 regulations that USAT had. And USAT was required 14 to respond in writing, and these -- this is -- 15 these documents indicate a sequence of examination 16 and responses by USAT. 17 Q. And the next item refers to performance 18 reports. What were those performance reports? 19 A. Are you referring to the memos to 20 boards of directors? 21 Q. That's the category you have. There is 22 a question in case of who they went to, but you 3022 1 have them categorized as memos to boards of 2 directors. 3 A. Yes. 4 Q. What are those performance reports? 5 A. Those are reports indicating USAT's 6 estimates of the performance of its financial 7 portfolio at various dates, regularly provided to 8 the board of directors. 9 Q. Now -- but you don't know whether or 10 not they were actually provided to the board of 11 directors? 12 A. No, I don't. 13 Q. Now, with regard to the next, it talks 14 about asset liability minutes and memos. What 15 does that refer to? 16 A. Well, like many financial institutions, 17 USAT's executives had a committee in charge of 18 managing assets and liabilities. And these 19 documents refer to the minutes of that committee's 20 meetings at various dates from 1985 through 1988. 21 Q. Now, there is referred to as investment 22 committee minutes and a number of dates here. 3023 1 What does that refer to? 2 A. These are like minutes of committee 3 meetings. In this case, it's the investment 4 committee of USAT. And as you can see, there are 5 regular meetings and regular minutes of those 6 meetings that are available in the record. 7 Q. Now -- then look at various 8 miscellaneous documents that you have listed in 9 this report. And I'd like to direct your 10 attention to a couple of those for the purposes of 11 clarification. One of them is -- at the bottom of 12 Page 2 is an asset liability presentation using 13 the Sendero AL model. What does that refer to? 14 A. Sendero is a commercially-available 15 source of software that would provide estimates 16 of, among other things, market values of interest 17 rate sensitive instruments and risks associated 18 with those instruments. A/L stands for asset 19 liability model, and this document refers to the 20 results of the Sendero model for USAT's portfolio 21 of assets and liabilities. 22 Q. Is that at a particular date? 3024 1 A. Yes. And I see from the document here 2 that the date is April 22nd, 1988. 3 Q. Now, in the next page, there is 4 reference to a number of computer printouts and 5 various information from Smith Breeden Associates. 6 Do you see that? 7 A. Yes. 8 Q. What use did you make of the computer 9 data that Smith Breeden had generated on the USAT 10 portfolio? 11 A. Well, first of all, it gave me the 12 estimates provided by Smith Breeden on the 13 sensitivity of that portfolio to changes in 14 interest rates and prepayment rates. And 15 secondly, it provided me with information about 16 what USAT was learning from Smith Breeden about 17 its own portfolio. 18 Q. Now, did you -- when you reviewed the 19 investment committee minutes, did you find in the 20 investment committee minutes any estimation of the 21 market value of the various mortgage-backed 22 securities portfolios in the -- owned by USAT or 3025 1 any of its subsidiaries? 2 A. Yes. During 1987 and most of 1988, up 3 to October '88, there were regular reports 4 included in the investment committee minutes that 5 provided by category and then cumulative by type 6 of portfolio both the embedded losses in those 7 mortgage-backed security portfolios as well as how 8 those losses would grow or shrink with changes in 9 interest rates of a particular variety. 10 Would you like me to describe the 11 particular interest rates? 12 Q. Not at this point in time. 13 A. Okay. 14 Q. Now, when you developed your data for 15 your report, what data did you utilize for your 16 report? 17 A. I utilized both the results of these 18 investment committee minutes showing USAT's own 19 estimates of its losses and interest rate 20 sensitivity as well as some analysis that I did 21 myself of the institution's -- that is USAT's 22 alone -- entire interest rate risk sensitivity. 3026 1 That analysis that I did was based on the Federal 2 Home Loan Bank Board quarterly reports provided by 3 USAT to the Federal Home Loan Bank Board. 4 Q. I'd like to direct your attention to 5 the portion of your report that starts with the 6 stock price data that's in the report. That's 7 after the main body of the report dated 8 January 11th. 9 Now, taking the stock price data 10 information that you have there, where did you 11 obtain that information? 12 A. It was obtained from a vendor of 13 on-line financial data, trade line. 14 Q. And that is reflected in charts or 15 Table 1.1 and Charts 1.1.1 through .1.32. 16 A. As well as in other tables it shows up 17 as well, yes. 18 Q. Now, if you look at Table 2.1, it says 19 "USAT balance sheet summary." Where did you 20 obtain that information? 21 A. Well, this information was also 22 obtained by the Federal Home Loan Bank -- pardon 3027 1 me -- the Federal Home Loan Bank Board reports as 2 well as from the Sheshunoff reports. 3 Q. And then Table 2.2 is the UFGI balance 4 sheet summary. Where did you obtain that 5 information? 6 A. This is obtained from a series of 7 reports provided by -- or pardon me -- by UFGI to 8 the Securities and Exchange Commission. These are 9 commonly called 10Q. "Q" for quarterly reports. 10 They are in a similar series to the 10K reports 11 that are provided to the SEC by publicly-traded 12 corporations. 13 Q. Now, I'd like to turn your attention to 14 Table 2.1. It says "Summary of the SBA analysis 15 summer 1986 USAT mortgages." Tell us what that 16 refers to. 17 A. This is analysis -- this is a report 18 from an analysis done by Smith Breeden showing by 19 quarter the impact of -- among other information, 20 the impact of change in the term structure of 21 interest rates on the market value of the mortgage 22 portfolio analyzed by Smith Breeden. 3028 1 Q. Now, there are a number of tables, the 2 3.1, 3.2, 3.3. Were those information that was 3 obtained from Smith Breeden? 4 A. Yes. Up to Table 3.2 and including all 5 those charts, that's -- this is simply capturing 6 the data provided by Smith Breeden and putting it 7 into tables and charts. 8 Q. Now, the Table 4.1 says "SBA annualized 9 long-term prepayment rates." Do you see that? 10 A. Yes, I do. 11 Q. And where did you obtain that 12 information? 13 A. This was obtained from information 14 provided by Smith Breeden showing the estimated 15 prepayment rates on mortgage-backed securities for 16 various categories. 17 Q. And then Tables 4.21 and 4.22, the 18 treasury yield curves, where did you obtain that 19 information? 20 A. Well, I obtained treasury yield curve 21 information from two sources. One was Smith 22 Breeden and another was a publication provided by 3029 1 Coleman, Fisher & Ibbotson who provide estimates 2 of yields to maturity for U.S. treasuries. 3 Q. Now, look at Chart 4.31, 4.32. I guess 4 it's 4.31 through 4.35. 5 A. Yes. 6 Q. Where did you obtain that information? 7 A. The U.S. Treasury Department provides 8 in its periodic bulletins information on treasury 9 rates of various maturities, and this information 10 I obtained from those treasury bulletins one at a 11 time. 12 Q. Then you have a Table 4.1, mortgage 13 prepayment data. What does that refer to? 14 A. I'm sorry. Could you repeat the name 15 of the -- the number of the table? 16 Q. It's Table 4.4, mortgage prepayment 17 data. 18 A. Okay. There's two sources of data 19 here. One is a journal publication by Doug 20 Breeden, the same Breeden of Smith Breeden, 21 published in the Journal of Fixed Income showing 22 prepayment rates for various categories of 3030 1 mortgage-backed securities. And the second source 2 is a Goldman Sachs monthly bulletin that provides 3 the same kind of information by category. 4 Q. Now, what are the categories? At the 5 top, there is -- like in the middle, there is O to 6 49 and then it goes 50 to 99 minus 49 to zero. 7 Can you tell us what that means in 8 terms of the meaning of that chart? 9 A. Well, the easiest way to do it is by 10 example. Let's suppose that one was investing in 11 a portfolio of 8 percent coupon mortgage-backed 12 securities pooled by, let's say, Ginnie Mae. And 13 let's suppose that the prevailing market interest 14 rates were -- for that kind of mortgage were 15 instead of 8 percent, let's say 7 percent. So, 16 that would be a hundred basis points, 1 percent 17 difference, between the coupon rate on the 18 mortgages one holds and the current coupon rates 19 that are available to homeowners in the mortgage 20 market. Because the currently available mortgages 21 are at a lower interest rate, one would expect 22 homeowners would respond by refinancing or at 3031 1 least some would. And so, one might be concerned 2 about the extent of prepayments on one's 3 mortgage-backed security portfolio. So, one would 4 go, for example, to the column labeled 100 to 149 5 basis points and if it was in 1984, using these 6 data, one would estimate a 12.7 percent rate of 7 prepayment. And that's so-called mortality rate. 8 It's basically the percentage of the remaining 9 principal that would be paid down beyond scheduled 10 prepayments. So, this indicates the speed at 11 which homeowners would choose to refinance from 12 their current high coupon mortgages to lower 13 coupon mortgages in response to the lower rates. 14 Q. Now, does that mean that that was the 15 prepayment rate for that mortgage in 1984 or for 16 that -- in 1984 -- mortgage-backed security in the 17 year 1991? 18 A. Well, first of all, this is a generic 19 category. So, it refers to a large class and 20 there is averaging over this class. So, it varies 21 by agency. The Ginnie Mae or Government National 22 Mortgage Association, as you can see, has 3032 1 different prepayment rates than other federal 2 agencies. So, one would first have to know what 3 agency issued one's mortgage pool and, secondly, 4 one would want to have additional information on 5 the mortgage pool. But in terms of averages, this 6 gives the average prepayment as of the year 1984. 7 And I presume that this is averaged through that 8 year, but I don't -- I didn't actually review the 9 data collection methods that were used by these 10 sources. 11 Q. So, these charts purport to be charts 12 of what the prepayment rate was in the year 13 mentioned on the left-hand column? 14 A. That's correct. 15 Q. Now, Table 4.5 is the -- it says 16 "Prevailing mortgage rates and interest rates." 17 What was the source of that information? 18 A. Okay. There -- again, there are a 19 number of sources. For the Ginnie Mae data, I 20 referred to the same article by Doug Breeden in 21 the Journal of Fixed Income. Some of the 22 information was obtained by the annual 3033 1 statistical -- from the annual statistical digest 2 of the board of governors of the federal reserve 3 system commonly known as the Fed. And the -- some 4 of the interest rate data were obtained from the 5 same interest rate source that I mentioned 6 earlier: Coleman, Fisher & Ibbotson. 7 Q. Now, then you have Table 5.1 through 8 7.62, various descriptions of different aspects of 9 the USAT portfolio. 10 A. Yes. 11 Q. What was the source for that 12 information? 13 A. The source was information provided by 14 USAT to the Federal Home Loan Bank Board in its 15 quarterly FHLBB reports. 16 Q. And did that information include the 17 United MBS portfolio? 18 A. No. The United MBS was a subsidiary of 19 UFGI. So, its assets and liabilities would not 20 have been included in the reports provided to the 21 Federal Home Loan Bank Board. 22 Q. Now, then you have a section of your 3034 1 report starting about -- it's Table 7.10.1 through 2 7.11.1. Do you see those two charts? 3 A. Yes, I do. 4 Q. What was the source of that 5 information? 6 A. Okay. Well, in broad terms, there were 7 two sources: The reports provided by USAT to the 8 Federal Home Loan Bank in its quarterly FHLBB 9 reports as well as the internal reports provided 10 to USAT personnel in its investment committee 11 through the investment committee minutes. 12 Q. Now, with regard to the United MBS 13 portfolio sensitivity analysis, did you adopt the 14 figures that have been used by the institution in 15 doing their sensitivity analysis? 16 A. Well, I didn't have any independent 17 information on the United MBS portfolios' interest 18 rate sensitivity. There were no other sources of 19 data. So, I did use the information provided by 20 USAT, its own estimates. 21 Q. Now, can you define for the Court what 22 "sensitivity" refers to when you're discussing 3035 1 sensitivity here? 2 A. Yes. In this case, the report refers 3 to the sensitivity of the market value to a change 4 of interest rates of a certain variety, a 5 conventional move in interest rates which is a 6 change in interest rates at every maturity 7 increasing by 1 percent. So, for example, if the 8 one-year rate was at 4 percent and the ten-year 9 rate was at 5 percent, then this would measure the 10 change in market value of that portfolio as the 11 short rate moves from 4 percent to 5 percent and 12 as the long-term interest rate moves from 13 5 percent to 6 percent and, likewise, for all of 14 the other maturities along the term structure of 15 interest rates. In other words, a parallel shift 16 of the term structure of interest rates by 100 17 basis points. 18 Q. Now, take a look at Table 8.1. It says 19 "Interest rate sensitivity summary from investment 20 committee minutes. Source: United MBS summary of 21 investment committee minute books." 22 Is this the USAT investment committee 3036 1 minutes that you referred to earlier that you had 2 relied upon? 3 A. These are the data in those minutes, 4 yes. 5 Q. Now, looking at this report, it starts 6 with January 21st, 1987. Why was that date -- why 7 did you pick that date? 8 A. I simply took all of the dates at which 9 I had data provided by USAT. So, I won't presume 10 to say that they didn't measure it before that; 11 but at least in terms of the documentation that 12 they provided, this is the first date at which I 13 had any information. 14 Q. And this is a sensitivity analysis 15 solely for United MBS? 16 A. That's correct. 17 Q. Do you know what United MBS was when it 18 was created? 19 A. Well, it was a subsidiary of USAT that 20 was created to hold mortgage-backed securities and 21 related securities. As to the date on which the 22 sub was incorporated, I couldn't give it to you. 3037 1 Q. Does this chart show when it first -- 2 or the earliest date that the minutes reflect 3 something? 4 A. It must have existed as of January '87; 5 and I'm presuming that if they were keeping these 6 reports regularly, that it was created just prior 7 to this. But I don't know the exact date. 8 Q. Now, look at Table 9.1. It's referred 9 to as "Joe's portfolio. Interest rate sensitivity 10 summary. Source: Investment committee minutes." 11 What was the source of that 12 information? 13 A. This is, again, investment committee 14 minutes. That is, the minutes of the investment 15 committee of USAT regularly provided this 16 information from January '87 onward. 17 Q. What's your understanding that the 18 reference to Joe's portfolio as the -- 19 A. Well, "Joe's portfolio" seemed to mean 20 different things at different times. But in this 21 particular case, I'm assuming this included all or 22 essentially all of the mortgage-backed securities 3038 1 held by USAT not including, of course, United MBS. 2 Q. Now, you took all of this data; and you 3 did your own sensitivity analysis of the entire 4 institution? 5 A. That's correct. 6 Q. And what methodology did you use to do 7 that analysis? 8 A. Okay. I mentioned earlier that I had 9 two sources of data. One was the sensitivities 10 provided by USAT itself for the United MBS 11 portfolio as well as Joe's portfolio. And the 12 other source was the Federal Home Loan Bank Board 13 reports. So, what I did was to take the USAT 14 estimates for the portion of the portfolio that I 15 couldn't measure on my own, which was the United 16 MBS portfolio portion, and added to my own 17 estimates of the sensitivity of USAT's portfolio 18 to changes of interest rates combining the two to 19 give the total sensitivity of the two portions. 20 If you like, I can describe how I estimated the 21 sensitivity of the USAT portion. 22 Q. Did you -- let me just ask you this 3039 1 question. Did you use the same methodology that 2 was used to analyze the United MBS portfolio by 3 USAT itself when you did an analysis of the USAT 4 portion of the portfolio? 5 A. Yeah. Let me address that in two 6 parts. First of all, with regard to the United 7 MBS portfolio, I only had information on the risk; 8 that is, the sensitivity associated with these 9 parallel under 100 basis point shifts in the term 10 structure of interest rates. So, that was the 11 measure that I used for the other portfolio in 12 order to be able to add the two together. Of 13 course, there would be other risks that one might 14 be concerned about or one would be concerned 15 about. 16 I'm sorry. Could you remind me of the 17 question? I know I had a second aspect of this 18 that I wanted to address. 19 Q. Well, in doing it, you used essentially 20 the methodology that USAT used? 21 A. Oh, yes. The methodology. The 22 methodology is essentially to estimate the changes 3040 1 in market value when one moves the hundred basis 2 point parallel shift in the yield curve by what 3 would normally be called a non-option adjusted 4 method which means that one takes projected 5 prepayment rates at a hundred basis points shift 6 in the term structure of interest rates and assume 7 that those prepayment rates would actually apply 8 for the life of those mortgages. One would then 9 discount the cash flows to present value and 10 obtain an estimate of the market value at the 11 hundred basis point higher term structure of 12 interest rates. One would then take that market 13 value, subtract the market value at the original 14 term structure of interest rates; and the 15 difference would be called, in this case, the 16 sensitivity. 17 Q. You didn't do any analysis of the 18 impact on the net margin interest between the 19 liabilities and the assets, did you? 20 A. If you mean, by "margin," yield income 21 or -- 22 Q. Yield income, book income. 3041 1 A. No. 2 Q. Why not? 3 A. Well, two reasons. One is with regard 4 to United MBS. They measure changes in market 5 value, not changes in yield. And secondly, if 6 that's the two choices, I think they made the 7 right choice in that if one's concerned about the 8 interest rate risk of this portfolio, one wants to 9 measure how much market value it loses and not 10 simply what the change in the yield of the 11 portfolio is. That would only account for current 12 yield income and not the entire change in market 13 value which reflects all the cash flows to the 14 maturity of the instruments. 15 Q. Now, you used a hundred basis points as 16 the -- for your sensitivity analysis. Turn to 17 Chart 4.31 which is the chart dealing with 18 treasury notes and bonds yield to maturity and the 19 charts that follow that. 20 Do those charts tell us anything about 21 why you picked a hundred basis points as the 22 benchmark for the measurement as opposed to 200 3042 1 basis points or 300 basis points? 2 A. Well, looking at those charts, I think 3 it's pretty clear that one is concerned not only 4 about 100 basis point parallel shifts in the term 5 structure but also substantially larger changes in 6 the term structure. As one can see, rates did 7 move up and down by hundreds of basis points over 8 short time periods in the early 1980s. 9 So, one would want to look at larger 10 shifts; and, secondly, one would also want to look 11 at changes in interest rates which were not 12 necessarily the same at every maturity. For 13 example, one can see from the chart that 14 short-term rates would sometimes move much more 15 than long-term rates. Sometimes vice versa. 16 Long-term rates would change by more than 17 short-term rates. So, non-parallel shifts where 18 the yield curve becomes more steep -- that is, 19 longer rates, higher change than shorter rates or 20 more shallow -- and other types of changes in the 21 term structure are apparent from this chart. 22 So, in principle, one would want to 3043 1 look at large changes and non-parallel changes. 2 However, the only data that I had from USAT for 3 its United MBS portfolio was its own measure of 4 hundred basis point shifts in the term structure. 5 So, that's -- I restricted my analysis for 6 purposes of getting the entire picture to that 7 particular measure of interest rate sensitivity. 8 Q. So, you essentially applied their 9 methodology? 10 A. That's correct. 11 Q. And you say their methodology did not 12 take into consideration non-parallel shifts in the 13 yield curve? 14 A. Nor large shifts in the yield curve nor 15 other types of interest rate and prepayment and 16 other forms of risk. 17 Q. Now, with regard to non-parallel 18 shifts, can you describe what you mean by that 19 term for the Court? 20 A. Yeah. I think I may have fleetingly 21 referred to it in my last response, but let me try 22 to be a little more deliberate. Let's suppose 3044 1 that the short rate was at 7 percent and the long 2 rate was at 10 percent. A type of interest rate 3 shift that would be particularly nasty for this 4 kind of portfolio would be one in which the 5 long-term rate moved not just a hundred basis 6 points but, let's say, 300 basis points and the 7 short rate, instead of going up and parallel with, 8 it went down or went up by a smaller amount. That 9 would cause the yield curve to steepen and thereby 10 cause greater problems for a portfolio such as 11 USAT which had exposures to long-term interest 12 rates. 13 By the way, I should say that that 14 wasn't necessarily the case in all cases, that 15 they were exposed only to steepening of the yield 16 curves. In some cases, because of their swap 17 portfolio, they would have been exposed, as well, 18 to movements that tilted the opposite way to raise 19 short rates and lower long rates. 20 Q. Now, I'd like you to turn to the chart 21 that's on Page 21 of your expert report. Is that 22 chart essentially the conclusions that you reached 3045 1 in assessing the impact of an adverse hundred 2 basis point parallel shift in the yield curve to 3 the entire USAT portfolio all assets and all 4 liabilities and hedges and the United MBS 5 portfolio, all assets, liabilities, and hedges? 6 A. Yeah. With regard to that particular 7 measure of interest rate risk, this summarizes my 8 conclusions at least insofar as I was able to 9 obtain systematic conclusions from this time 10 period with regard to the loss that United would 11 record with a hundred basis point parallel shift 12 including losses in the swaps portfolio that were 13 already embedded in the portfolio but not yet 14 recorded. 15 Q. So that when you start with March of 16 1985, you have a projected loss of $40 million 17 with a hundred basis point adverse shift in the 18 yield curve? 19 A. That's correct. That's the estimate. 20 Q. And that's based on the sensitivity 21 analysis that you did utilizing the methodology 22 that USAT had relied upon? 3046 1 A. Yes, it is. And I want to emphasize 2 that this is just an estimate. The number could 3 be a little higher, a little lower. This is just 4 my best estimate available. 5 Q. Now, their methodology, was that what 6 is referred to as an OAS model that you used to do 7 this? 8 A. No. You would not refer to that as an 9 OAS model. 10 Q. What would you refer to that as? 11 A. Sometimes it's called a static 12 valuation model. Sometimes it's called a variable 13 prepay model. The idea is that with this type of 14 modeling, the type that was used by United and, 15 therefore, by me in obtaining these estimates, one 16 does not attempt to account for additional risks 17 associated with the fact that homeowners will 18 prepay adversely to the interest of USAT as 19 interest rates change. It assumes that 20 homeowners' prepayment behavior will not change 21 during the course of the maturing of these assets 22 as interest rates move. 3047 1 Q. So that if there is one prepayment rate 2 for that asset, it'll remain that prepayment rate 3 through all interest rate changes; or is there a 4 set of changes that are incorporated in the model? 5 A. There's simply -- when you initially 6 raise the hundred basis point -- raise the yield 7 curve by 100 basis points, one re-estimates what 8 the static prepayment rates will be for the life 9 of the mortgage portfolio and assumes that those 10 prepayment rates will be sustained for the life of 11 the portfolio. That's the methodology. Of 12 course, one wouldn't want to do that in practice. 13 Q. Now, is the OAS model something 14 different than this model in significant ways? 15 A. Yes. OAS, which refers to option 16 adjusted, will attempt to account for the fact 17 that as interest rates change during the course of 18 the maturing of these assets, homeowners will 19 prepay their mortgages to the disadvantage of the 20 owner of these assets. 21 So, for example, if interest rates 22 during one possible scenario were to come down, 3048 1 homeowners would recognize the opportunities to 2 stop paying high interest rates on their current 3 mortgages and refinance their homes at lower 4 interest rates. Taking advantage of that option 5 would be reflected in an option adjusted spread or 6 OAS model, but it would not be reflected in the 7 static model that's used for this purpose here. 8 Q. Now, what was the composition of the 9 risk-controlled arbitrage portfolio that USAT had 10 in place? 11 A. The portfolio was made up of -- largely 12 of fixed rate mortgage-backed securities and I can 13 describe those in more detail if you like on the 14 asset side. These investments in mortgage-backed 15 securities were funded by borrowing at short-term 16 interest rates using the mortgages themselves as 17 collateral for that borrowing. And the portfolio 18 included swap positions which are contracts, 19 derivative security contracts that exchange fixed 20 for floating interest rates as an attempt to hedge 21 against changes in interest rates. 22 Q. Now, what is your understanding of the 3049 1 composition of the United MBS portfolio that was 2 created sometime around January of 1987? 3 A. It was of a similar variety except that 4 it did not have substantial hedges of the sort 5 that I described a moment ago; that is, swaps to 6 offset rate interest rate sensitivity. It was 7 essentially exposed to changes in interest rates 8 in a way that the -- that the mortgage-backed 9 portfolio that I described earlier was not. 10 Q. What were the risks that the USAT and 11 the United MBS mortgage-backed security 12 risk-controlled arbitrage portfolios faced? 13 A. Okay. In my report, I broke them down 14 basically by category. The first general category 15 is interest rate risk and, as already was 16 illustrated, this portfolio was subject to losses 17 associated with parallel shifts of the term 18 structure of interest rates. 19 Secondly, I also mentioned in the 20 category of interest rate risk the fact that the 21 term structure of interest rates does not just 22 move in a parallel fashion. It also moves in 3050 1 non-parallel fashions. And that's an additional 2 source of risk on top of that. 3 Thirdly, there are changes in the 4 volatility of interest rates. From time to time, 5 interest rate markets become more volatile. 6 Changes in rates up and down become larger and 7 more rapid. And this type of volatility is 8 adverse to an owner of mortgage-backed securities 9 of the sort held by USAT. So, increase in 10 volatility would be an additional risk to this 11 portfolio, a very material one. Thirdly -- 12 Q. Now -- 13 A. Yes. 14 Q. Now, with regard to those, are those 15 addressed on Pages 6 through 10 of your report? 16 A. I'll just check the page numbers. 17 Referring specifically to interest rate risk, the 18 various types are addressed on Pages 6 and 7 and 19 then the other types of risks are addressed on 20 Pages 8 through the top of Page 12. 21 Q. Now, with regard to the interest rate 22 risk, you indicated that one of them was just a 3051 1 change in interest rates? 2 A. Of different varieties, yes. 3 Q. And one of those was a parallel shift 4 in the yield curve? 5 A. That's right. 6 Q. That is, the short-term interest rates 7 and long-term interest rates go up or go down the 8 same amount. And why was that a risk to the USAT 9 mortgage-backed security portfolio? 10 A. Well, because if one considers the 11 change in market value of the mortgages themselves 12 or the mortgage-backed securities, those changes 13 in values would not be fully offset by the changes 14 in values of the other instruments held in the 15 portfolio such as swaps. So, as interest rates 16 move, the assets and liabilities and hedges of 17 this portfolio are not moving in such a way as to 18 insulate this portfolio from changes in interest 19 rates of that variety as can be seen from the 20 numbers that I've provided on Page 21. 21 Q. Put the swaps aside for a minute. 22 A. Okay. 3052 1 Q. And just take the mortgage-backed 2 securities and the liability that was used to 3 finance them. 4 How would a parallel shift in the yield 5 curve impact USAT if interest rates rose? 6 A. Okay. Well, as interest rates go up, 7 the market value of any loan will go down. 8 Mortgages are just a type of loan. So, if I 9 promise to pay you, Mr. Guido, if you don't mind 10 my using you as an example, $100 million next year 11 when interest rates are at 5 percent meaning 12 roughly that I could invest 95 million now and pay 13 you 100 million next year, if interest rates go 14 from 5 percent to 10 percent, I have -- my 15 obligation to pay you is now less valuable to you. 16 Instead of owning $95 million worth of loan, you 17 own roughly $90 million worth of loan because it 18 only takes 90 million now to get back 100 million 19 in a year at 10 percent interest. 20 So, generally speaking, as interest 21 rates go up, the market values of loans such as 22 mortgages go down. This portfolio was 3053 1 substantially long-term loans of this variety. 2 Now, it is true that on the liability side, as 3 interest rates go up, the liabilities are moving 4 in the opposite direction but they are very, very 5 insensitive because they are all short-term 6 liabilities. So, the market values of those 7 liabilities are very insensitive to changes in 8 interest rates whereas on the asset side, 9 mortgages themselves were quite sensitive. So, 10 there was a mismatch there between the interest 11 rate sensitivity of the assets, which were very 12 exposed to general increases in interest rates, 13 and the liabilities which were relatively 14 unexposed to decreases in interest rates. 15 Q. Now, you mentioned volatility risk. 16 How does that come into play in valuing a 17 mortgage-backed security portfolio? 18 A. Okay. Volatility, as I mentioned a 19 moment ago, when it changes and interest rates 20 move up and down more rapidly or with higher 21 degrees of uncertainty, it becomes more and more 22 likely that homeowners will find it to their 3054 1 advantage to prepay their loans. Of course, if a 2 big interest rate move up happened, then they 3 would choose not to refinance. But if a big 4 interest rate move down happened, much more likely 5 to refinance their home mortgages. And simply 6 increasing the amplitude, that is the sizes of 7 these up and down moves, increases the likelihood 8 that rates will move down enough to cause them to 9 refinance to the disadvantage of the investor in 10 those mortgages. So, increasing volatility is bad 11 news for an investor of mortgage-backed 12 securities. 13 Q. Is that because the market anticipates 14 future moves in interest rates and if there is an 15 increase in volatility, the value of the bond will 16 go down in anticipation of accelerated 17 prepayments? 18 A. Yeah. The market wouldn't necessarily 19 assume that increased volatility means rates are 20 going down. But they know it's more likely that a 21 sufficient downward move would occur so as to 22 cause prepayments. And when the market sees that 3055 1 increased likelihood of prepayments, they will say 2 "these mortgages are worth less" and they will 3 write down the market value of these mortgages in 4 the market. They will be cheaper. 5 Q. Now, you also addressed liquidity risk. 6 What is the liquidity risk in a mortgage-backed 7 security? 8 A. Well, if one is investing one's capital 9 to hold the securities come what may, then 10 liquidity risk is not such a big issue. It means 11 that going in there is a transaction cost but 12 other than that, it's not a big deal. 13 If on the other hand, as was the case 14 here with USAT, one is borrowing money to invest 15 in these mortgage-backed securities portfolios, 16 one is taking the risk that the people lending 17 money will demand to be repaid when the mortgages 18 which are used to collateralize those loans are 19 not of sufficient value to repay them. 20 So, just to give you an example, 21 suppose I borrowed a hundred million dollars to 22 invest in mortgage-backed securities and let's 3056 1 suppose that I invested in a hundred million 2 dollars' worth of mortgage-backed securities. 3 Between today and tomorrow, those hundred million 4 dollars of mortgages may go down to, say, 5 98 million. The people that lent me money are 6 going to say, "I want my hundred million back." 7 And I look at my portfolio and it's only worth 8 98 million. The only way that I can pay them back 9 is to liquidate a large portion of that portfolio. 10 Now, if you're forced to liquidate a 11 mortgage portfolio quickly, you will typically 12 receive less for it because there is a 13 transactions cost associated with trying to sell 14 something quickly, sometimes called a fire sale, 15 or a need to immediately liquidate a portfolio 16 will cause those securities to be worth less in 17 the market than they would be if you were able to 18 calmly liquidate them over a long period. And 19 that illiquidity risk is a substantial risk when 20 the portfolio, as in USAT's case, was almost 21 entirely funded by borrowing. 22 Q. Now, are some mortgage-backed 3057 1 securities or their derivatives less liquid than 2 others? 3 A. Absolutely. 4 Q. And what sorts of instruments are less 5 liquid? 6 A. Well, the more, let's say, interest 7 rate risk sensitive they are, the more illiquid. 8 And the prime examples are IOs which refers to 9 interest only strips of mortgages, POs which 10 refers to principal only strips, certain types of 11 CMOs or collateralized mortgage obligations which 12 are derivatives of mortgages, residuals of CMOs, 13 and the like. 14 Q. Well, can you define for us what you 15 mean by "IO strips"? 16 A. Okay. Consider a pool of mortgages. 17 Now, as homeowners pay their mortgage payments, 18 they are paid in two portions. One of them pays 19 down some of the principal. Another portion pays 20 off some of the interest. Those two payments 21 would go to the pooler of the mortgages, but they 22 could also be split apart. One could invest in a 3058 1 security which receives only the principal 2 payments of the mortgage pool. That would be 3 called a PO. Or one could invest in a security 4 which receives only the interest portions of those 5 homeowners' mortgage payments, and that would be 6 called an IO or interest only strip. 7 Q. Did USAT hold any IO or PO strips? 8 A. Yes. Substantial numbers. 9 Q. Now, were those held in the United MBS 10 portfolio as opposed to the USAT portfolio? 11 A. I'm just going to refresh my memory to 12 make sure about that. 13 Q. It's Table 8.1 and 9.1. 14 A. Okay. Those were held in the United 15 MBS portfolio. 16 Q. Now, going back to the table on Page 17 21 -- I'm sorry. Before we do that, you mentioned 18 on Page -- I think it's Page 8 of your report, 8 19 and 9, prepayment risk. 20 What was the prepayment risk that 21 you're referring to in your expert report that 22 USAT encountered in this portfolio? 3059 1 A. Okay. Well, I'm going to make a 2 two-part answer again. The first and obvious 3 answer is one that I brought up earlier which is 4 that as interest rates go down, homeowners are 5 likely to -- more likely to prepay their 6 mortgages. And that's to the disadvantage of the 7 investor such as USAT because USAT's now going to 8 be receiving cash just when interest rates are 9 very low. That's bad news. 10 On the other hand, if interest rates go 11 up, then prepayment rates are going to slow down 12 and that means that USAT will be receiving its 13 cash later than expected just at a time when 14 interest rates are very high in the market denying 15 USAT the opportunity to take those cash flows and 16 invest them at high interest rates. They are 17 still only receiving low interest rates at farther 18 and farther up maturity as prepayments slow down. 19 So, either way, whether interest rates 20 go down or up, there is a risk associated with the 21 change in prepayment behavior of the homeowners 22 that prepay those mortgages. That was the first 3060 1 part of the answer. 2 The second part of the answer is that 3 prepayment risk is even worse than I just 4 suggested because it's very difficult to estimate 5 how that prepayment will change as interest rates 6 change. For example, if one knew precisely for 7 each type of interest rate change exactly how 8 prepayments would adjust, one could at least 9 attempt to try to offset those effects. And one 10 could also attempt to account for that when 11 deciding how much to pay for the security. 12 However, it's very difficult to know what 13 prepayment behavior will be. It changes from time 14 to time as the transactions technology changes as 15 there are, for example, lower and lower costs to 16 changing one's mortgages as homeowners become more 17 and more aware of the advantages of prepayment and 18 all of that sort of demographic social/cultural 19 change in the mortgage market is very difficult to 20 predict. So, that additional aspect of prepayment 21 risk is also very material. 22 Q. Okay. Now, with regard to the USAT 3061 1 portfolio, was that a portfolio that had swaps on 2 it as a hedge against interest rate changes? 3 A. Yes. 4 Q. And how did the prepayment risk 5 manifest itself in the USAT portfolio exclusively? 6 A. Okay. Well, the prepayment risk was in 7 the -- anything that was related to mortgages, 8 anything that had a prepayment option in it. 9 So, while the swaps had some offsets 10 associated with interest rate risk, they didn't 11 help at all with regard to prepayment risk. In 12 fact, they added to the types of risk associated 13 with prepayment risk. 14 Q. And why was that that they added to the 15 risk? 16 A. Okay. Do you want the short answer or 17 the long answer? 18 Q. The short answer and then we'll see 19 whether or not we need the long answer. 20 A. Okay. I mentioned earlier that as 21 interest rates go down, things get worse relative 22 to a non-prepayable bond and interest rates go up, 3062 1 they also are worse relative to a non-prepayable 2 bond. Now, the swaps that were held by USAT were 3 actually equivalent for economic purposes to short 4 positions in non-prepayable bonds. 5 So, if you take the advantages of 6 non-prepayable bonds that I described earlier and 7 then reverse that, take short positions, you're 8 actually adding to the types of risks associated 9 with mortgage-backed securities when you use that 10 as a hedge. And the terminology that's used on 11 the street to describe this effect is called 12 negative convexity. The mortgage-backed portfolio 13 had negative convexity. Negative means bad in 14 this case. And the swap position that they had 15 also had negative convexity. So, combining the 16 two created even greater negative convexity. 17 Q. Is that something that was known in 18 1984 or five? 19 A. Yes. 20 Q. Now, let's move over to Page 13 of your 21 report. You address in your report USAT's 22 investment risk management sophistication. 3063 1 What is it about the USAT's risk 2 management sophistication that you found 3 objectionable? 4 A. Okay. First of all, let me be clear. 5 I wasn't objecting to their own level of 6 sophistication in its own right. I was objecting 7 to the relative sophistication of USAT relative to 8 those with whom it was trading. And in this case, 9 the sophistication dealt with both the systems 10 that were available to USAT's trading partners 11 which were vastly superior to USAT's systems and 12 also, the information available to USAT's trading 13 partners which were the sophisticated 14 broker/dealers on Wall Street like Salomon and 15 Goldman, Merrill-Lynch, Morgan Stanley, First 16 Boston, and the like. 17 Q. What sort of information was available 18 to Wall Street, the people that USAT was dealing 19 with, that wasn't available to USAT? 20 A. Okay. Several types. First of all, at 21 the time in question, there were models available 22 on Wall Street called option adjusted spread 3064 1 models that would allow Wall Street broker/dealers 2 in mortgage-backed securities that were trading 3 with USAT to estimate the impact of this adverse 4 prepayment behavior by homeowners when they 5 estimated both the prices and the risks associated 6 with these mortgages. This technology was not 7 available to USAT. 8 The next part is that at a 9 broker/dealer -- at a major broker/dealer in 10 mortgage-backed securities on Wall Street, there 11 are desk after desk of traders that are closely 12 monitoring not only the mortgage-backed securities 13 markets but also related markets like government 14 bond markets, government option markets -- that is 15 bond option markets and the like. And this 16 information which is constantly available to these 17 traders allows the traders very up-to-date 18 knowledge of what the fair market value of a 19 mortgage-backed security is. 20 So, essentially, the Wall Street 21 broker/dealers had their pulse on the 22 mortgage-backed security market in a way that USAT 3065 1 with its limited information and limited personnel 2 did not. 3 Q. Well, these investment houses were 4 selling mortgage-backed securities and swaps 5 financed with reverse repos to USAT. Weren't they 6 providing that information to USAT? 7 A. Well, I think the saying goes caveat 8 emptor, meaning that anyone's whose buying from a 9 more sophisticated seller should be aware of that 10 relative sophistication. Now, if one wants to 11 invest in the securities, then one incurs that 12 cost once. But if one is going to trade 13 repeatedly in the securities, one is going to 14 incur those costs of trading against more 15 sophisticated counter-parties more frequently. 16 This is just on the trading side and then there is 17 the risk management aspect of it, as well. 18 Q. Now, you indicate that the Wall Street 19 firms had the OAS model available to them which 20 USAT didn't have as reflected in the methodology 21 that they utilized? 22 A. That's correct. 3066 1 Q. And when was the OAS model available to 2 Wall Street? 3 A. Well, I was asked that question in my 4 deposition and I said it was available at the time 5 in question and I was asked in the second round to 6 actually provide some information on why I believe 7 that. I had assumed it was common knowledge. 8 So, I actually went back and I found an 9 article that referred to its availability as of 10 1985 and I also contacted someone at a Wall Street 11 investment bank and asked for his recollection. 12 He was a player in this market -- for when it 13 became available and that also indicated that it 14 was available at the major players by 1985. 15 MR. NICKENS: Your Honor, I object to 16 that testimony. I asked him about that, about 17 what was said in that conversation. And he said 18 he couldn't remember or couldn't tell me, couldn't 19 provide the notes. Mr. Senft is not available and 20 I believe, based upon the information that we have 21 checked, that it's directly contrary to what this 22 witness has just said as to what he was told by 3067 1 Mr. Senft. So, I object to that hearsay testimony 2 and inaccurate testimony. 3 MR. GUIDO: Your Honor, I -- 4 MR. NICKENS: If we need to bring 5 Mr. Senft here as to what was said in that 6 conversation and whether or not this witness has 7 just truthfully related that, then we will do so. 8 MR. GUIDO: Your Honor, Mr. Carron is 9 in the room and listening to this testimony. I 10 think also spoke with Mr. Senft. I think that the 11 testimony will resolve that and, if need be, we 12 will call Mr. Senft, but I think there is plenty 13 of evidence that will be available to Your Honor 14 to ascertain the accuracy of this witness' 15 statements about his conversation. 16 THE COURT: All right. I'll deny the 17 objection. 18 Q. (BY MR. GUIDO) Now, what was the 19 article that you just referred to that led you to 20 conclude that the OAS models were available to the 21 Wall Street firms that were dealing with USAT in 22 trading mortgage-backed securities? 3068 1 A. This was an article -- I can't give you 2 the exact title. Its authors were a Mr. Waldman 3 and a Mr. Modzelewski that described the model 4 that was used by Salomon at the time. The article 5 was published in a book whose title I believe is 6 "The Handbook of Mortgage-backed Securities." In 7 fact, I think I see a copy of it over there on 8 that table or at least some edition of it. 9 Q. And is that -- 10 MR. NICKENS: Tell us where it is. 11 Q. (BY MR. GUIDO) -- edited by 12 Dr. Fabozzi? Are either of these the correct 13 edition of that book? 14 A. I can't say without looking, but I can 15 look if you like. 16 Q. Please. 17 A. Well, unless I've missed one of the 18 entries in this, I stand corrected. This is 19 probably not the same book, but it is a book 20 edited by Frank Fabozzi either called The Handbook 21 of Mortgage-backed Securities or The Handbook of 22 Fixed-Income Securities. And if necessary, I can 3069 1 obtain a copy of the article for the Court. 2 Q. Okay. Do you recall it was an article 3 that was in one of Frank Fabozzi's handbooks of 4 mortgage-backed securities? 5 A. Yes, dated 1985. 6 Q. Now, with regard to the OAS model, you 7 have a number of paragraphs on Page 13 through 14 8 describing what it is about an OAS model that is 9 superior to the static model that USAT had 10 available. And what factors go into the OAS model 11 that don't go into the static model that was 12 utilized by USAT? 13 A. Well, I'm not sure I'd use the word 14 "factors." It's more a question of methodology. 15 The OAS model will actually simulate what 16 scenarios will occur between now and the maturity 17 of the mortgages. Time by time, changes in 18 interest rates are simulated. Responses of those 19 interest rate changes to -- that is, prepayments 20 that respond to those changes in interest rates 21 are also simulated. And in each scenario from now 22 till maturity, one obtains the actual cash flows 3070 1 associated with those -- those prepayments by 2 homeowners. In that scenario, one knows what the 3 interest rate path that was taken was and one can 4 discount the cash flows to present value. That 5 all occurs scenario by scenario for hundreds, 6 perhaps thousands, of scenarios and then the 7 average effect of all of these scenarios will 8 provide a more accurate estimate of the market 9 value than simply an assumption in a static model 10 which says that prepayment rates will stay 11 constant or will change in a fashion which 12 doesn't -- is not affected by interest rates 13 through the life of the mortgages. The latter 14 model is inferior with regard to measuring the 15 market value of the portfolio. 16 Q. Now, is that because the OAS model 17 includes data that is not included in the static 18 model? 19 A. Yeah. I should be very clear that, of 20 course, the model is only as good as the data 21 available and the way that it's used. 22 So, another aspect of an OAS model is 3071 1 data by type of mortgage, by coupon rate 2 historically that shows how those prepayments are 3 sensitive to changes in interest rates so the 4 model itself is not of any use without good data 5 and the OAS models available would have had to 6 have access to historical data as well as 7 knowledgeable users of those models. 8 Q. Now, did the constant model that was 9 used by USAT, did that use averages of prepayment 10 rates? 11 A. Well, they used projected prepayment 12 rates as of that time. Now, where they obtained 13 those projections I couldn't really say. They 14 could have been based on industry averages. 15 Q. Does the OAS model use averages, or 16 does it use something other than averages? 17 A. It doesn't use average prepayment 18 rates, per se. It does use historical prepayment 19 rates to estimate what future prepayments would 20 occur for each interest rate scenario. So, 21 instead of taking an average to apply from now 22 till maturity, changes in prepayments are 3072 1 estimated to depend on changes in interest rates. 2 This is using statistical methods. 3 Q. Now, were the models that were used by 4 USAT to determine what hedges to put on, were 5 those comparable to an OAS model? 6 A. Well, I couldn't really say how the 7 model that was used by USAT, which was not an OAS 8 model, how it was used for purposes of choosing 9 hedges or whether, in fact, they used a model to 10 choose hedges. So, I mean, one could have 11 attempted to use their model to select hedges or 12 at least to get approximate hedges. But as to 13 whether they did or not, I couldn't really say. 14 THE COURT: We'll take a short recess. 15 16 (A break was taken at 10:32 a.m.) 17 18 THE COURT: Be seated, please. We'll 19 be back on the record. Mr. Guido, you may 20 continue. 21 MR. GUIDO: Thank you, Your Honor. 22 (10:52 a.m.) 3073 1 Q. (BY MR. GUIDO) Now, turning to Page 2 21 of your report, the table and the figure on the 3 mark to market loss to USAT with a hundred basis 4 point parallel shift in the yield curve, that 5 shows a total loss figure in the right-hand column 6 which you have graphed in Figure No. 3; is that 7 correct? 8 A. That's correct. 9 Q. Now, you included the swaps mark to 10 market in the calculation of the losses. That 11 actual market loss isn't really a sensitivity. 12 It's an embedded loss, is it not? 13 A. That's correct. 14 Q. Why did you include that in this mark 15 to market loss? 16 A. Well, I wanted to give an impression to 17 the extent that I had available data about the 18 total loss including embedded losses that would 19 occur with a hundred basis point shift in the term 20 structure of the yield curve. This would be one 21 signal of the degree of risk associated with the 22 portfolio relative to capital. 3074 1 Q. Now, why didn't you also include the 2 gains or losses for the MBSs for that same time 3 period? 4 A. Well, I had USAT's estimates of those 5 gains and losses but only for the period from 6 beginning in 1987 through late 1988, about 7 October 1988, and I didn't think it would be 8 appropriate to show those losses only for that 9 period and not for other periods. There were 10 substantial losses in addition. The other 11 information that I have to the extent that I know 12 is consistent through the whole period. It's 13 referring to the same number. 14 Q. Now, what conclusions do you draw from 15 this set of data when you show the results of a 16 hundred basis point parallel shift in the yield 17 curve? 18 A. Well, certainly, the losses that were 19 already there and would be added by a relatively 20 modest shift in the term structure of the yield 21 curve were very, very large relative to USAT's 22 capital and, therefore, indicated extreme danger. 3075 1 And this is just one of those sources of risk that 2 I mentioned earlier today. 3 Q. How significant is a hundred basis 4 point parallel shift in the yield curve? 5 A. Well, it's the type of shift that 6 occurs not infrequently. I can refer, for 7 example, to -- let's find -- if you'll just give 8 me a moment, I'll give you a chart that might 9 illustrate. 10 Q. Is that Chart 4.3.1? 11 A. Yeah. I'm looking at Chart 4.3.1 and 12 if you want to look just for a hundred basis point 13 moves, you could turn to the next page, Chart 14 4.3.2 which shows the magnitudes of changes in the 15 term structure of interest rates by date. And one 16 can see that a hundred basis point shifts, while 17 they don't happen all the time, are definitely 18 there and something that would be modest relative 19 to some of the shifts that actually have occurred, 20 particularly in the early 1980s. 21 Q. Now, did you assess the degree of risk 22 that this -- these projected mark to market losses 3076 1 would have for USAT? 2 A. I'm sorry. Could you repeat the 3 question? 4 Q. Did you attempt to ascertain what the 5 degree of risk that these projected losses would 6 have on USAT's operations? 7 A. Well, not in any numerical way but 8 certainly that was one of the key considerations 9 in my opinions about the extreme level of risk. 10 Q. Now, was it -- was your conclusion 11 about the extreme level of risk based on just 12 these absolute numbers or these numbers in 13 relationship to something else? 14 A. Oh, well, certainly, it was the -- the 15 question was how big were these risks relative to 16 the amount of capital that USAT had to sustain 17 them. I mean, for example, if this institution 18 had had significant amounts of capital, hundreds 19 and hundreds of millions of dollars, then it could 20 have withstood these kinds of losses indicated on 21 Page 21 of my report. However, they didn't have 22 lots and lots of capital. They had essentially no 3077 1 capital at all. 2 Q. And how do you come to that conclusion? 3 A. Well, we could turn, for example, to 4 Table 2.4. Now, the right-hand column of this 5 table gives the total stock market capitalization 6 of UFGI and shows that that capital, first of all, 7 started at a very small value, on the order of 8 $50 million in early 1984, and declined to 9 negligible eligible amounts in 1987. 10 Now, if one considers $50 million 11 relative to the total assets and liabilities of 12 the institution which were on the order of, say, 13 $5 billion or more, if you count United MBS, 7 14 billion if you count that, then this is absolutely 15 miniscule. It's a small fraction of 1 percent. 16 Q. And what were the-dollar risks that 17 were associated with that capital? 18 A. Well, among all the other risks that I 19 mentioned, if one just considers alone this 20 hundred basis point loss that's illustrated on 21 Page 21 of my report, one can see that it's 22 multiples of the total capital of the institution 3078 1 as measured by the stock market. So, no question 2 at all these -- there's essentially no capital and 3 plenty of risk. 4 Q. Now, do you have any figures that 5 reflect the book value of USAT at that time? 6 A. Yes, I do. For example, you could 7 refer to Table 2.1 which has the USAT balance 8 sheet summary or Table 2.2 which has the UFGI 9 balance sheet summary. These provide book data on 10 USAT's assets and liabilities. 11 Q. Okay. Take Table 2.1 -- 12 A. Okay. 13 Q. -- which is the book value. Can you 14 describe to us what the far right-hand column 15 refers to? 16 A. Okay. Well, first of all, I caution 17 these are only accounting numbers. But they show 18 that if one does not include goodwill, that is if 19 one takes the regulatory net worth and takes out 20 the goodwill, which is I think the appropriate 21 measure if you're stuck with book accounting 22 numbers, one sees that, indeed, there was no 3079 1 capital. I mean, essentially, it's zero or 2 significantly negative relative to the liabilities 3 of the institution. 4 Q. Now, if you look at the second column 5 which is regulatory net worth -- is that what it 6 is -- including goodwill? 7 A. The second column from the right is 8 regulatory net worth including goodwill, that's 9 right. 10 Q. Now, apparently, the chart doesn't have 11 the dates on it. My copy of the chart doesn't 12 have the dates. Does your copy of the chart have 13 the dates? 14 A. In the left-most column, I'm showing 15 the quarterly periods, March '84, June '84, 16 September '84, and so on. 17 Q. Well, take December 1986. 18 A. Okay. 19 Q. What was the regulatory net worth at 20 that point in time? 21 A. The book regulatory net worth counting 22 the goodwill was 249 million plus or minus cents. 3080 1 Q. And if you look at your chart on Page 2 21, what was the risk of loss with a 100 basis 3 point parallel shift that was projected there? 4 A. December '86, 234 million including the 5 losses embedded in the swap portfolio. 6 Q. Was that September or December '86? 7 A. You asked for December '86? 8 Q. Uh-huh. December 1986. 9 A. Yeah. 10 Q. 203 as opposed to 261 in terms of 11 projected loss? 12 A. This says "total loss"? 13 Q. Yes. 14 A. On Page 21? 15 Q. Yes. 16 A. I'm showing 234. 17 Q. Okay. And that there was 248 or 18 49 million in capital? 19 A. That's correct. 20 Q. What conclusions do you draw from that 21 in terms of the risk of loss that USAT incurred in 22 its portfolio at that point in time? 3081 1 A. Well, if you count the goodwill, which 2 I think is not the right thing to do because it 3 wasn't capital that would be actually there to 4 sustain that loss, it shows that the loss is on 5 the same order of magnitude as the total net worth 6 of the institution. And clearly, since a hundred 7 basis point shift in the term structure of 8 interest rates is modest relative to the types of 9 risk one would want to guard against when 10 protecting the entire existence of the 11 institution, that indicates an extremely dangerous 12 situation. And I want to keep adding emphasis 13 that this is only one of the types of risks that I 14 described earlier today. This does not include 15 all the other kinds of risks. 16 Q. And was one of the other kinds of risks 17 non-parallel shifts in the yield curve? 18 A. Non-parallel shifts, increases in 19 volatility, illiquidity problems, and prepayment 20 risks were the key ones in this regard. And this 21 is only measuring modest parallel shifts of the 22 term structure of interest rates and already we 3082 1 have a very large number relative to book capital. 2 Q. Now, you didn't do an analysis of the 3 mortgage-backed securities portfolios in 4 relationship to the capital over the 1984, 1988 5 time period, did you? 6 A. I looked at the whole institution 7 relative to capital. 8 Q. Now, Tables 8.1 and 9.1 are tables that 9 show the performance based on USAT's own figures 10 of the mortgage-backed security portfolios between 11 January '87 and October 1988. 12 What do those figures show you in terms 13 of the risk that each of those portfolios imposed 14 on USAT given its capital levels? 15 A. Well, as one can see from the 16 right-most columns of those tables -- let me just 17 give an example. If we go to, say, Table 8.1 and 18 just go to the bottom row, the very right-hand 19 bottom corner, one sees that including the losses 20 in the mortgage-backed securities in the United 21 MBS portfolio and the effects of IOs, POs, and 22 hedges, with a hundred basis point shift in the 3083 1 term structure, there is a loss of 162 million 2 from that portfolio alone. 3 And if one goes to the corresponding 4 Table 9.1, same date, bottom right-hand corner, 5 that's another example of an additional loss of 6 131 for a total of 294, roughly, by my addition. 7 Q. And how does that compare to the value 8 of the company represented by its market value? 9 A. Well, relative to market value at the 10 end of September '87 -- let me go to my market 11 value chart. 12 Q. It's Table 2.4. 13 A. Okay. So, Table 2.4. As of 14 September '87, the market value was 18 million. I 15 believe I just said that the losses at a hundred 16 basis points in those two portfolios including the 17 hedges was about 290 million. So, vastly larger 18 than the stock market capitalization which, I 19 would add, is larger than the effective economic 20 capital of the institution to protect itself 21 against adverse changes. And I can elaborate on 22 that if you want me to. 3084 1 Q. Now, compared to the regulatory net 2 worth, how significant is that risk? 3 A. Well, I guess it's a question of just 4 going to the right column and row. Relative to 5 regulatory net worth at September '87, 247 million 6 if you include the goodwill, which I again caution 7 is not a prudent thing to do here since it wasn't 8 actually available to cushion the institution 9 against losses. But if you do include it, 10 247 million was the book net worth whereas the 11 losses just in those two MBS portfolios at a 12 hundred basis point shift in the yield curve was 13 290 million, 50 million larger. 14 Q. Now, how significant did USAT view a 15 hundred basis point shift in interest rates? 16 A. Well, there is not extensive 17 documentation on what they viewed as large, but 18 there is a reference in a memo that was written by 19 Mike Crow -- I don't see a copy of it on my table. 20 But there is a memo by Mike Crow to the management 21 of USAT summarizing a presentation by Mike Giarla 22 of Smith Breeden. And in that presentation, the 3085 1 institution is shown to lose no matter whether 2 interest rates go up or down except for what they 3 describe as a modest shift of a hundred basis 4 points. 5 So, if one takes that terminology as I 6 do to be -- to be reasonable, that shows that a 7 hundred basis point shift is not a truly dramatic 8 shift. One should be able to protect an 9 institution against far more dramatic shifts than 10 a hundred basis points. 11 Q. So, even using the goodwill as capital, 12 they would not have been able to withstand that 13 shock? 14 A. Even if we went so far as to include 15 the goodwill as capital, that's right. 16 Q. I'd like to hand you what's been marked 17 as Exhibit A10649 which has previously been 18 admitted, Your Honor, and ask you: Is that the 19 memorandum that you just testified about? 20 A. Yes, it is. 21 Q. And where in that memorandum do you 22 find the reference to the hundred basis point 3086 1 risk? 2 A. On the second page -- I don't think 3 these pages are numbered except with Bate stamps. 4 But the second page in starting on the bottom line 5 is the sentence that I was referring to. And I'll 6 just quote it if you want me to. 7 Q. Please. 8 A. And I quote, "Similar to the overall 9 conclusions of Goldman Sachs' analysis, we appear 10 to lose in the overall structured arbitrage 11 portfolio in both increasing and decreasing 12 interest rate environments with the exception of a 13 modest decline in rates (a decline of 100 basis 14 points)." 15 Q. Now, you testified, I think, that USAT 16 had purchased interest only and principal only 17 strips. Do you recall that testimony? 18 A. Yes, I do. 19 Q. How did you ascertain that they had 20 purchased interest only or principal only strips? 21 A. Well, there are at least two sources 22 that I can recall at the moment. One is the 3087 1 summary of USAT's investment committee of the 2 sensitivity of its portfolio to changes in 3 interest rates, Table 8.1 that I referred to 4 earlier, where substantial embedded losses and 5 substantial interest rate risk on IOs and POs are 6 indicated. And secondly, there are internal 7 memoranda of USAT, for example, from Mr. Dominic 8 Bruno that indicate problems associated with the 9 illiquidity of IOs and POs in their portfolio. 10 MR. GUIDO: Your Honor, at this point 11 in time, I'd like to offer as Exhibit T6079 a 12 memorandum from Dominic Bruno to Mike Crow dated 13 May 17th, 1988. The subject, the MBS portfolio 14 objectives, direction, and strategy. 15 MR. NICKENS: No objection. 16 THE COURT: Received. 17 Q. (BY MR. GUIDO) Is that the -- one of 18 the documents, the internal documents that you're 19 referring to that you reviewed that made reference 20 to the IOs and POs? 21 A. Yes, it is. Would you like me to quote 22 the reference that I was referring to? 3088 1 Q. What page is that on? 2 A. It's the second page in, the first 3 sentence in the first full paragraph. And I'll 4 quote, "United's MBS portfolio is distinguished by 5 the fact that it possesses significant positions 6 in the riskiest MBS products created by Wall 7 Street over the past two years -- i.e., low coupon 8 IOs, high coupon POs (the same kinds of securities 9 Merrill-Lynch lost millions on several months 10 ago), and residuals off of floater type CMOs." 11 Q. Now, in that memorandum, Mr. Bruno 12 talks about low coupon IOs. Why are low coupon 13 IOs particularly risky? 14 A. Well, I'm not sure how I can 15 characterize it except to say that prepayment risk 16 for IOs is extremely severe because with an IO 17 security, one is, as you may recall from my 18 earlier description, receiving only the interest 19 payments. Now, in the event that a homeowner 20 decides to prepay a mortgage earlier than 21 suspected, one receives no further cash flow at 22 all as opposed to the situation in which one owns 3089 1 the entire mortgage portfolio, mortgage itself, 2 and at least in that scenario if the homeowner 3 prepays, one receives the principal at the time of 4 prepayment. But with IOs, you stop receiving any 5 money at all. And these IOs are noted to be 6 specific -- especially sensitive to prepayment 7 risk. 8 Q. And what about the POs that made them 9 particularly sensitive to interest rate risk? 10 A. Well, POs have essentially the same 11 problem, just moves in the opposite direction. If 12 a homeowner decides to delay prepayment, one 13 simply gets the cash flows much later. And this 14 makes POs particularly sensitive to prepayment 15 risk relative to the entire mortgage-backed 16 security which at least continues to pay interest 17 in the interim. The PO, you just simply get cash 18 flows that you were anticipating to get shortly at 19 a much later date when prepayments decline. 20 Q. Now, you in your report address what 21 you have characterized as the yield enhancement 22 trading of USAT as evidence that USAT was not 3090 1 operating the risk-controlled arbitrage in a 2 reasonable manner. 3 MR. NICKENS: Your Honor, I object to 4 leading. I don't have any objection to referring 5 to the specific words of the report. But to have 6 Mr. Guido characterize them and then ask the 7 question, I object. 8 MR. GUIDO: I'm sorry, Your Honor. 9 I'll rephrase the question. 10 Q. (BY MR. GUIDO) You address the 11 question of trading in your report, do you not? 12 A. Yes, I do. 13 Q. Okay. And is that on Page 19 of your 14 report? 15 A. Yes, it is. 16 Q. Okay. And what is it that you 17 ascertained from your review of the documents and 18 records USAT was doing that reflected trading in 19 your opinion? 20 A. Well, both the description of the 21 trading strategy as well as the trades that 22 occurred that were documented indicate that there 3091 1 was an attempt to identify securities that were 2 underpriced by the market and buy those or to 3 identify securities that were overpriced by the 4 market and sell those, thereby attempting to 5 speculatively capture trading profits by buying 6 and selling securities, guessing whether they were 7 over or underpriced. 8 Q. Now, what was wrong with that strategy 9 in USAT's risk-controlled arbitrage portfolio? 10 A. Well, the primary problem is that USAT 11 was not in a position to profit by speculative 12 trading in its mortgage-backed security portfolio. 13 It was trading at a disadvantage relative to its 14 trading partners; that is, the major 15 broker/dealers on Wall Street in mortgage-backed 16 securities. And by entering into this kind of 17 speculative trading program, USAT was essentially 18 betting that it could out-perform the remainder of 19 the market; that is, that it could guess better 20 than its trading partners what was over and 21 underpriced. So, this was simply adding risk and 22 trading with trading partners who were more 3092 1 sophisticated. 2 Q. Did you attempt to ascertain what the 3 impact of that trading was on USAT's portfolio? 4 A. I did not attempt to split out how much 5 of USAT's embedded losses in its mortgage 6 portfolio were associated with the -- this trading 7 strategy that I just described. I only have 8 information on the total losses. 9 Q. And did the -- if you look at Tables 10 8.1 and 9.1, did the losses that were incurred by 11 USAT and United MBS increase between January of 12 1987 and the last date of the report, September of 13 1988? 14 A. Yes. If in Table 8.1 one refers to the 15 second column from the right, one sees beginning 16 in 1987 at the start no losses and then over time, 17 the embedded losses growing quite dramatically to 18 over $100 million. 19 Q. Now, looking at the Table 8.1, can we 20 draw any conclusions about whether those 21 portfolios were properly hedged against parallel 22 shifts in the yield curve? 3093 1 A. Okay. Well, in terms of hedging, one 2 sees the interest rate sensitivity -- that is, the 3 losses associated with a parallel shift of the 4 yield curve -- which are -- well, there's a fair 5 bit of noise in here -- which are more or less 6 similar to the losses associated with the hundred 7 basis point decline in interest rates until 8 roughly the end of nineteen -- pardon me -- the 9 end of May 1987 when we start to see very 10 significant mismatch in that the losses associated 11 with a hundred basis point rise in interest rates 12 are large and substantial whereas as interest 13 rates decline, the portfolio gains in value 14 indicating that this portfolio is positioned to 15 profit for a decline in interest rates and to lose 16 in the case of a rise in interest rates. So, this 17 is what would be called a speculative stance, 18 basically a bet that interest rates would decline. 19 Q. Do you know what the size of the 20 mortgage-backed security portfolio was through 21 that time period? 22 A. It was on the order of billions, but I 3094 1 couldn't tell you exactly how much without looking 2 it up. 3 Q. Was it in the hundreds of millions? 4 A. I'm not going to blurt out an answer 5 without looking it up. I just make the risk of 6 just forgetting exactly the right number. But it 7 was certainly many, many hundreds of millions. I 8 just don't know -- 9 Q. At this point? 10 A. -- at this point. 11 Q. We'll wait for another witness to 12 provide us with that figure, Mr. Duffie. 13 A. Thank you. 14 Q. But this report reflects that there was 15 a portfolio that appeared not to be adequately 16 hedged against rising interest rates? 17 A. Oh, certainly not -- 18 MR. NICKENS: I object to the leading 19 nature of these questions. 20 THE COURT: Sustained. 21 MR. GUIDO: I'll rephrase. 22 A. Is there a question pending? 3095 1 Q. (BY MR. GUIDO) Does this report 2 reveal anything about the extent to which the 3 portfolio was hedged through that time period? 4 A. And I'll just summarize my previous 5 response, which is to say that it was not hedged 6 through the entire period and beginning in roughly 7 mid-1987. Not only was it not hedged but it was 8 positioned speculatively to gain with a decline in 9 interest rates and lose with a rise in interest 10 rates, at least according to USAT's own estimates. 11 Q. Now, you indicated in your report that 12 you reviewed the Joe Phillips deposition. Do you 13 recall that? 14 A. Yes. 15 Q. And did you see in the Joe Phillips 16 deposition that he made reference to a roll-down 17 strategy? 18 A. Yes. 19 Q. And what was your understanding that he 20 was referring to? 21 A. Well, the idea of a roll-down strategy, 22 putting aside the question of USAT's execution of 3096 1 it, would be as interest rates change to 2 restructure the portfolio so as to attempt to 3 maintain some balance between the assets and the 4 liabilities and hedges -- that is, to at least 5 partially insulate the portfolio from interest 6 rate risk. 7 So, "roll down" would refer to that 8 rebalancing as interest rates go down and some of 9 the high coupon mortgage-backed securities payers 10 decide to prepay their mortgages leaving less and 11 less mortgages on the asset side. One would 12 either want to rebalance the hedges or to replace 13 those high coupon mortgage-backed securities that 14 were paying down with new mortgage-backed 15 securities presumably of a lower coupon so that 16 they wouldn't pay-down so quickly at that point. 17 That was the general idea of a roll 18 down. Now, with regard to the execution at USAT, 19 it was a somewhat different matter. 20 Q. Now, were there other alternatives 21 besides rolling down the portfolio that were 22 available to Joe Phillips to adjust that portfolio 3097 1 in response to a decline in interest rates? 2 A. Yes. There were several other 3 alternatives that could have been taken. For 4 example, I mentioned a moment ago one could 5 readjust the hedges so in this case, as the high 6 coupon mortgages disappeared -- that is, they paid 7 down -- one could have taken the large swap 8 portfolio and cut it down in size as well in order 9 to maintain some degree of balance between the 10 two. 11 One could also have adjusted the 12 maturity structure of the swap portfolio so as, 13 again, to maintain some balance between the assets 14 and liabilities in each maturity bucket. 15 Q. Now -- 16 A. Among other -- among other strategies. 17 Q. How would he have adjusted the swaps to 18 make it conform to the pay-down mortgage-backed 19 securities? 20 A. Well, first of all, one wants to be 21 careful because, as I mentioned earlier today, 22 there is no way that a swap portfolio which has 3098 1 negative convexity is going to fully offset or 2 even largely offset the interest rate risk in a 3 mortgage-backed security portfolio. But one can 4 do the best one can do. For example, one can 5 adjust the swap portfolio so that at least with 6 respect to parallel shifts of the interest rate 7 curve, it would have the same sensitivity as the 8 mortgage portfolio. 9 In order to do that, Joe Phillips could 10 have done two things. First of all, as I 11 mentioned, cut down the size of the portfolio by 12 offsetting the swaps that they had, meaning 13 signing new agreements with the same 14 counter-parties to reduce the sizes of those swap 15 positions, or signing new swap agreements which 16 would move in the opposite direction, sometimes 17 called mirror swaps, and changing the maturity 18 structure of the swaps so that instead of all 19 being in the longer maturity, Joe Phillips could 20 have added swaps of shorter maturity to, in a 21 sense, leaven the mixture so that there would be 22 less interest rate risk. And those additional new 3099 1 swaps could have been newly created swaps or they 2 could have been arranged through agreements that 3 adjusted the pre-existing swaps. 4 Q. What would have been the transaction 5 costs of doing so? 6 A. Well, I'm not going to give you a 7 precise bid spread in that market. Basically the 8 cost would be the transaction cost for entering 9 into the swap agreements. One does not pay 10 initially to enter into a swap agreement, at least 11 a new swap agreement. So, essentially, it would 12 be the transaction cost associated with doing that 13 which would be -- well, let's suppose there were 14 just for round numbers, say, a billion in swaps. 15 Transactions costs nowadays are on the order of 16 three basis points or less, probably half that for 17 plain vanilla swaps like these. But in those 18 days, let's say it was much larger. Let's say it 19 was 10 basis points. That would be one tenth of 20 1 percent of a billion dollars which, let's see -- 21 1 percent would be 10 million. So one tenth of 22 that would be 1 million. So, let's say $1 million 3100 1 if they signed new swaps of the same magnitude 2 but, of course, they didn't have to do all of 3 that. They were only trying to reduce the size, 4 not eliminate it. 5 So, let's say something under a million 6 to reduce the swap positions they had and if they 7 wanted to layer on mirror swaps or different 8 maturity swaps, maybe at most another million to 9 do that. 10 So, on the order of under a couple of 11 million dollars. That's not counting, of course, 12 the fact that these swaps were already underwater 13 by tens and tens of millions of dollars. If they 14 wanted to at the same time recognize the losses in 15 those embedded swaps, of course they would have 16 had to pay for the privilege of wiping out those 17 embedded losses. And that would have been losses 18 that had already been occurred, but it would have 19 required cash payments to recognize them at that 20 date. 21 Q. I'd like to show you Exhibit T4310 22 which has been admitted into evidence. And it's a 3101 1 review of the MBS swap arbitrage activities as of 2 11-24-1986 as reflected in the memoranda from 3 Bruce Williams to Jenard Gross, Gerald Williams, 4 and Mike Crow. 5 MR. NICKENS: Your Honor, Mr. Guido has 6 mentioned this document to me. We have not been 7 able to find it be as being admitted. 188. I 8 apologize. 9 THE COURT: Would you give the number 10 again, Mr. Guido? 11 MR. GUIDO: Yes. T4310, Your Honor. 12 Q. (BY MR. GUIDO) Dr. Duffie, that 13 report summarizes the results of the USAT 14 mortgage-backed security risk-controlled arbitrage 15 portfolio at that date. And it shows that there 16 is an embedded loss of $122 million in the swaps 17 portfolio. Do you see that? 18 A. Yes, I do. 19 Q. And how much would your transaction 20 costs have added to that embedded cost if the 21 swaps had been readjusted in the way that you have 22 just described to better duration match the 3102 1 portfolio to a reduced size of the mortgage-backed 2 security portfolio because of prepayments? 3 MR. NICKENS: Your Honor, I object to 4 the form of the question because the document does 5 not refer to embedded losses. It merely refers to 6 mark to market losses. 7 MR. GUIDO: Okay. Mark to market 8 losses, Your Honor. I always thought that meant 9 embedded losses, but mark to market. 10 MR. NICKENS: The market can change. 11 And so, the implication of embedded is that that 12 loss cannot change. And so, I object to the use 13 of that term. 14 THE COURT: All right. Let's use mark 15 to market. 16 Q. (BY MR. GUIDO) Mark to market. 17 A. Okay. Well, as you indicated, the 18 document indicates a mark to market loss of 122 19 million. I mentioned in round numbers something 20 on the order of 2 million or less, probably less, 21 for the new swaps which would have been put in 22 place in order to adjust the swap portfolio so 3103 1 that at least with regard to parallel shifts of 2 interest rates, they would maintain a better 3 balance between assets and hedges. So, that would 4 have increased that mark to market loss, if one 5 counts the transactions cost, from 122 million to 6 let's say something on the order of 124 million. 7 Q. Does that memorandum indicate the 8 amount of money that USAT realized as a gain on 9 the sale of the mortgage-backed securities when 10 the roll down occurred? 11 A. Yeah. It indicates recognized gains of 12 67 million and as yet to be recognized gains of 13 12 million for a total of 79 million in gains on 14 the mortgage-backed side. 15 Q. And how large is the transaction cost 16 to readjust the portfolio by readjusting the swaps 17 in relationship to the gain that it had realized? 18 A. Well, again, using the same ballpark 19 numbers that I gave you a moment ago, let's say 20 something under 2 million relative to those gains 21 of 79 million. So one-fortieth order of 22 magnitude. 3104 1 Q. Okay. Now, were there other methods 2 other than the methods that were adopted by Joe 3 Phillips to readjust the portfolio after the 4 decline in interest rates? 5 A. Yes. One could have -- I mean, 6 actually, my first reaction would be in this 7 scenario to liquidate the portfolio and take the 8 losses at that moment. Given the minimal or 9 non-existent capital of the institution, I think 10 that might have been the most prudent course to 11 recognize those losses at that time. But if one 12 insists on maintaining the mortgage-backed 13 securities in addition to adjusting the swap 14 portfolio, one could, for example, buy certain 15 options which are designed to take that negative 16 convexity and eliminate some or all of it. 17 Q. Now, why do you say that the most 18 prudent might have been to liquidate the portfolio 19 and then take the losses? 20 A. Well, even with the best possible 21 hedging, as I mentioned earlier, it's very 22 difficult to know exactly what's going to happen 3105 1 with mortgage-backed securities portfolios. 2 Prepayment behavior changes unexpectedly. And so, 3 no matter what they could do in terms of hedging, 4 so long as they had those mortgage backs on their 5 books, they were going to be subject to additional 6 risks, non-trivial risks. And given the amount of 7 capital that they had, I think it was very 8 imprudent not to recognize the losses and get out 9 of a situation where there is a risk of additional 10 losses of capital. 11 Q. What had the Smith Breeden people told 12 them would happen to that portfolio if there were 13 changes -- additional changes in interest rates? 14 A. In fact, I think I quoted earlier today 15 from a memo by Mike Crow of USAT summarizing the 16 Smith Breeden presentation and indicating that 17 except for a modest decline of a hundred basis 18 points in interest rates, that portfolio would 19 lose no matter which direction interest rates 20 moved, whether up or down. 21 Q. How does the loss figure in 22 Exhibit T4310 compare to the loss figure in your 3106 1 Table 8.1 -- I mean 9.1 at October 1988? 2 A. As of October 1988, the embedded losses 3 in the mortgage-backed security portfolio are 4 shown as on the order of $100 million. 5 Q. Okay. 6 A. These are USAT's numbers. 7 Q. Okay. Now, with regard to the 8 roll-down itself, was the roll-down performed in a 9 manner in which an attempt was made to maintain 10 the spread income? 11 A. Well, let me first of all say that I 12 don't think there would have been any strategy 13 which would -- which could have been assured of 14 maintaining the spread income but there could have 15 been strategies that mitigated the loss of income 16 which would have been better than the ones that 17 were actually conducted. 18 For example, putting aside the issue of 19 the swaps and what would one do to try to maintain 20 a better balance there, one would want to replace 21 those high coupon mortgage-backed securities with 22 lower coupon mortgage-backed securities in a more 3107 1 timely manner. That is, rather than waiting for 2 interest rates to come way down and suffering 3 substantial losses on the swap side while holding 4 a low amount of mortgage-backed securities, one 5 could have replaced those mortgage-backed 6 securities more -- in a more timely manner so as 7 to maintain a better balance on the asset side. 8 Q. Were the profits, to your knowledge, 9 reinvested in mortgage-backed securities in an 10 attempt to maintain as much of the net income 11 spread as possible? 12 A. Well, it indicates that the profits 13 were recognized. I can't tell you in terms of 14 dollars how much of those profits were actually 15 plowed back into mortgage-backed securities. 16 Q. Were all of the -- were all of the 17 proceeds plowed back into mortgage-backed 18 securities? 19 A. No. 20 Q. Now, was the roll-down initiated at a 21 time that interest rates indicated -- the decline 22 in interest rates indicated that prepayments were 3108 1 going to accelerate? 2 A. Well, again, there is an issue of 3 timeliness here. The short answer is no. 4 Interest rates came down significantly and only at 5 that point, after they had already come down, was 6 the roll-down strategy pursued. 7 Q. So that when the roll-down strategy was 8 initiated, the losses had already been incurred in 9 a mark to market basis in the portfolio? 10 MR. NICKENS: Your Honor, I object to 11 this continuing leading of this witness. He's 12 perfectly capable -- in fact, in many instances he 13 doesn't require a question. So, we shouldn't have 14 this leading to clutter up the record. 15 MR. GUIDO: I'm sorry, Your Honor. 16 Q. (BY MR. GUIDO) Were there -- you 17 testified that the roll-down was a question of 18 timeliness. What was it about the roll-down that 19 raised in your mind a question of timeliness? 20 A. Well, the term structure of interest 21 rates showed dramatic declines during early 1986 22 on the order of more than 200 basis points or so 3109 1 before significant replacement of the 2 mortgage-backed securities occurred. And that -- 3 I mean, basically to put it in colloquial terms, 4 the horses were already out of the barn. There 5 were too few mortgage-backed securities sitting 6 against that huge swap portfolio by the time that 7 a significant replacement of those paid down 8 mortgages occurred. 9 Q. Now, what would have been the 10 appropriate time to consider rebalancing in a 11 declining interest rate environment given that 12 portfolio? 13 A. Well, if one was doing the job in a 14 timely fashion, one would every day, if not 15 intraday, monitor the prepayments, replace the 16 mortgages that were prepaid with new mortgages -- 17 again, if one is insisting on this strategy of 18 maintaining the mortgage backs and the swaps. One 19 would continue to maintain a balance in that 20 portfolio always replacing the mortgages as paid 21 down with new mortgages or new mortgage-backed 22 securities. 3110 1 Q. So, by -- I mean, were there any sort 2 of guideposts in terms of changes in interest 3 rates that one would look to to ascertain when it 4 was time to rebalance? 5 A. Well, as I mentioned, as frequently as 6 possible, I noted an article that I understand was 7 available to USAT by Terry Smith of the Federal 8 Home Loan Bank of Dallas indicating minor 9 increments of interest rates such as 25 or 50 10 basis points. I don't have the exact numbers at 11 hand. 12 Q. Do you have any knowledge other than 13 those articles of what would be the appropriate 14 time? 15 A. Well, certainly in the broker/dealer 16 community, if one is maintaining a balance between 17 hedges and mortgage-backed securities portfolios, 18 it would be done essentially on a daily basis. 19 So, interest rates wouldn't change -- typically 20 wouldn't change to such a great degree on a given 21 day that one would need to be out of balance by 22 any significant degree with regard to that measure 3111 1 of interest rate risk. 2 Q. What is your understanding of the 3 stated objectives of USAT for its risk-controlled 4 arbitrage portfolios? 5 A. Well, the stated objectives were to 6 earn a positive income hedged, meaning that the 7 income would not be affected by changes in 8 interest rates or any other changes for the 9 portfolio. And that spread income would be stated 10 objectives, the difference between the yields on 11 the mortgage-backed securities and the payments on 12 the swaps and short-term loans that were used to 13 finance that position. 14 Q. Now, did USAT manage its portfolio in a 15 way to achieve those objectives? 16 A. No. In fact, I would maintain that it 17 couldn't have been assured of meeting those 18 objectives no matter what hedging strategy it 19 pursued. But in any case, it didn't pursue a 20 hedging strategy which even came close to meeting 21 those objectives. As we know, they did, in fact, 22 not maintain positive income. They lost 3112 1 substantial amounts on that strategy. 2 MR. GUIDO: No further questions, Your 3 Honor. 4 THE COURT: Are you going to 5 cross-examine, Mr. Nickens? 6 MR. NICKENS: Yes, sir, I am. 7 THE COURT: Do you want to start before 8 lunch or later? 9 MR. NICKENS: Whatever your wishes are, 10 Your Honor. It'll take us a few minutes to get 11 set up, but whatever you wish. 12 THE COURT: All right. Let's adjourn 13 until 1:15. 14 MR. NICKENS: Thank you. 15 16 (Luncheon recess.) 17 18 THE COURT: Be seated, please. The 19 hearing will come to order. I believe the direct 20 examination of Mr. Duffie is complete and 21 Mr. Nickens is going to cross examine. 22 MR. NICKENS: Thank you, Your Honor. 3113 1 2 CROSS-EXAMINATION 3 4 (1:23 p.m.) 5 Q. (BY MR. NICKENS) Dr. Duffie, we have 6 met before, have we not? 7 A. Yes, Mr. Nickens. 8 Q. On the occasion of your depositions or 9 two sessions of your deposition, I asked you some 10 questions? 11 A. That's correct. One day in May, I 12 think it was, and two days in July. 13 Q. Have you reviewed your deposition? 14 A. I've gone over it. 15 Q. Did you correct it in those areas where 16 there were mistakes? 17 A. I've gone over it so recently that I 18 haven't taken the opportunity to make any 19 corrections. 20 Q. And so, you haven't signed it? 21 A. No, I haven't signed it. 22 Q. But you have reviewed it and is it 3114 1 basically accurate? 2 A. I couldn't vouch for it line by line. 3 When I reviewed it, I didn't notice anything that 4 I disagreed with with regard to my answers. 5 Q. Now, I'd like to ask you a few 6 questions to clarify your testimony. Are you 7 willing and prepared to do that? 8 A. I'll do my best, sir. 9 Q. Thank you. Now, would you say that you 10 know a lot more today about risk-controlled 11 arbitrage and mortgage-backed securities than you 12 knew, say, in 1984? 13 A. I think that's a fair assumption. 14 Q. Where were you in 1984? 15 A. I was just beginning my job at Stanford 16 University on the finance faculty. 17 Q. And have you made any effort in 18 connection with this work to try to research what 19 the state of knowledge was concerning 20 mortgage-backed securities and risk-controlled 21 arbitrage in 1984 and 1985? 22 A. Well, we had discussion earlier about 3115 1 the state of knowledge, for example, in OAS 2 modeling. So, I did take the opportunity at least 3 incidentally to refresh my recollection through 4 the literature on that. 5 Q. On OAS only? 6 A. Not only. But just recently, I was 7 looking at that. I also read, for example, 8 Mr. Carron -- some of Mr. Carron's writings on 9 mortgage-backed securities around that time, the 10 status in the mid-Eighties. 11 Q. Have you made any effort to put 12 yourself in the position of the person making the 13 decision at the time? 14 A. Well, I would say as I went through the 15 various things that I did to prepare for my 16 testimony that I attempted as much as possible to 17 think in terms of the risks that were present at 18 that time, yes. 19 Q. Isn't it true that what you've 20 basically done is come in some 10 to 12 years 21 later and second-guessed those decisions? 22 A. Well, I wouldn't put it in those terms. 3116 1 I'm certainly here 10 to 12 years later. I can't 2 dispute that. And I've done the best job that I 3 can with the information that I have available 4 now. I will say that, of course, I wish I had 5 more information that was current at the time, but 6 I've done the best job I could. 7 Q. So, you have tried to put yourself in 8 the position of the person making the decisions 9 back there in 1984, '85, and '86? 10 A. What decisions are you speaking of 11 right now? 12 Q. How to structure the risk-controlled 13 arbitrage, how to deal with the roll-down, how to 14 deal with unexpected events like interest rates 10 15 or 15 times what you had been -- or repayment 16 rates 10 or 15 times what had been projected. 17 Have you tried to put yourself in the 18 position of that person or persons? 19 A. I haven't verified the aspect of your 20 question dealing with how many times prepayment 21 rates were but I believe, for example, I answered 22 some questions this morning about alternatives 3117 1 that were in place at the time and in general 2 terms, yes, I tried to understand the risks that 3 were faced by the institution at that time. 4 Q. Have you ever actually managed a 5 risk-controlled arbitrage? 6 A. No, I haven't. 7 Q. Have you ever actually managed a 8 mortgage-backed securities portfolio? 9 A. It depends on what you mean by 10 "manage." I've certainly advised banks on the 11 management of mortgage-backed securities 12 portfolios. 13 Q. You've been a consultant to 14 institutions that are considering investments in 15 mortgage-backed securities? 16 A. Not only considering but are in the 17 active state of managing such investments, yes. 18 Q. But have you ever had to make the 19 day-to-day decisions about what to do with the 20 portfolio as events unfold? 21 A. No, sir. That's not my job. 22 Q. Now, this isn't the first time that you 3118 1 have testified concerning a savings and loan 2 institution. You've told us that? 3 A. Yes, sir. 4 Q. And, in fact, you have worked for the 5 OTS on how many prior occasions? 6 A. Well, I have to be careful as to 7 exactly what branch of the government I'm 8 literally working for. In the case of the 9 Franklin Savings, I believe it was the OTS 10 although there were other parties to that 11 litigation. And in the case of Far West, I was 12 actually working for the Department of Justice who 13 were the attorneys on behalf of, I think, more 14 than one government agency. But in general terms, 15 I have worked with OTS before, yes. 16 Q. And, in fact, you've worked for the 17 FDIC in this case or in a related case? 18 A. Well, I guess it depends on what you 19 mean by "have worked for." I understand that they 20 have an interest in this case. I'm not sure in 21 terms of the legal situation what that means. I 22 haven't had that explained to me. 3119 1 Q. Well, did you prepare a report for the 2 FDIC in the matter pending here in Houston? 3 A. That's not correct. I prepared a draft 4 of a report. 5 Q. Oh, I'm sorry. You didn't reach a 6 final conclusion? 7 A. No. I didn't reach a final conclusion. 8 I drafted a report. 9 Q. So, you did draft a report and -- 10 A. That's correct. 11 Q. Okay. Now, have you ever looked at an 12 S&L institution that didn't have interest rate 13 risk? 14 A. No. 15 Q. That's sort of in the nature of the 16 lending business, isn't it? 17 A. Generally in the thrift industry, yes. 18 Q. And have you ever considered an 19 institution in which -- well, let me ask you this. 20 Have you ever testified on behalf of any of the 21 people accused of mismanaging a savings and loan 22 institution? 3120 1 A. No. 2 Q. Have you ever -- in all the cases 3 you've been asked to look at, have you ever found 4 that it was properly managed? 5 A. Well, there is a total of two cases and 6 in both of those cases, I did not provide opinions 7 on whether the institution was properly managed 8 but, rather, whether the risks associated with the 9 mortgage-backed securities investments were large 10 relative to the capital of those institutions. 11 Q. But what you did in Franklin Savings is 12 pretty much what you've done in this case, isn't 13 it? 14 A. I guess it depends on the term "pretty 15 much." There is certainly very similar aspects to 16 the risks that that mortgage-backed securities 17 portfolio was subject to and the risks that the 18 USAT mortgage-backed securities portfolio was 19 subject to. 20 Q. And the judge that heard your testimony 21 in that case was whom? 22 A. I believe it was Judge Saffels. 3121 1 Q. And he did not accept your testimony, 2 did he? 3 A. I'm not sure what you mean by "accept." 4 I believe that I was accepted by the Court as an 5 expert witness in that matter and I believe that 6 my testimony was taken. As to whether Judge 7 Saffels agreed with what I had to say on every 8 count, I don't have specific knowledge. 9 Q. Well, you know that he published an 10 opinion in which he said basically that he didn't 11 find your testimony persuasive. Right? 12 A. I was told that there were words in his 13 opinion that suggested that a government -- an 14 expert witness working for the government's 15 testimony was not persuasive on some points. 16 Q. You didn't look at that opinion? 17 A. No. I had a copy of it. I didn't read 18 it. 19 Q. You haven't read it? 20 A. No, I haven't. 21 Q. Not the least bit curious about what he 22 had to say? 3122 1 A. I was curious. 2 Q. I'm going to hand you what's been 3 marked as Exhibit B3773. 4 MR. NICKENS: B3773, Your Honor, 5 purports to be the decision of Judge Saffels 6 reported at 742(f) sub 1089 District Court of 7 Kansas 1990. And we offer B3773. 8 MR. GUIDO: Objection, Your Honor. The 9 opinion has been overturned and it doesn't stand 10 as a statement of fact. So, therefore, if this 11 counsel is offering it for proof as to the truth 12 of the facts asserted therein, we object to it as 13 being unreliable and not admissible under the 14 rules of this court. 15 THE COURT: Well, it's a public 16 document. I can take official notice of it even 17 if it's not offered. I don't see the point in 18 objecting. Obviously, the document for whatever 19 status it may have -- I know the decision was 20 overruled. Let's proceed. 21 Q. (BY MR. NICKENS) Dr. Duffie, if you 22 could turn over to Paragraph 205 which is at 3123 1 Page 1123. Do you have that? It's in the lower 2 right-hand corner of the page. Do you see 3 Paragraph 205? 4 A. Yes, I do. 5 Q. Why don't you read it for us? 6 A. Paragraph 205 and I quote, "Dr. Darrell 7 Duffie, professor of finance at Stanford 8 University, was called as a witness by the OTS to 9 testify about the types of risks facing Franklin's 10 operations, his reservations with Franklin's 11 computer models, and the changes in Franklin's net 12 equity value. Dr. Duffie's testimony consisted of 13 theoretical evaluations of the risks presented to 14 Franklin's operation and other thrifts in general. 15 Moreover, Dr. Duffie's testimony was largely an 16 attempt by the OTS to provide some post hoc 17 justification for the regulators' action. The 18 Court finds that Dr. Duffie's testimony is of 19 little help in this case and grants it little 20 weight regarding the practical reality of 21 Franklin's economic condition and operations and 22 the necessity for pointing -- for appointing a 3124 1 conservator." 2 Q. And the Court indicates at Footnote 11 3 that it had a motion pending to strike certain 4 parts of your testimony because of errors that had 5 been made which it was -- which motion was denied? 6 A. I'm just reading that footnote. Yes. 7 I just read that. 8 Q. It says "The Court is fully aware of 9 the errors in the numbers used by Dr. Duffie as 10 pointed out in plaintiff's letter response, 11 Document 383, and has considered the corrected, 12 accurate numbers in consideration of this 13 exhibit and of Dr. Duffie's testimony." 14 Now, isn't that pretty much the role 15 that you played in this case? 16 A. Isn't what -- which -- 17 Q. As described by Judge Saffels in the 18 Franklin case. 19 A. Well, let me just go back. I was 20 called as a witness by the OTS, that's correct. I 21 did testify about the types of risks facing 22 Franklin's operations, that's correct. I did 3125 1 testify about my reservation with Franklin's 2 computer models, that's correct. And I also 3 testified about the changes in Franklin's net 4 equity value. All that's correct. 5 Q. And how would you compare that to your 6 role in this case? 7 A. Well, there's certainly some of these 8 issues that I've addressed in this case, as well. 9 Q. And in the Far West case, you had 10 pretty much the same role, did you not? 11 A. Well, I was again called by the 12 government to testify about the risks facing Far 13 Wests institution through mortgage-backed 14 securities investments. 15 Q. And you reached basically the same 16 conclusions? 17 A. Well, the matter was substantially 18 different. For example, significant differences 19 in the modeling used by Far West and by Franklin 20 were in place. I would more liken the Far West 21 situation to the USAT situation treating those two 22 as somewhat different than the Franklin situation. 3126 1 Q. So, to be clear, was the Far West 2 situation like this one or not? 3 A. It was more like this one than the 4 Franklin case was. 5 Q. Would you say that they -- you reached 6 basically the same conclusions? 7 A. No. I did in both cases reach the 8 conclusion that there were substantial risks and 9 that those risks were excessive for the capital 10 available in those institutions. 11 Q. So, you reached a somewhat different 12 conclusion in Far West? 13 A. Yeah. In the course of reaching those 14 conclusions, there were different issues that were 15 to be considered in the case of Franklin and Far 16 West. 17 Q. Well, Dr. Duffie, let me ask you some 18 questions about the preparation of your report in 19 this case. Now -- and that was Exhibit 5 to your 20 deposition and is in the record as Exhibit A11012. 21 You're familiar with your report. Right? 22 A. Yes. 3127 1 Q. Would you like to have -- you have it 2 in front of you. Now, describe for me the process 3 of preparing your report. What did you do? 4 A. Well, as I testified in my deposition 5 when you asked this, I went back at the beginning 6 just in terms of collecting data on the 7 institution. Anything that would provide me with 8 information on the assets, liabilities, and hedges 9 of the institution, anything about the market 10 values of those positions, the nature of their 11 coupon and maturity structure, anything with 12 options embedded in them, collected all of that 13 information, and then started to organize the 14 information in terms of the balance sheet of the 15 institution, any data that I had bearing on that I 16 attempted to collect. Then the composition of the 17 portfolio with regard to its exposure to interest 18 rates. 19 Q. Sir, I don't -- I'll come back to that. 20 I don't mean to interrupt. But I'm interested in 21 the mechanical aspects of the preparation of the 22 report. 3128 1 A. Well, okay. To be very brief, review 2 of documents, collection of data, data analysis, 3 and presentation of results. 4 Q. And then you sat down at a keyboard and 5 started typing it. Right? 6 A. The report itself? 7 Q. Yes, sir. 8 A. Yes, that's correct. 9 Q. And so, you typed this report yourself 10 personally? 11 A. This typescript is not my typescript. 12 I wrote my own typescript and then had my research 13 assistant do optical character recognition of my 14 typescript in order to be able to merge it with 15 the graphical illustrations that are shown in the 16 report. My software did not allow me to input 17 graphical illustrations. 18 Q. Well, let me show you Exhibit 5 to your 19 deposition. Is there any difference from 20 Exhibit 5 and the exhibit that you have in front 21 of you? 22 A. Well, the one that I'm holding is 3129 1 marked "attorney work product," but that's not the 2 distinction that I was making earlier. 3 Q. With regard to Exhibit 5 to your 4 deposition which I have handed you and which is 5 another copy of your report, did you type that one 6 out yourself? 7 A. Well, as a technical distinction here, 8 I typed the report myself and then as I mentioned 9 a moment ago, I gave the report to my research 10 assistant and had him scan it by optical character 11 recognition into his computer, insert the charts 12 which I had instructed him to do, and then produce 13 this hard copy version which was then subsequently 14 Xeroxed and provided to you, among others. 15 Q. And as you told me in your deposition, 16 you typed some parts and somebody else typed 17 others. Right? 18 A. I'm not sure exactly what you meant by 19 that. I typed the original draft myself and then 20 as I revised the report, I had my research 21 assistant type in the corrections that I -- 22 adjustments that I asked him to make. 3130 1 Q. Well, let me ask you to focus on Page 2 356 of your deposition. I'm going to hand that -- 3 put that in front of you. And if we could go to 4 the last line of Page 355. And if it's okay with 5 you, Dr. Duffie, I'll just read the questions and 6 then you read out your answers. 7 A. Okay. 8 Q. And if you recall it any differently, 9 please tell us. 10 A. That's fine. 11 Q. At Line 22, Page 355, I asked you, 12 "This is something you typed yourself?" 13 And you said -- 14 A. "This" meaning this particular draft? 15 Q. "Exhibit 5." 16 A. "I'm going to give you a strange 17 answer. Yes and no." And then, in parens, it 18 indicates that I chuckled or laughed. 19 Q. And I said, "You typed some portions, 20 and other portions were typed by others?" 21 A. Answer, "That's correct." 22 Q. "Okay. What portions can you identify 3131 1 as being your work?" 2 A. My answer, "It was all originally typed 3 by me and then -- and I've forgotten why. It 4 could have been because I wanted the charts 5 inserted electronically and I didn't have the 6 means to do that myself. I sent my typed script 7 to Tom Varner and he actually used scanning, if 8 I'm not mistaken. Either that or I sent it to him 9 on a floppy disk and then he captured the 10 electronic document and inserted the figures." 11 Q. And I asked you, "And what source 12 materials did you use for preparing your 13 conclusions?" 14 A. Answer, "Everything that I could lay my 15 hands on, all of the documents and analysis that I 16 had done." 17 Q. "For USAT?" 18 A. "For anything that related to USAT, 19 yes." 20 Q. "Right. Well, I mean, this is 23 21 single-spaced pages that -- that indicates you 22 spent 14 hours to prepare?" 3132 1 A. Answer, "Are you indicating that I may 2 have spent only 14 hours in preparing this 3 report?" 4 Q. "I'm just trying to get an idea of the 5 process again and you're saying that you sat down 6 and just typed this out and produced one draft or 7 maybe more drafts which I gather were then 8 transmitted to -- well, I'd better ask the 9 question. Were the earlier drafts transmitted to 10 anyone?" 11 A. "No." 12 Q. "Did you discuss the drafts with 13 anyone?" 14 A. "Yes." 15 Q. "And that was Mr. Guido, Mr. Veis, and 16 Mr. Rogers?" And you said -- 17 A. "No. Only Mr. Guido." 18 Q. Then if we can go over to Page 359 at 19 Line 18, I asked you, "Did anyone else participate 20 in the drafting of the report?" And you said -- 21 A. I'd just like to read the intervening 22 text to make sure I'm with you. 3133 1 Q. Okay. 2 A. Okay. I'm with you again. 3 Q. Any other information you'd like the 4 Court to have there in the intervening pages? 5 A. No. I just wanted to read everything 6 to where we left off. 7 Q. Line 18, Page 359. "Did anyone else 8 participate in the drafting of the report?" 9 A. "Not other than what I've already 10 mentioned about Mr. Varner assembling the 11 electronic version of the report." 12 Q. "And in terms of the actual sitting 13 down there and drafting it, you just drafted an 14 original report?" 15 A. Answer, "With my own ten fingers." 16 Q. "Okay. And the source materials were 17 the USAT materials that you had received from 18 either Mr. Rogers or Mr. Guido?" 19 A. Answer, "The raw source materials, 20 yes." 21 Q. "Were there any other source 22 materials?" 3134 1 A. Answer, "Pardon me. Other source 2 materials that I had gotten on my own were 3 interest rate data, stock market data, and the 4 like which we've already discussed." 5 Q. "Right. Anything else as source 6 materials for your report?" 7 A. "Not other than information in the 8 public records such as I've mentioned." 9 Q. Now, is there anything about your 10 deposition testimony that -- anything new that you 11 have recalled as you've reviewed it? 12 A. This testimony right here? 13 Q. Yes, sir. 14 A. Not as I sit here now. Not that I 15 recall. 16 Q. Now, Mr. Guido asked you some questions 17 about the documents that you relied on, did he 18 not? 19 A. Yes. 20 Q. And that was Exhibit 3 to your 21 deposition and is identified in the record as 22 A11015. 3135 1 MR. NICKENS: May I ask where the trial 2 exhibits are from this morning? 3 Q. (BY MR. NICKENS) A11015 you 4 identified and went through with Mr. Guido. 5 Right? 6 A. Not entirely, but we did cover some of 7 the documents listed in this -- in this exhibit. 8 Q. And this is supposed to be a 9 representation of all the documents that you 10 relied upon in preparing your report? 11 A. Well, it says reviewed and relied upon. 12 But if you take that -- yes. 13 Q. And there are no others? 14 A. Well, I can't give you a categorical 15 "yes" answer. But as far as I knew when I 16 prepared this document, yes. 17 Q. Now, I'm going to hand you a document 18 that's been marked as B3772, Dr. Duffie. Do you 19 recognize that as your report in the Far West 20 case? 21 A. I'd like to review it just for a 22 moment, please. It's both a copy of my report as 3136 1 well as a journal article. 2 MR. NICKENS: We offer B3772, Your 3 Honor. 4 MR. GUIDO: No objection, Your Honor. 5 THE COURT: Received. 6 Q. (BY MR. NICKENS) Dr. Duffie, I'd like 7 for you to take your report in this case and your 8 report in the Far West case. Can you look at 9 Paragraph No. 1? 10 A. Of which report? 11 Q. Of both reports. Are they essentially 12 identical? 13 A. There are some changes and, if I may 14 say, I haven't mischaracterized the preparation of 15 my report. You will find that there are 16 substantial similarities between lines in my 17 report on the Far West case and lines in my report 18 on USAT with regard to certain matters. 19 Q. Yes, sir. Are you telling this Court 20 you did not use your report in Far West in 21 preparing your report in this case? 22 A. No. I did use it, in fact. 3137 1 Q. And you just merely forgot that when I 2 asked you about whether there were any other 3 documents that you had relied upon? 4 A. No, I did not forget that. This -- 5 Q. So, you told the Court -- 6 A. Excuse me, sir. I haven't finished my 7 response. This exhibit which you were asking me 8 about a moment ago is marked USAT documents 9 reviewed and relied upon. 10 Q. Well, had you forgotten about it when I 11 asked you at your deposition whether you had 12 relied on any other documents? 13 A. Frankly, yes. 14 Q. And when we read your testimony from 15 just a few minutes ago about what you used in 16 preparing your report, you had forgotten about it? 17 A. When we were speaking a few minutes ago 18 about documents that I used in preparing for my 19 report, we were speaking about USAT documents. 20 Q. Did I not ask you whether there were 21 any other documents that you relied upon as source 22 materials for your report? 3138 1 A. I believe you did ask that and, yes, I 2 had forgotten. 3 Q. And you not only forgot about it at the 4 time of your deposition, you forgot about it a few 5 minutes ago? 6 A. I want to review with regard to what 7 questions you asked in my deposition. Can you 8 remind me what the question was? 9 Q. Page 360. 10 A. Yes. When you asked about any other 11 source materials, I should have mentioned that I 12 referred to my own report in the Far West case. 13 I'm sorry about that. It was an oversight. 14 Q. And the similarities between the 15 reports go not only just to your background, et 16 cetera. They go to the conclusions, don't they? 17 A. Well, I haven't compared the reports; 18 but they certainly use similar language in 19 describing the types of risks facing 20 mortgage-backed securities investments such as 21 risks associated with changes in interest rates, 22 prepayment rates, volatility, and so on because 3139 1 these risks are generic to mortgage-backed 2 securities investments. 3 Q. Let me ask you to look over at 4 Paragraph 9 of your Far West report and compare it 5 to Paragraph 10 of your report in this case. And 6 why don't I read the Far West one while you -- "My 7 conclusions are summarized as follows." That 8 line's the same. Right? 9 A. That's correct. 10 Q. And you say "Through the RCA Far West 11 assumed" -- and here you say "through its MBS 12 portfolio, USAT assumed." Right? 13 A. That's correct. 14 Q. Basically changed the name. Right? 15 A. The name of the portfolio and the 16 institution, that's correct. 17 Q. "Assumed material risks that created a 18 significant and substantial possibility that the 19 institution, in an effort to increase its expected 20 return" -- and then in the Far West, you say "in 21 the short run." That's left out? 22 A. That's correct. 3140 1 Q. "Would ultimately incur major losses." 2 And then in Far West, "on its RCA investments"? 3 A. That's correct. 4 Q. You stopped the sentence there. Right? 5 A. Yes. 6 Q. And you were looking at your Far West 7 report when you drafted these conclusions, weren't 8 you? 9 A. Absolutely. 10 Q. And you said, "Moreover, the steps 11 taken" -- and here it's by Far West and here it's 12 USAT -- "to limit the risk associated with" -- and 13 in Far West, it's "RCA investments" and here, 14 it's, what, "MBS"? 15 A. That's right. 16 Q. "To limit the risk associated were 17 generally unsophisticated and ineffectual," 18 exactly how you described it in Far West. Right? 19 A. Correct. 20 Q. "In essence, the RCA investments" -- 21 and here it's what, in essence? 22 A. The MBS investments. 3141 1 Q. "Made by" -- and it was Far West and 2 here it's USAT -- "amounted to little more than a 3 wager." Right, Dr. Duffie? 4 A. That is correct. 5 Q. And -- now, did you tell us a few 6 minutes ago that the Far West case was different 7 or the same as this? 8 A. I said it was more similar to this case 9 than it was to the Franklin case, than either of 10 the two institutions were to the Franklin case. 11 Q. Yes, sir. And you went on to tell us 12 that it was different? 13 A. Yes, it was. 14 Q. But you reached this very same 15 conclusion? 16 A. That's not correct. I reached similar 17 conclusions with regard to the points you just 18 raised. If you compare other aspects of the 19 report, you'll find that there are a number of 20 other conclusions that are quite different. 21 Q. Sir, we can go through here and look at 22 the ones that are exactly the same; but there 3142 1 isn't any question that you had this in front of 2 you as you drafted your USAT report, is there? 3 A. No. There is no question. 4 Q. And just as a parenthetical, in Far 5 West, you criticized their use of scenario 6 analysis, didn't you? 7 A. I'd have to review my report on that 8 account. 9 Q. Paragraph 70 on Page 32. 10 MR. GUIDO: Paragraph 70. 11 MR. NICKENS: 7-0, 70. 12 A. Yes, I did. 13 Q. (BY MR. NICKENS) Now, have you been 14 told that Mr. Smith, who testified yesterday, 15 indicated that scenario analysis was the preferred 16 methodology at least by him? 17 A. I believe that scenario analysis that 18 he's discussing is not the same as the scenario 19 analysis that was -- as the term was used by Far 20 West and was used by them to analyze their 21 interest rate exposure. 22 Q. Well, how would you know that? Were 3143 1 you here yesterday? 2 A. No, but I'm familiar with the writings 3 of Mike Giarla and the documents put out by Smith 4 Breeden with regard to scenario analysis. 5 Q. I'm sorry, sir. I'm talking about 6 Mr. Terry Smith -- 7 A. Pardon me. 8 Q. -- of the Federal Home Loan Bank of 9 Dallas. 10 A. No, I'm sorry. I wasn't referring to 11 that Mr. Smith at all. 12 Q. So, you don't know what he meant by 13 "scenario analysis"? 14 A. No. 15 Q. Let me turn to a different topic, 16 Dr. Duffie. You testified earlier that you didn't 17 believe that USAT's reports accurately reflected 18 their economic situation? 19 A. That's correct. 20 Q. Now, you are not telling the Court that 21 they didn't accurately report their situation 22 according to the accounting rules, are you? 3144 1 A. No. I wasn't opining on what the rules 2 of accounting are. I was comparing the 3 information provided by the accounting reports in 4 comparison with the actual economic condition of 5 the institution. 6 Q. But you knew that USAT was required to 7 use certain rules for making its financial 8 reports? 9 A. Well, technically speaking, UFGI is 10 required in its reports to use certain rules as a 11 publicly-traded corporation. 12 Q. And you're not telling the Court that 13 they didn't follow those rules? 14 A. I'm not opining on whether they did or 15 did not follow the accounting rules. I'm 16 discussing the manner in which the accounting 17 performance did not reflect the true economic 18 performance of the institution. 19 Q. Okay. But when you're saying they 20 didn't report it correctly, you're not saying they 21 didn't report it according to the rules that 22 govern their activities? 3145 1 A. I'm not providing an opinion on whether 2 they -- whether the accounting reports provided by 3 USAT were consistent with accounting rules. 4 Q. In effect, you're criticizing the 5 accounting rules, aren't you? 6 A. As measures of economic performance, 7 yes, I am. 8 Q. And you prefer or would prefer a 9 mark-to-market system? 10 A. Well, insofar as accounting information 11 is used both by USAT and by regulators, I would 12 prefer to see the information on an economic 13 basis; that is, using mark-to-market, for example. 14 Q. Now, you're not telling the Court that 15 USAT didn't report both, are you? 16 A. Well, its accounting statements are 17 accounting statements. It did not provide 18 supplemental reports with its accounting 19 statements that I know of that provided the actual 20 economic performance. 21 Q. Aren't you aware of the fact that they 22 reported in their 10Ks and other SEC and Federal 3146 1 Home Loan Bank filings mark-to-market information? 2 A. But not with the accounting reports. 3 There were separate notes indicating some 4 information on market values, yes. 5 Q. What do you mean by "separate notes"? 6 You mean it was a footnote disclosure as opposed 7 to some other form of disclosure? 8 A. Either a footnote disclosure or 9 incidental information in the description of the 10 institution. 11 Q. So, am I to understand your criticism 12 now is that they didn't have a line item in their 13 financial reports for the mark-to-market 14 information? 15 A. I am not criticizing the institution 16 for the manner in which they provided information 17 to the public. My opinion was that the 18 information in the accounting reports does not 19 accurately reflect the true economic performance 20 of the institution. 21 Q. Now, if your system of mark-to-market 22 reporting were adopted, what percentage of the 3147 1 S&Ls in the United States would have been solvent 2 in 1985? 3 A. Well, first of all, the notion of 4 mark-to-market is not my system. It's a widely 5 recognized manner in which to keep track of the 6 economic performance of a financial portfolio. 7 Q. Regardless of whose it is, what 8 percentage of S&Ls would have been solvent if the 9 powers that be had seen fit to adopt 10 mark-to-market accounting? 11 A. I can't give you an opinion. I haven't 12 studied that. 13 Q. Well, you know it's well over 14 80 percent, don't you? 15 A. No, I don't know that. I just haven't 16 studied that. 17 Q. Are you familiar with the literature 18 indicating that the entire industry was insolvent 19 by billions of dollars on a mark-to-market basis? 20 A. No, I'm not familiar with that 21 literature. 22 Q. Well, don't you think that you ought to 3148 1 look at the practical consequences of your 2 proposals in evaluating whether they are a good 3 idea? 4 A. My proposal, which is to provide 5 information which accurately reflects the economic 6 performance of the institutions, would not change 7 the consequences. That is, providing the 8 information on how much losses there were in those 9 portfolios would not change the fact that the 10 losses were there one way or the other. So, 11 whether the institutions were insolvent or not was 12 not a question of how the information was 13 reported. 14 Q. Well, you as an academic then wouldn't 15 feel it necessary to take into account that 16 hundreds of institutions would be insolvent under 17 this system? 18 A. They would be as insolvent under one 19 accounting reporting regime as they were under 20 another. All I'm arguing is that the information 21 on the economic performance of the institution was 22 not available in the accounting numbers. 3149 1 Q. But you're not saying that about USAT? 2 A. Yes, I am. 3 Q. Well, I thought -- I don't want to 4 cover the same ground, but I thought you told us 5 that you recognized that they had made at least 6 footnote disclosure of mark-to-market losses? 7 A. And I said a moment ago that that 8 information was not recognized in the accounting 9 numbers. 10 Q. Well, could they have recognized such 11 numbers according to the accounting rules? 12 A. I'm not sure what the accounting rules 13 say. I believe it's allowed, but I'm not sure. 14 Q. Well, I'm trying to get at -- are you 15 suggesting that they do something lawless -- that 16 is, use something that would have been against the 17 rules? 18 A. I -- pardon me. 19 Q. I mean, you're not suggesting that 20 surely? 21 A. I wasn't suggesting anything with 22 regard to what information they ought to put into 3150 1 their accounting reports in terms of meeting 2 accounting requirements. Rather, I was suggesting 3 that the accounting numbers themselves do not 4 accurately reflect the true economic condition of 5 the institution which was substantially worse than 6 indicated by their accounting. 7 Q. And that would be true for many, many 8 other institutions, as well, correct? 9 A. I wouldn't be surprised if there were 10 many institutions at that time that were in a 11 similar situation, but I have no idea what the 12 numbers would be. 13 Q. Let me ask you a few questions about 14 your sensitivity analysis which is at Page 21 of 15 your report. That's where the table is. Now, in 16 making those calculations, you have to make a 17 prepayment assumption to calculate them, don't 18 you? 19 A. For the middle column which is marked 20 "USAT sensitivity," I did have to make prepayment 21 assumptions. 22 Q. That's the key column, isn't it? 3151 1 A. Well, with regard to where one uses 2 prepayment assumptions, it's the only column in 3 this report -- in this particular table for which 4 one would use prepayment assumptions. 5 Q. And you made prepayment -- the 6 prepayment assumptions you used were -- because it 7 happened in the past -- were the actual 8 prepayments. Right? 9 A. I used the historical average 10 prepayment behavior. 11 Q. That actually occurred? 12 A. That actually occurred. 13 Q. And so -- now, if you had been in the 14 position of the person trying to make the 15 decision, you wouldn't have had the benefit of 16 knowing what had happened because it was in the 17 future. Right? 18 A. That's correct. 19 Q. And that person has to make a 20 projection. Right? 21 A. That's correct. 22 Q. Because it's in the future? 3152 1 A. That's correct. 2 Q. And if the projection is incorrect 3 because of what then happens later, he can be in 4 trouble. Right? 5 A. It depends on what you mean by "in 6 trouble." 7 Q. Well, things will not happen as he 8 projected because his projection is wrong. It's a 9 fairly straightforward -- 10 A. Any projection of prepayments will not 11 turn out to be correct. 12 Q. And prepayments in this particular time 13 period -- that is, '85, '86, and '87 -- were 14 particularly volatile, weren't they? 15 A. Yes. 16 Q. Unexpectedly volatile, weren't they? 17 A. I don't think that I have enough 18 information to say how unexpected they were; but 19 certainly, they were volatile. 20 Q. Well, I can show you a document that we 21 discussed yesterday that indicated that persons 22 that were making projections on Ginnie Mae 12s in 3153 1 January of 1985 were using an SMM of .6 and that 2 in August of that year, SMM prepayments reached 3 levels over 8 percent. That is, the monthly 4 prepayments were over 8 percent. 5 Would you say that's a volatile market? 6 A. An 8 percent monthly mortality for 7 mortgage rates at that time, certainly very high 8 prepayment rates. I'm not sure what the 9 volatility of the prepayment rates themselves was. 10 Q. And if a person had eight months before 11 been projecting prepayment rates of .6, they would 12 have been wildly off. Right? 13 A. If that was -- in fact, if someone was 14 projecting .6 and the answer turned out to be 8, 15 they would have been surprised, yes. 16 Q. And this is First Boston. They are a 17 well-known investment banking firm. Right? 18 A. That's correct. 19 Q. And they are one of those people that 20 had all the superior information about 21 prepayments? 22 A. Superior to USAT. 3154 1 Q. Well, how much superior does it matter 2 in those if you're off by that much? 3 A. I believe I, in fact, testified this 4 morning that even the best prepayment models are 5 not reliable in controlling prepayment risks 6 because prepayment rates will change unexpectedly 7 and dramatically as market conditions change. 8 Q. Wouldn't you say that if you knew that, 9 as history unfolded, that a loss eventuated and 10 then you use historical figures as opposed to 11 projections, that it's sort of a self-fulfilling 12 prophecy? 13 A. What's the prophecy? 14 Q. Sir, if you use the historical 15 information knowing that that history resulted in 16 losses, aren't you going to inevitably produce a 17 loss? 18 A. Well, first of all, let me say that 19 there is no attempt here to mislead by using 20 information that was averaged through this period. 21 This was the only information available to me at 22 the time that would cover the range of instruments 3155 1 in this portfolio. So, I provided the best 2 available estimate that I could get as of today 3 with regard to the interest rate sensitivity of 4 that portfolio and found what projections were at 5 the time. 6 Q. Sir, couldn't you have gone back to 7 research this? You could have found projections, 8 couldn't you? 9 A. I don't think -- in fact, I think I 10 actually looked and I don't think I found 11 projections that handled the securities invested 12 in by USAT for the period covered from March '84 13 through March '88. 14 Q. But if you're trying to judge the 15 conduct of the people that are making the 16 decisions, wouldn't it be fairer to look at what 17 projected numbers were and not the historical 18 reality? 19 A. Well, certainly if I was attempting to 20 judge conduct, that would be one issue. But I 21 wasn't attempting to judge anyone's conduct here. 22 Q. You haven't been suggesting to this 3156 1 Court that the people involved in this decision 2 did so recklessly? 3 A. I didn't use those words. If you'd 4 like, I can give you an opinion on that. 5 Q. I think I know what your opinion is, 6 Dr. Duffie, but the question is: Have you been 7 suggesting to the Court and have you offered this 8 chart in support of your suggestions that the 9 respondents in this case were reckless in some way 10 or another? 11 A. I have not testified to that. Would 12 you like me to answer that question? 13 Q. No, sir. I think I know your answer. 14 I'm asking whether that has been what you have 15 been up to in your testimony to the Court. 16 A. I testified this morning that the 17 information provided in this chart among the other 18 information in my report definitely indicates that 19 the risks that this institution had were extremely 20 large relative to its capital. 21 Q. Without any suggestion as to what the 22 people who were there at the time should have 3157 1 appreciated. Is that what you're saying? 2 A. I didn't address that question this 3 morning. 4 Q. And so, instead of using projections, 5 you used the historical figures knowing all along 6 that those historical figures were errant? 7 A. No. That's not correct. 8 Q. You knew that they were wildly 9 different than the projections that everybody had 10 made? 11 A. Well, you mentioned one particular 12 number; but, no, for a fact, I don't know that 13 they are wildly different than projections that 14 people were making. 15 Q. So, you haven't studied that? 16 A. I used the best available information I 17 had. If you have more information, I'll be happy 18 to consider it. 19 Q. Okay. I'll come to that. But how do 20 you see that as being consistent with trying to 21 put yourself in the position of the persons that 22 were making the decisions at the time? 3158 1 A. I'm not placing myself in a situation 2 of the people that were making the decisions at 3 the time with regard to their measures of interest 4 rate risk. I'm attempting to estimate what the 5 interest rate risk was. 6 Q. Well, let's see what you're estimating 7 there. You said that you were estimating the 8 portfolio. Right? 9 A. When you say "the portfolio," I'm not 10 sure what you're referring to. 11 Q. Well, I thought that was the term that 12 you and Mr. Guido used frequently in your direct 13 testimony. I'm a little unsure, and that's what 14 I'm asking you about. And let me -- let me be 15 more specific. 16 When you say "the portfolio," you don't 17 mean Joe's portfolio? 18 A. I don't know what you mean at this 19 point by "the portfolio." 20 Q. Well, I mean the risk-controlled 21 arbitrage that was put into effect beginning in 22 late 1984 and which had been put in place by 3159 1 mid-1985 consisting of approximately $500 million 2 face amount of MBS and approximately the same 3 amount of swaps financed by reverse repos. That's 4 Joe's portfolio for this -- purposes of this 5 question. 6 A. Okay. 7 Q. You have made no effort to analyze that 8 portfolio standing alone, have you? 9 A. In fact, I made an effort but was 10 unable to get the necessary data in order to 11 analyze that portfolio. So, I analyzed the entire 12 institution. In fact, that's the appropriate 13 measure of interest rate risk facing USAT's 14 capital, is the entire institution. 15 Q. But if the judge has been asked to 16 consider the action with his regard to Joe's 17 portfolio, your information is no good, is it? 18 A. Well, with regard to Joe's portfolio 19 during that time period, I did not have 20 information that I could bring to bear on Joe's 21 portfolio. I had information on the institution 22 as a whole. 3160 1 Q. So, your testimony hasn't related to 2 Joe's portfolio as I have defined it for you? 3 A. I wouldn't say that entirely. This 4 table does not specifically refer to Joe's 5 portfolio or break those numbers out. It would 6 have been impossible to do so with the data that I 7 had. 8 Q. Now, you said you read Dr. Carron's 9 report. Right? 10 A. I believe -- I can't remember whether I 11 said that I've read it, but I did read it. 12 Q. And in his -- in his report, he breaks 13 up Joe's portfolio. Right? 14 A. He does refer to what he calls Joe's 15 portfolio. 16 Q. And he gives specific numbers and 17 specific coupon MBS, doesn't he? 18 A. Well, yes. 19 Q. So, when you say you didn't have the 20 information, you mean you didn't have the 21 information before you read Dr. Carron's report? 22 A. If one takes that portfolio that's 3161 1 described in Carron's report, that certainly was 2 not information available to me at the time that I 3 prepared my report, that's correct. 4 Q. Well, I was asking you about your 5 testimony just a few seconds ago. When you said 6 you didn't have the information, you didn't mean 7 to imply that you don't have it now? 8 A. Well, let's take it as a premise -- and 9 I'm not sure I'm going to agree with this -- that 10 the information provided in Carron's report is 11 actually information about Joe's portfolio as 12 you've described it. With that information in 13 hand, in principle, I could take that information 14 and do the same kind of analysis that I did for 15 the institution as a whole, remembering that I am 16 actually more interested in the interest rate risk 17 of the institution as a whole. I'm not sure 18 that's relevant for this discussion but if it was 19 necessary to do that, it could be done. 20 Q. And you didn't perform any stand-alone 21 analysis for United Mortgage Finance? 22 A. No. Same answer. 3162 1 Q. And when you say "take the institution 2 as a whole," are you including their portfolio of 3 whole loans? 4 A. Yes. 5 Q. So, you're looking at this institution 6 and measuring its sensitivity including the whole 7 loans that they had long before these respondents 8 had anything to do with the management of the 9 institution? 10 A. Well, certainly as one is investing in 11 mortgage-backed securities, one wants to consider 12 the impact of that investment on the entire 13 institution's risk relative to its capital. In 14 order to make a judgment about that, one needs to 15 know all of the interest rate sensitive 16 instruments on the portfolio, including the whole 17 loans. 18 Q. So, your sensitivity analysis includes 19 the whole loans? 20 A. Yes. 21 Q. And it includes mortgage-backed 22 securities that were owned in the liquidity 3163 1 portfolio? 2 A. I didn't break out anything 3 specifically with regard to liquidity portfolio. 4 I certainly included everything that was provided 5 by USAT to the Federal Home Loan Bank Board in its 6 reports. 7 Q. Well, were you told that there was a 8 liquidity portfolio that didn't have anything to 9 do with risk-controlled arbitrage? 10 A. I wasn't asked to consider such a 11 portfolio. 12 Q. And if that portfolio didn't have any 13 associated hedges, what effect would that have on 14 your sensitivity analysis? 15 A. Well, in principle, it shouldn't effect 16 it because I included everything. So, if it 17 infers a liquidity portfolio, whether or not there 18 were hedges attached to that portfolio all of it, 19 if it was correctly reported to FHHLBB, would be 20 corrected. 21 Q. So, we should be looking at the 22 changes, not the absolute amount? 3164 1 A. Changes in terms of -- 2 Q. The changes on your -- in your analysis 3 because we don't know how much of it is 4 attributable to, for example, the liquidity 5 portfolio? 6 A. I'm sorry. When you refer to changes, 7 could you tell me what you mean? 8 Q. Yes, sir. You've got a report there in 9 front of you. On Page 21 of your report where you 10 list some numbers and those numbers include the 11 sensitivity created by whole loans, 12 mortgage-backed securities owned outside of 13 risk-controlled arbitrage. It's the entire 14 institution, as you put it? 15 A. To the extent that it was reported by 16 USAT. 17 Q. Well, do you have any information that 18 they didn't record what -- properly record what 19 they had? 20 A. No. 21 Q. You weren't trying to suggest by your 22 answer that they had left something out? 3165 1 A. No. I'm relying on their information. 2 Q. Well, how about the speculative 3 portfolio? Are you familiar with that one? 4 A. Which one are you speaking of? 5 Q. I'm talking about the one that USAT set 6 up for the express purpose of trading 7 mortgage-backed securities. 8 A. Are you referring to what's been 9 referred to as the United MBS portfolio? 10 Q. No, sir. Are you unfamiliar with the 11 speculative portfolio? 12 A. I'm certainly familiar with the 13 speculative activity that occurred with United's 14 portfolio. But as to which specific portfolio 15 you're referring to right now, no, I'm not sure 16 what you're referring to. 17 Q. Do you know about the speculative 18 portfolio or not? 19 A. The speculative portfolio? 20 Q. Yes, sir. 21 A. No. Unless you enlighten me further, I 22 couldn't identify which of the portfolios held by 3166 1 USAT was the speculative portfolio. 2 Q. So, you don't know whether the trading 3 that you've identified was done in something 4 called a speculative portfolio? 5 A. Well, the -- 6 Q. Do you know that one way or the other? 7 A. The bulk of the trading in terms of 8 volume occurred in the United MBS portfolio. If 9 that's what you're referring to, then I do know 10 about that one. 11 Q. Now, let's talk a little bit about the 12 risk factors. And I'd like, if we could, to 13 compare MBS or risk-controlled arbitrage, if you 14 wish, to owning whole loans. 15 With regard to the interest rate risk, 16 is owning whole loans better or worse, risky or 17 less risky, than owning MBS? 18 A. Are you going to tell me what kind of 19 whole loans and what kind of MBS? 20 Q. Well, explain to us why that would make 21 a difference to your answer. 22 A. Well, just to take a hypothetical, if 3167 1 you gave me a 20-year whole loan and we're 2 comparing it with five-year adjustable rate 3 mortgages, then the 20-year whole loan would be 4 more interest rate risk sensitive. It might have 5 more credit risk, as well. 6 Q. I'm not -- 7 A. If we're comparing 30-year fixed rate 8 mortgages with six-month whole loans, then the 9 conclusion would be the opposite. 10 Q. Well, let's say that we're talking 11 about a whole loan of 30 years with a coupon of 12 10 percent and we're talking about an MBS that 13 comprises whole collection of those very same 14 loans. Okay? 15 A. Okay. 16 Q. Now, in terms of that comparison, which 17 one is riskier to own? 18 A. If they are identical -- that is, if 19 the MBS is literally holding the same portfolio as 20 in the whole loan portfolio -- then presumably 21 they are the same -- they represent the same risk. 22 Q. Well, aren't there different risks -- 3168 1 the same interest rate risk? 2 A. Well, is there more information here 3 about what the whole loans are? Are they 4 prepayable, non-prepayable? Are they 5 mortgage-backed securities, non-prepayable, 6 prepayable? 7 Q. You know -- I'm just asking you to 8 compare in a theoretical way -- and I want to talk 9 about identical -- in other words, you take a 10 collection of whole loans and they are the 11 underlying collateral for the MBS. And let's just 12 say that it happens that one of them got loose and 13 is owned out there free and clear. Okay? 14 A. Okay. 15 Q. That's what I'm trying to focus on. 16 A. Okay. 17 Q. Now, which one is riskier to own? 18 A. Well, if they are otherwise identical 19 and the MBS was agency pooled, then I would rather 20 go with the MBS. 21 Q. Okay. And the reason for that is 22 because you have done away with default risk. 3169 1 Right? 2 A. Yes. 3 Q. And you have added liquidity -- 4 A. Yes. 5 Q. -- with the MBS. Right? 6 A. Yes. 7 Q. Are there any other differences that 8 would make the ownership of the whole loans 9 preferable? 10 A. Of the whole loans portfolio or the 11 whole loan that got away from the MBS? 12 Q. Either one. Any other differences that 13 we should focus on? 14 A. Well, if it was a sufficiently 15 diversified MBS portfolio that was pooling this 16 mortgage-backed security, then there is another 17 reason there to go for the MBS and that is 18 diversification of some of the risks associated 19 with it. 20 Q. Now, as to prepayment risk with this 21 same situation, which is preferable? 22 A. With regard to prepayment risk, well, 3170 1 there is no credit risk. We're not talking 2 about -- that is, if we're not talking about 3 credit risk differences, liquidity risk 4 differences, the interest rate risks are roughly 5 the same putting aside prepayment risks. So, the 6 prepayment risk in the mortgage-backed security 7 portfolio, we'd have no reason to believe unless 8 we looked into the composition of the portfolio 9 that it would be any different than the whole 10 loan. 11 Q. Unless you had the diversification that 12 you mentioned which would make it preferable even 13 on prepayment risk. Right? 14 A. If you didn't have any other 15 information, you'd prefer the diversification, 16 that's right. 17 Q. How about liquidity? We talked about 18 that. But the MBS is more liquid? 19 A. Well, certainly if it's agency pooled. 20 Q. Now, Dr. Duffie, I'd like to ask you 21 some questions about this issue of sophistication. 22 Now, you're familiar with the qualifications of 3171 1 Sandy Lorenson. Right? 2 A. Well, I read her qualifications in her 3 deposition. 4 Q. You would -- based upon her 5 qualifications, you would surely expect that she 6 was competent to manage a risk-controlled 7 arbitrage. Right? 8 A. Depends on the arbitrage, depends on 9 the institution, depends on the tools she had 10 available. In terms of her personal 11 qualifications, I wouldn't have anything to argue 12 about that. 13 Q. Well, let's say you're going to hire 14 somebody to run your risk-controlled arbitrage and 15 you interview them and they seem fine and you look 16 at their resume and it says that they have 17 graduated from the Ph.D. program at MIT, that they 18 have worked under Louis Ranieri at Salomon 19 Brothers, and that they have been in charge of the 20 hedging activities at Freddie Mac. 21 Now, those are pretty strong 22 qualifications, aren't they? 3172 1 A. Are you referring to those at Ms. 2 Lorenson's qualifications? 3 Q. I am. I think the record will bear 4 that out unless I've made a mistake. 5 A. Well, I could stand corrected. I 6 didn't know that she actually had a Ph.D. from 7 MIT. I thought it was an MBA, but let's suppose a 8 that it was a Ph.D. for sake of discussion. 9 Q. Well, if I made a mistake -- for the 10 sake of discussion, assume it's either. 11 Would it make a material difference to 12 your answer? 13 A. Well, it could depending on the 14 institution and the risks involved and the types 15 of securities that are going into the portfolio. 16 Q. So, she -- let me get this right. It 17 would make a difference to your answer if you knew 18 that she was an MBA as opposed to a Ph.D. despite 19 the fact that she had gone on to be the head of 20 the hedging program at Freddie Mac and worked 21 under Mr. Ranieri? It would make a difference to 22 you? 3173 1 A. It could. 2 Q. Well, what kind of qualification -- I 3 mean, let's talk about hiring somebody. That's, 4 after all, what these respondents did. They hired 5 somebody. 6 What further qualifications would you 7 look for before you could hire them? 8 A. Mr. Nickens, I'm not taking issue with 9 the choice to hire Ms. Lorenson. I don't have any 10 disagreement, as I mentioned earlier, with her 11 personal qualifications. 12 Q. So, you wouldn't blame -- if the issue 13 is whether somebody was imprudent in hiring her, 14 you wouldn't blame them for that? 15 A. I don't have enough information at this 16 point without -- unless you tell me which RCA 17 portfolio and which institution to make a judgment 18 on that. I could design hypotheticals under which 19 perhaps her qualifications would not be 20 sufficient. But I don't have enough information 21 to say that she was unqualified. I just haven't 22 looked into her resume sufficiently deeply beyond 3174 1 the qualifications that you just mentioned. So, I 2 won't disagree with that if it helps things to 3 proceed. 4 Q. Based upon her qualifications, would 5 you expect her to be competent to handle a 6 risk-controlled arbitrage such as -- make it 7 easier for you -- such as United MBS? 8 A. I don't think anyone could have handled 9 that risk-controlled arbitrage if by "handled" you 10 mean managed it in a safe and sound manner with 11 regard to interest rate, prepayment, volatility 12 risk given the little capital that USAT had. 13 Q. Okay. Now we're getting to the nub of 14 it. No one could have managed that portfolio? 15 A. We testified this morning that the 16 institution had essentially zero capital and we're 17 talking about assets on the order of 5 to 18 $7 billion which are subject to the kinds of risks 19 that I just mentioned. So, I don't think any 20 amount of management of that particular portfolio 21 could have kept this institution from being in a 22 very risky situation. 3175 1 Q. And your preferred solution was to 2 liquidate? 3 A. Well, that was the first one that came 4 to mind, yes. 5 Q. Go out of business? 6 A. Yes. 7 Q. Right? 8 A. That would be the most prudent thing to 9 do. 10 Q. Fire the employees. Right? 11 A. Well, if that involved -- if the 12 portfolio had to be liquidated and at the same 13 time the institution had to be dissolved, it's 14 potentially the case that the institution could be 15 reconstituted under new organizational -- under 16 new organization and employees might not need to 17 be fired. 18 Q. To hell with the depositors? 19 A. I didn't say to hell with the 20 depositors. 21 Q. Well, depositors are FSLIC. Right? 22 A. No. Depositors and FSLIC are two 3176 1 different things. 2 Q. Well, one or the other of them are 3 going to take a loss if you take liquidation and 4 go out of business. Right? 5 A. Pay me now or take the risks of paying 6 me more later. 7 Q. So, your solution that you're offering 8 to the Court is that they should have gone out of 9 business? Have we got that straight? 10 A. With regard to this MBS portfolio, yes. 11 If that was -- if I was asked what I would do in 12 that situation, I would have reduced the 13 portfolio. 14 Q. Now, you talked about the comparative 15 disadvantages of an institution such as USAT in 16 buying MBS from broker/dealers. Right? 17 A. Yes, I did discuss that. 18 Q. Now, you indicated that USAT was an 19 institution of $6 billion or so and this may -- 20 how much -- how big do you have to be to do 21 business with these broker/dealers? 22 A. Well, if you're providing the 3177 1 mortgage-backed securities that you're purchasing 2 as collateral, you don't have to be that big. 3 Q. Well, I mean, without having a 4 competitive disadvantage? 5 A. Well, it's not a question of size. 6 Q. Well, how is it that anybody buys 7 anything from these broker/dealers? I mean, isn't 8 everybody according to your lights at a 9 competitive disadvantage? 10 A. Yes. Everybody -- well, I want to 11 qualify that. Certainly everyone in a similar 12 situation to USAT. 13 Q. Well, I have this misunderstanding 14 perhaps that this MBS was a large market, 15 something like a trillion dollars? 16 A. It's a very large market. 17 Q. And it's traded out there with 18 thousands of people in it, the market? 19 A. That's correct. 20 Q. And isn't it the marketplace that 21 determines the price? 22 A. Well, no. People put bids and offers 3178 1 on prices. 2 Q. Well, isn't the solution to your 3 problem is that you do a little comparative 4 shopping? Couldn't that just sort of even out 5 these disadvantages? 6 A. Yes, it could. 7 Q. You could call up two or three of the 8 broker/dealers and say, "What's your price on 9 Ginnie Mae 12s today?" 10 A. That's correct. 11 Q. And do you think you could rely on the 12 fact that after you had called five or six of 13 them, that you had gotten a good price? 14 A. You could rely on the fact that if you 15 did your work right and if the price quotes were 16 all fairly similar, that you got the best price 17 you could get. 18 Q. Well, how is that situation with this 19 market any different from the stock market? If I 20 want to buy stock and I go to a broker/dealer, is 21 it likely that the broker/dealer's going to know 22 more about the company than I do? 3179 1 A. Yes. 2 Q. And so, does that mean I can't buy? 3 A. No. 4 Q. It just means that I'm taking a risk? 5 A. It means that you're trading at a 6 competitive disadvantage and that you, on average, 7 will suffer a transactions cost associated with 8 that disadvantage. That's not to say that you 9 shouldn't make the investment. If you're 10 investing for the long term -- that is, to capture 11 eventual expected gains in a position, then that 12 cost of trading is something that you take into 13 account. That's different than speculative 14 trading where you're going back in and out of the 15 market attempting to identify -- with trading 16 partners that are much more sophisticated than you 17 are, which securities are underpriced relative to 18 the market and which securities are overpriced 19 relative to the market. 20 Q. I have a spot in my outline to discuss 21 that with you and if you can just hold off right 22 before we get there, I'll ask you some questions 3180 1 about an exhibit. It's at Tab 236. 2 MR. NICKENS: It's identified, Your 3 Honor, as B2684 and it's the top 100 thrifts in 4 mortgage-backed securities holdings that is we 5 discussed yesterday. This has already been in 6 evidence, Dr. Duffie. 7 Q. (BY MR. NICKENS) Can you identify any 8 one of these hundred largest thrift institutions 9 that wasn't at a competitive disadvantage with 10 broker/dealers? 11 A. No. 12 Q. You said that you were using the stock 13 market capitalization in order to measure the 14 capital of the institution. Right? 15 A. Yes. 16 Q. Now, do you know of any institution -- 17 A. Excuse me. I want to continue my 18 answer by saying that I took that as an upper 19 bound in the sense that the stock market 20 capitalization is including the market value of 21 the opportunity to take upside gains but not to 22 suffer from downside losses. So, that would be 3181 1 more than the actual capital available in the 2 institution in the market's estimate. 3 Q. Well, do you know of any institution of 4 this nature that would use that method of looking 5 at their capital? I mean, themselves? 6 A. Certainly. 7 Q. For what purpose? 8 A. Institutions look at their own share 9 price all the time to determine how the market 10 values their firm. 11 Q. But for accounting purposes and all 12 these other -- for regulatory purposes, these 13 regulated institutions have specific rules that 14 they have to adhere to in measuring their capital. 15 Right? 16 A. In measuring their accounting capital, 17 yes. 18 Q. And it does not include the market 19 price of their -- the stock of their company or 20 the holding company or whatever? 21 A. No. The regulatory capital 22 requirements don't refer specifically to the stock 3182 1 market valuation. 2 Q. And you're suggesting that this 3 methodology is superior to that required by the 4 regulations? 5 A. For purposes of measuring what capital 6 was available in the institution as a buffer 7 against losses, yes, definitely. 8 Q. Why is it that the people that are in 9 the business of making these decisions haven't 10 gone to stock market capitalization as a 11 measure -- such a measure? 12 A. Well, first off, not all institutions 13 have publicly-traded shares. 14 Q. That would be a problem, wouldn't it? 15 A. That would be a problem. Secondly, 16 some institutions have -- are owned by 17 publicly-traded institutions which have 18 substantial additional assets. So, it would be 19 difficult to split out what portions of the assets 20 were associated with the thrift institution 21 institution itself. 22 Q. And in addition to that, you'd have the 3183 1 problem of just market fluctuations. Right? 2 A. I'm not sure what problem you're 3 referring to. 4 Q. Well, the stock market price can go up 5 and down for all kinds of reasons having nothing 6 to do with the health of the institution? 7 A. Why would you say that? 8 Q. Well, somebody might be interested in 9 buying it for any number of purposes. Right? 10 A. Well, why would they be interested in 11 paying more than what it was worth? 12 Q. Well, I'm not suggesting that, 13 Mr. Duffie. I'm just saying that if you use stock 14 market capitalization, you'd have all kinds of 15 extraneous factors affecting the capital of the 16 institution, wouldn't you? 17 A. Well, it's rather the other way around. 18 The capital of the institution is reflected in the 19 stock price. 20 Q. Well, let's take this particular 21 institution. You've looked at the stock market 22 price, haven't you? 3184 1 A. Yes, I have. 2 Q. It went up and down substantially, 3 didn't it? 4 A. During what period? 5 Q. Let's take '85, '86. 6 A. If you take that period as a whole, it 7 went down substantially. 8 Q. Well, okay. Went down substantially. 9 It changed. Right? 10 A. Certainly did. 11 Q. And so, the people running the 12 institution wake up one day and the stock 13 market -- their stock market price has gone down 14 and they are insolvent. How do they plan? 15 A. How do they plan for -- 16 Q. How do they make the plan -- how do 17 they plan the activities of the institution if 18 their capital is dependent upon their stock market 19 price? 20 A. I mentioned before that it's rather the 21 other way around. The stock market price reflects 22 all of the information available to the stock 3185 1 market including assets, liabilities, any other 2 information that's provided to the public. 3 MR. NICKENS: Your Honor, if it's 4 convenient, this would be a good time for us to 5 take a break. 6 THE COURT: All right. We'll take a 7 short recess. 8 9 (A break was taken at 2:33 p.m.) 10 11 THE COURT: Be seated, please. We'll 12 be back on the record. Mr. Nickens, you may 13 continue. 14 MR. NICKENS: Thank you, Your Honor. 15 (2:56 p.m.) 16 Q. (BY MR. NICKENS) Dr. Duffie, in 17 reviewing my notes, I over -- I notice that I 18 failed to ask you a question that I almost 19 invariably ask in this situation. 20 Would you tell the Court what you're 21 charging for the services that you're providing to 22 the OTS in this matter? 3186 1 A. Yes, sir. I am charging $500 per hour. 2 Q. Now, let's talk a little bit about this 3 OAS model and this issue about what the status of 4 it was. I'm going to hand you a book. It's not 5 the one we were looking for, but I think it has 6 the reference that we were looking for. And 7 identify for the record what the book is and what 8 page that I've referenced for you. 9 A. This is a book titled "Mortgage and 10 mortgage-backed Securities Markets" by 11 Frank Fabozzi and Franco Modigliani. 12 Q. Now, Mr. Modigliani is a Nobel prize 13 winning economist? 14 A. That's correct. 15 Q. And Mr. Fabozzi is one of the most 16 widely published in the United States on 17 mortgage-backed and fixed income securities, is he 18 not? 19 A. If by "widely published" you mean the 20 number of articles or books published, I couldn't 21 say he's most widely published, but he's certainly 22 very widely published. 3187 1 Q. Now, what does he indicate and identify 2 on the page for us with regard to the OAS 3 modeling? 4 A. I'm sorry. The question is? 5 Q. There is a footnote there that I 6 thought that was what we were looking for with 7 that little -- 8 A. You tagged it? 9 Q. Yes. Isn't that the one you had 10 reference to earlier about your research? 11 A. I don't believe that I referred in my 12 testimony this morning to this particular 13 footnote, but it does refer to the article in an 14 earlier publication that I referred to this 15 morning, yes. 16 Q. So, why don't you read the footnote in 17 the record so that we can get this behind us? 18 A. Okay. It says -- Footnote No. 5 on 19 Page 280. And I quote, "In a critical review of 20 this methodology, Finnerty and Rose (John D. 21 Finnerty and Michael Rose, arbitrage-free spread: 22 A consistent measure of relative value, Journal of 3188 1 Portfolio Management, [Spring 1991], Pages 65 to 2 81) state that the earliest published work 3 involving this methodology appeared in Michael 4 Waldman and Stephen Modzelewski, quote, 'A 5 Framework for Evaluating Treasury-based Adjustable 6 Rate Mortgages,' unquote, in Frank J. Fabozzi 7 edited The Handbook of Mortgage-backed Securities 8 (Chicago: Probus Publishing, 1985), Pages 307 to 9 346." 10 Q. Now, what you have said in your report 11 among other places but at Paragraph 59 is that 12 since before the period in question, sophisticated 13 MBS traders such as Wall Street trading houses 14 have used option adjusted spread models to predict 15 the cash flows from their MBS investments. Right? 16 A. That's correct. 17 Q. Now, what is the period in question? 18 A. Well, 1985 to 1988, let's say. 19 Q. Well, that would mean that since before 20 1985, you are stating that sophisticated MBS 21 traders have used option adjusted spread models to 22 predict the cash flows? 3189 1 A. That's the opinion that I stated. 2 Q. Since before 1985? 3 A. That's correct. 4 Q. And what Mr. Fabozzi and 5 Mr. Modigliana's book says is that the earliest 6 published work involving this methodology appeared 7 in a book published in 1985? 8 A. That's correct. 9 Q. So, how do we push it back to before 10 1985? 11 A. Well, those facts are consistent -- 12 that is, an article published in 1985 was, first 13 of all, written presumably before 1985. It takes 14 a while to publish something. And secondly, is 15 the first published version of this modeling 16 strategy. One doesn't infer from the publication 17 date that the modeling strategy was not available 18 before 1985. 19 Q. Well, this isn't the only information 20 that we've got on this subject, is it? 21 A. No, it's not. 22 Q. And what it says is that the earliest 3190 1 published work involving this methodology -- now, 2 have you read this article -- pardon me -- by 3 Waldman and -- I can't pronounce the name. 4 A. Modzelewski. 5 Q. Yes, sir. 6 A. I reviewed the article. 7 Q. Did it indicate that the model was in 8 use anywhere? 9 A. Yes. It indicated it was in use at 10 Salomon. 11 Q. Anywhere else? 12 A. The article did not refer to its use 13 anywhere else. 14 Q. And what kind of use? 15 A. To be quite frank, I don't remember how 16 it said it was being used at Salomon. 17 Q. Now, in its early stages -- ever? 18 A. I should amend my answer. It did say 19 it was being used to value adjustable rate 20 mortgages. 21 Q. But not for trading purposes, did it? 22 A. I don't recall whether the article 3191 1 stated that it was being used for trading purposes 2 or not. 3 Q. In its early stages, these models took 4 about 30 minutes to a hour to push out a result? 5 A. I don't want to testify to the amount 6 of time, but it did in its very early incarnations 7 take more than a few seconds to compute the prices 8 of mortgage -- prepayable sensitive securities. 9 Q. And that would make it very difficult 10 to use for the traders' purposes, wouldn't it? 11 A. No. 12 Q. Have you ever talked to a trader that 13 indicated that this tool was in use prior to 1985? 14 A. Well, I talked to a trader who said it 15 was in use by 1985. 16 Q. Who was that? 17 A. Dexter Senft. 18 Q. Let me ask you to look at a document 19 that we have identified as B586, and I believe 20 it's been introduced if I can get the tab number. 21 Now, to use an OAS model for 22 mortgage-backed securities, you also have to have 3192 1 a prepayment model. Right? 2 A. That's one of the ingredients. 3 Q. And the truth is that the OAS model is 4 only as good as that prepayment model. Right? 5 A. Well, a necessary condition for the 6 model to work well is that the prepayment model 7 part of it does work well. 8 Q. This article is by Mr. Waldman who is 9 one of the ones that you indicated and it's dated 10 September 4th, 1985, and it indicates in the third 11 paragraph "In order to create a logical framework 12 for valuing mortgage issues, Salomon Brothers has 13 developed a statistical model of prepayment rates 14 based upon historical prepayment experience." 15 Now, this is Salomon Brothers 16 announcing that they have a prepayment model. 17 Right? 18 A. I just read it for what it says. It 19 has developed a statistical model of prepayment 20 rates. 21 Q. And you couldn't possibly have a 22 workable OAS model before you had a prepayment 3193 1 model? 2 A. I think we agreed on that a few minutes 3 ago. 4 Q. And it says above -- in the sentence 5 just above -- this is September 1985 at Salomon 6 Brothers. "It has recently been common practice 7 for market participants to evaluate yields and 8 yield spreads based upon the past 12 months' 9 prepayment experience." 10 Do you see that? 11 A. Yes, I do. 12 Q. Now, that's certainly not an OAS model, 13 is it? 14 A. Just on its face, it doesn't look like 15 it's based on OAS modeling. 16 Q. Then if we turn over -- just a second. 17 If you turn over to the second page under the 18 Salomon Brothers prepayment model, do you see -- I 19 guess it's the second sentence. "The parameters 20 of the model have been estimated by regression 21 techniques using data on mortgage prepayments for 22 the 1979 to '85 period." 3194 1 So, their prepayment model is using 2 historic prepayments? 3 A. Yes. 4 Q. And then if you go over to Page 5, 5 under high coupon issues, if you look at the third 6 sentence, it says "First, the prepayment estimates 7 for the high coupon category carry the greatest 8 margin of error." 9 And that's informing people that they 10 should take some care in using the prepayment 11 models for high coupon issues. Right? 12 A. It doesn't say specifically how one 13 should take care, but that's one of the 14 conclusions that I would draw from that. It 15 doesn't say specifically that. 16 Q. Now, how can one have a workable OAS 17 model before having a prepayment model? 18 A. Well, I believe we agreed that one 19 needs a prepayment model in order to have an OAS 20 model. 21 Q. Now, let me ask you to look at B1619 22 once I can find it for you. This also is in 3195 1 evidence as Tab 237 and has been identified as a 2 working paper by Mr. Hjerpe dated May 11th, 1987, 3 published by the Federal Home Loan Bank Board, 4 their office of policy and economic research. 5 Let me ask you to look over to Page 6 nine. And if you can read along with me at the 7 bottom of the page, "The reason for less than 8 perfect efficiency in the marketplace for pricing 9 MBS may be the result" -- 10 A. I'm sorry. I'm not quite there yet. 11 Q. Okay. 12 A. Bottom of Page 9? 13 Q. Yes, sir. 14 A. Okay. Thank you. 15 Q. And by the way, up above, you'll see 16 that Mr. Hjerpe is saying that the buyers of this 17 MBS may believe that they have a comparative 18 advantage in pricing MBS and able to recognize 19 underpriced securities in the market. Right? 20 A. I can read the sentence. It says 21 "First, the institution may believe it has a 22 comparative advantage in pricing MBS and thus be 3196 1 able to recognize underpriced securities in the 2 market." 3 Q. So, in comparison to your evidence 4 about the comparative disadvantage in the market, 5 Mr. Hjerpe is indicating in 1987 that the 6 institutions thought they had the upper hand? 7 A. He indicates that he thinks that they 8 thought that, yes. 9 Q. Okay. Then if we go down here to the 10 last paragraph, "The reason for less than perfect 11 efficiency in the marketplace for pricing MBS may 12 be the result of a lack of consensus as to the 13 appropriate theoretical and practical models for 14 pricing these securities. Jaocb and Toves, 1987, 15 have recently analyzed the various approaches 16 currently being used to price MBS in the 17 marketplace. They conclude that even though 18 several different approaches are currently being 19 used by investors, a market consensus will develop 20 in the next year the options based modeling for 21 pricing MBS or the most appropriate." And the 22 authors argue that the current static model would 3197 1 depend on duration and convexity will become 2 obsolete as the marketplace realizes the 3 disadvantages of such models. 4 So, as late as 1987, there is still a 5 debate going on as to whether the OAS model is the 6 appropriate model, correct? 7 A. Depends on what you mean. There was -- 8 according to Mr. Hjerpe, there was discussion 9 among investors about different approaches. And 10 according to him, consensus would develop around 11 the OAS technology. 12 Q. Well, actually, he's referring to Jacob 13 and Tolls' article, isn't he? 14 A. Pardon me. He's summarizing their 15 assumptions with regard to that, yes. 16 Q. Yes, sir. And then if I can ask you to 17 look over at Page 18 on the same subject in the 18 middle paragraph -- are you with me? 19 A. The middle paragraph? 20 Q. Yes, sir. 21 A. Yes. 22 Q. "Several different schools of thought 3198 1 exist as to the most appropriate method of 2 estimating prepayments. Some investors argue that 3 the most recent historical prepayment experience 4 of the particular mortgage pool is the best 5 indicator of estimated future prepayments on MBS. 6 Others argue that a national average for mortgages 7 of similar coupons is a good estimate of expected 8 prepayments. Still others claim that an options 9 based pricing approach that uses an average of all 10 possible outcomes is the best estimator. There is 11 no consensus as to the best method of prepayment 12 predictions." 13 Now, do you have any information to 14 disagree with Mr. Hjerpe about the condition in 15 May of 1987? 16 A. I'm not sure what group that he's 17 drawing a consensus from, but I'm quite confident 18 that if one goes to the broker/dealers on Wall 19 Street that are dealing with mortgage-backed 20 securities in 1987, there's very broad consensus 21 around the fact that failure to use an option 22 adjusted spread model would not be a satisfactory 3199 1 way to account for prepayment risk. 2 Q. Would you include First Boston among 3 those group of knowledgeable investment bankers? 4 A. When drawing a consensus, yes. I don't 5 know their particular opinions. 6 Q. Are you aware of the fact that 7 Mr. Carron, who is here in the courtroom, was with 8 First Boston in 1986 in their analytics 9 department? 10 A. I know that he was with First Boston. 11 I can't state specifically which dates. 12 Q. And so, he would likely have personal 13 knowledge as to when OAS models became used on 14 Wall Street? 15 A. Well, if he was at First Boston and he 16 was actively involved in mortgage-backed 17 securities trading at that time, I would expect 18 that he would know when OAS was available. 19 Q. Right. 20 A. Or at least within -- 21 Q. Do you know who he was working with at 22 First Boston? 3200 1 A. No. 2 Q. Would the name Dexter Senft come to 3 mind? 4 A. As a matter of fact, I did hear that he 5 was working with Dexter but I didn't realize it 6 was at First Boston. 7 Q. All right. And that's the person 8 you're relying upon for your conclusion as to when 9 it became used. Right? 10 A. I'm not relying upon it. I polled 11 Dexter on his opinion on that. 12 Q. If you would, although this is a 13 slightly different point, if we can look over at 14 Page 30, the conclusion of Mr. Hjerpe's article. 15 I'm going to just ask you whether you agree or 16 disagree with the statements here made by Mr. 17 Hjerpe in his conclusion. 18 "Though RCA is a complex series of 19 financial transactions, it does not include any 20 transactions that are new to the thrift industry. 21 The packaging of all the transactions into a 22 highly leveraged growth strategy is the only new 3201 1 component. At present, it appears that a 2 reasonable spread can be attained by proper 3 implementation of such a strategy. However, these 4 opportunities have diminished and will probably 5 continue to diminish in the future to a point 6 where the costs meet or exceed the return. The 7 concept of an RCA strategy can, in fact, be quite 8 appropriate for a thrift. It includes active 9 participation in the mortgage markets, putting 10 long-term assets on the books of an institution, 11 and attempting to fund those assets with lower 12 cost liabilities. This is the typical mission of 13 a thrift. In addition, RCA is a relatively low 14 cost and rapid method for inflating the balance 15 sheet of an organization. The strategy can help 16 problem institutions generate income with little 17 cost that may help to offset negative earnings of 18 other assets and, in turn, improve overall return 19 on assets and equity for these institutions." 20 Now, do you agree or disagree with the 21 sentiments expressed by Mr. Hjerpe in May of 1987? 22 A. I disagree with -- most definitely with 3202 1 the last sentence saying "the strategy can help 2 problem institutions generate income with little 3 cost that may help to offset negative earnings on 4 other assets and, in turn, improve overall return 5 on assets and equity for these institutions." 6 Q. Okay. I'm going to hand you a 7 document, Dr. Duffie, that has been identified as 8 Exhibit B2802 and which on its face appears to be 9 a capital markets release from the Office of 10 Thrift Supervision Release No. 5, date August 3, 11 1989. And it is entitled "Option Adjusted Spreads 12 and Duration Valuable Tools for Portfolio 13 Management." 14 MR. NICKENS: We offer Exhibit 2802, 15 Your Honor. 16 MR. GUIDO: No objection, Your Honor. 17 THE COURT: Received. 18 Q. (BY MR. NICKENS) Now, you've seen 19 this document before, have you not, Dr. Duffie? 20 A. If I've seen it, it's been long enough 21 so that I actually would need to review it in 22 order to refresh my memory on the document. 3203 1 Q. Well, let's just focus on a couple of 2 points if we might. On the very first sentence, 3 the introduction, this is August of 1989. It says 4 "Option based models have recently become the 5 preferred tool for analyzing the relative value 6 and duration of mortgage-backed securities." 7 A. That's correct. 8 Q. Now, how far back would you think that 9 term "recently" should be extended? 10 A. A few moments ago, you quoted to me a 11 1987 article that referred to the recent advent of 12 these models. So, going from '89 back to '87, 13 taking another step backwards from 1987, let's say 14 around 1985. 15 Q. Well, actually, what it said was in 16 1987 that Mr. Toves and the other gentleman in his 17 article were predicting that a consensus was 18 developed later. Right? 19 A. It also referred to his view that in 20 1987, these models had recently come into play. 21 Or perhaps that was one of the other articles that 22 you asked me to review. 3204 1 Q. What this is reporting is that, in 2 fact, it has become the preferred tool by 1989? 3 A. That's what it says. Doesn't say what 4 the tool among what community. 5 Q. Well, this is -- where do you think 6 these capital markets release go to? 7 A. Since it's the Office of Thrift 8 Supervision, I suppose the people that would be 9 concerned with thrifts, which is a different 10 community than the one that I was painting as 11 having more sophistication -- that is, the 12 broker/dealer community on Wall Street. I 13 wouldn't take issue with the claim -- I don't know 14 one way or the other -- that between 1987 and '89, 15 the use of option adjusted spread models was 16 becoming more common in the thrift industry. 17 Q. Now, if you'd look over at the bottom 18 of the page, it says "In the past, static cash 19 flow yield analysis was a widely used method to 20 price and compare the relative value of different 21 securities. The more" -- 22 A. I'm sorry. I didn't catch where you 3205 1 were just starting. 2 Q. It's the next-to-the-last paragraph 3 right before the carry-over paragraph. 4 A. Okay. I'm with you now. Thanks. 5 Q. "In the past, static cash flow yield 6 analysis was a widely used method to price and 7 compare the real value of different securities. 8 The more sophisticated variance of this method 9 incorporated prepayments into the model by 10 projecting a path of prepayments into the future." 11 You see that? 12 A. Yes, I do. 13 Q. Now, those are the same static models 14 that you've been critical of. Right? 15 A. I was critical of them with regard to 16 valuation and hedging of mortgage-backed 17 securities, yes. 18 Q. Well, and this was talking about 19 used -- "was a widely used method to price and 20 compare the relative value of different 21 securities." That's the very thing you were 22 talking about. Right? 3206 1 A. I was not talking about wide use of it. 2 I was critical about the use of such a model when 3 attempting to value trade or speculatively trade 4 or identify under and overpriced assets when one's 5 trading partners have superior technology. 6 Q. Well, isn't that comparing the relative 7 value? 8 A. Yes, absolutely. 9 Q. Do you see any criticism in here of 10 that methodology? 11 A. I'll have to read it. 12 Q. Well, let me withdraw the question. If 13 you can go to the next page, you'll see the first 14 complete sentence there on the top, "Analysts use 15 these cash flow yields to compare the values of 16 securities with similar characteristics." 17 A. I'm sorry. I'm just not quite keeping 18 up with you. 19 Q. It's at the top of the Page 2 and I 20 started -- there is a partial sentence there and I 21 started the first full sentence. 22 A. I'm with you now. 3207 1 Q. "Analysts use these cash flow yields to 2 compare the values of securities with similar 3 characteristics. In addition, they would use 4 historical spreads of mortgage cash flow yields 5 over similar treasuries to determine which 6 securities were currently rich, cheap, or fairly 7 valued." 8 Now, is that what you've told the Court 9 could not be done without an option adjusted 10 spread model? 11 A. I think I was pretty clear that any 12 model that attempts to value securities with 13 prepayment risk in them is not going to be able to 14 do it in a fully accurate manner because it's 15 impossible to completely predict what prepayments 16 will be in any scenario. Rather, what I said is 17 that these types of static valuation models are at 18 a distinct disadvantage relative to models that 19 have the ability to simulate changes in prepayment 20 rates as interest rates are changing. 21 Q. Let me ask you to go over to Page 6. 22 And at the bottom of that -- of the first 3208 1 paragraph, carry-over paragraph, it starts "For a 2 thrift manager who has a portfolio exposed to 3 rising rates -- that is, the duration of assets 4 exceeds that of liabilities -- a sale of the rich 5 10s to purchase the cheap 12s would not only 6 enhance return but reduce the portfolio's interest 7 rate risk, as well. For a pension fund manager 8 with a portfolio exposed to falling rates, the 9 duration of assets is less than that of 10 liabilities. The sale of the rich 10s to purchase 11 the cheap 8s would again not only enhance return 12 but reduce interest rate risk, as well." 13 Now, is that what you're describing as 14 speculative trading? 15 A. Well, it's quite clear that this is 16 writing about situations in which one is 17 attempting to do two things at the same time. 18 That is, to identify over and underpriced 19 securities and, at the same time, to restructure a 20 portfolio so as to protect it more from interest 21 rate risk. So, I certainly did not suggest that 22 one cannot restructure a mortgage portfolio in 3209 1 order to improve its interest rate risk exposure. 2 In fact, just the opposite. This morning in my 3 testimony, I suggested that one should restructure 4 a portfolio in order to reduce its exposure to 5 interest rate risk. However, this refers as well 6 to identification of rich and cheap securities and 7 certainly, one can attempt to do so with the best 8 techniques that one has available. But again, 9 when one is conducting a program such as conducted 10 at USAT in 1987 and '88 in its United MBS 11 portfolio of repeated trading in and out of the 12 portfolio attempting to identify over and 13 underpriced securities when trading with 14 sophisticated Wall Street counter-parties, I would 15 disagree that this is going to be an effective 16 strategy and, as I suggested, I consider it a 17 risky and loss-generating strategy. 18 Q. In fact, a lot of people did. Right? 19 A. Did what USAT did? 20 Q. Yes, sir. 21 A. I was just looking at the volumes of 22 trade in that portfolio. I'm not familiar with 3210 1 institutions that had such a significant degree of 2 replacing securities within such a short period of 3 time. There was a large amount of activity. 4 Certainly securities were purchased by other 5 institutions using these kinds of static valuation 6 models. 7 Q. Are you familiar with the publication 8 Institutional Investor? 9 A. Yes, I am. 10 Q. Let me show you an article that is in 11 evidence, Dr. Duffie, entitled "Cashing In on 12 Risk-controlled Arbitrage" published, the record 13 will show, in February 1986. And I would draw 14 your attention to the third page of B823. In the 15 right-hand column a little below halfway down 16 where it says "most thrifts" -- 17 MR. GUIDO: Your Honor, if this 18 question is a question that assumes the accuracy 19 of the statement for purposes of 20 cross-examination, I object to the use of the 21 document for that purpose. 22 MR. NICKENS: Your Honor, the 3211 1 document -- 2 MR. GUIDO: I did not agree -- I 3 objected to the use of the document. I did not 4 object to the use of the document to ascertain 5 what it is the portfolio managers at USAT knew. I 6 objected, Your Honor, to the introduction of the 7 document for the truth of the matters asserted 8 therein. This question is premised upon the truth 9 of the matter asserted in the article and for that 10 reason, I object to the question. 11 MR. NICKENS: Your Honor, he hasn't 12 heard the question yet and this document is in 13 evidence. 14 MR. GUIDO: I heard the question twice 15 already. You've just gone to the document to read 16 the document. 17 THE COURT: Well, let's have the 18 question. 19 MR. NICKENS: My question, Your Honor, 20 is whether he agrees or disagrees with the 21 statement or claim made in this exhibit about 22 trading. 3212 1 THE COURT: There's nothing wrong with 2 that. Can you answer the question? 3 Q. (BY MR. NICKENS) Let's make sure what 4 we're talking about. I'll read it, if that's 5 okay. "Most thrifts that pursue risk-controlled 6 arbitrage also actively trade pass-throughs, 7 buying those they think are undervalued and 8 selling the ones they think are too dear to add 9 incremental returns. They view this as another 10 means of obtaining insurance against prepayment 11 risk. They constantly sift through the 12 particularities of different pass-through issues 13 and track the pattern of supply and demand in the 14 market much the same way active managers of fixed 15 income portfolios do to boost returns." 16 It goes on to say in a quote, "It's not 17 difficult to turn over the entire portfolio just 18 in the process of taking advantage of added value 19 available in the marketplace to increase the 20 spread." It says "William Siemens, chairman of 21 Overland Park Savings and Loan, located about 22 40 miles south of Kansas City." 3213 1 Now, do you agree or disagree that that 2 was what people were doing back in 1986? 3 A. Well, there are several assertions here 4 so I'd prefer to take them one at a time. The 5 first -- 6 Q. Okay. But we're talking about a 7 historical fact. Do you have any historical 8 information to indicate that this is not correct? 9 A. Well, again, there are several facts 10 that are asserted here, and I'd prefer to do them 11 one at a time rather than treating them as one. 12 MR. GUIDO: Your Honor, I object to 13 that last question because he's now changed the 14 question. If the question to this witness is: Do 15 you agree with the particular facts, that's 16 different than the question: Do you have anything 17 to contradict the fact? 18 MR. NICKENS: That's a fair 19 observation, and I will withdraw the question and 20 return to the one I had before. 21 THE COURT: All right. Let's let the 22 witness answer the question. 3214 1 THE WITNESS: Thank you, Your Honor. 2 A. So, the first assertion is that most 3 thrifts that pursue risk-controlled arbitrage also 4 actively trade pass-throughs, buying those they 5 think are undervalued and selling the ones they 6 think are too dear to add incremental returns. 7 And, frankly, I don't have enough information to 8 either contradict or to agree with this assertion. 9 They view this as another means of obtaining 10 insurance against prepayment risk. And here, it's 11 not only an assertion of what's done but what the 12 views of others may be. I certainly disagree with 13 that view that he is asserting or she is 14 asserting -- pardon me -- she is asserting that 15 may be held by others. But as to whether she 16 believes those views are held or not, I haven't 17 got any information to contradict that. 18 "They constantly sift through the 19 PHRAEURTS of different pass-through issues and 20 track the pattern of supply and demand in the 21 market much the same way that active managers of 22 fixed income portfolios do to boost returns." 3215 1 Again, I don't have any information 2 that would contradict that assertion or be 3 consistent with it one way or the other. And then 4 the next statement is a quote. So, I won't 5 dispute the fact that Mr. Sieman said this. And 6 then I guess that was it. 7 Q. Okay. Moving on to another subject, 8 Dr. Duffie, you mentioned in your direct testimony 9 something about IOs and POs and those were held in 10 United MBS portfolio, correct? 11 A. That's correct. 12 Q. And that was a portfolio managed by Ms. 13 Sandy Lorenson, correct? 14 A. Well, it was listed under United MBS 15 and my understanding is that she was charged with 16 that portfolio, yes. 17 Q. And IOs and POs can and are frequently 18 used as hedging instruments; isn't that correct? 19 A. They can be under certain 20 circumstances. 21 Q. And you don't know when they were 22 bought. Right? 3216 1 A. No. I don't have the trade tickets on 2 these securities. 3 Q. Well, I'm not asking for a specific 4 date, but you don't know approximately even when 5 they were bought? 6 A. No. That's not true. I'm just going 7 to refer to the table. I'm referring to Table 8.1 8 in my -- in the exhibit from my report this 9 morning. According to this, since there is no 10 record of any embedded losses or interest rate 11 sensitivity prior to April 1987, just as I this 12 morning was relying on the first records of the -- 13 of these investment committee minutes as an 14 indication of the beginning of this portfolio, I 15 would similarly assume that these IOs and POs 16 weren't added to the portfolio until at least 17 early 1987. 18 Q. Well, do you know for what purpose they 19 were bought? 20 A. Well, they certainly were not acting as 21 hedges -- 22 Q. Sir -- 3217 1 A. -- in the manner that you suggested 2 since they, in fact, add to the interest rate risk 3 of the institution. 4 Q. Well, how do you know they add to the 5 interest rate sensitivity of the institution? 6 There may have been other things going on at the 7 same time. Right? 8 A. Yes. But we do know that the 9 institution was exposed to -- 10 THE COURT: Finish your answer. 11 A. Yes. We do know that the institution 12 was exposed to a general rise in the term 13 structure of interest rates and we also know that 14 these IOs and POs, according to USAT's 15 information, were substantially exposed to the 16 extent of $20 million at times to the same kind of 17 rise in the term structure of interest rates. 18 Q. (BY MR. NICKENS) In the overall 19 scheme of the portfolio, these are a fairly small 20 amount of holdings. Right? 21 A. Well, I wouldn't classify a loss of 22 $25 million that was already in the portfolio plus 3218 1 another 25 million as small relative to USAT's 2 capital which during this period was, let's say, 3 between zero and $10 million. 4 Q. Is it in the nature of IOs and POs that 5 they tend to move in opposite directions in terms 6 of their price or value with the movement of 7 interest rates? 8 A. Depends on the IOs and POs and what 9 amounts you held of each. According to USAT's 10 information, they were moving in the same 11 direction in their interest rate sensitivity as 12 mortgage-backed securities in total. 13 Q. Now, did you look at the investment 14 committee minutes to determine the -- when they 15 were bought and for what purpose? 16 A. I certainly went through the investment 17 committee minutes. I can't say that I recall 18 specifically what the strategy was in buying IOs 19 and POs. 20 Q. Let me ask you a few questions about 21 your testimony that the portfolio, meaning the 22 entire institution, was positioned speculatively. 3219 1 Do you recall your testimony earlier 2 about that to the effect that it was positioned in 3 such a way that there would be a gain if interest 4 rates go down and a loss if interest rates went 5 up? 6 A. During a significant period of time to 7 a significant extent, yes. 8 Q. And you termed that position as being 9 speculative? 10 A. Yes, it is. 11 Q. And I gather, then, that it would also 12 be speculative if they were positioned in such a 13 way that if interest rates went down -- went up, 14 rather, there would be gains and if they went 15 down, there would be losses? 16 A. Well, let me premise this by saying any 17 portfolio is going to suffer some loss or gain 18 whether interest rates go up or down. There will 19 always be some loss or gain one way or the other. 20 However, in this case, there was a dramatic shift 21 about mid-1987 to a situation in which going from 22 losses and gains which were in general terms 3220 1 fairly similar going up or down to a stance in 2 which, for example, if I take August 5th, '87, 3 there was a gain of about $50 million of interest 4 rates went down by a hundred basis points but a 5 loss of about $80 million if interest rates went 6 up a hundred basis points. I would definitely 7 call that a speculative position with regard to 8 changes in interest rates. 9 Q. Well, and you would regard it if it 10 were the other way around as far as the movement 11 of interest rates would also be speculative. 12 Right? 13 A. Very much so. 14 Q. And so -- now, do you know of any 15 circumstances in which you can gain if interest 16 rates go either up or down? 17 A. Not unless you have superior 18 information to the market and not unless you're 19 talking about relatively trivial amounts. 20 Q. All right. So, if interest rates -- if 21 you're positioned in such a way to benefit from 22 rising rates, it's speculative. If you're in a 3221 1 position to benefit from lower rates, it's 2 speculative. It's not possible to benefit from 3 rates at the same time going up and down. Right? 4 A. Well -- 5 Q. Right? 6 A. I qualified that by saying in typical 7 situations, no, you can't profit no matter what 8 happens. 9 Q. And so, the only non-speculative 10 situation is one in which you lose going either 11 way? 12 A. I didn't say that. 13 Q. Well, what is the other circumstance? 14 I would have thought there was only four. 15 A. Well, first of all, speculation refers 16 to the manner of trading; that is, whether it's 17 designed for long-term investment, for example, or 18 designed to hedge the portfolio on the one hand 19 versus attempting to identify under and overpriced 20 securities with heavy volumes of trading as one 21 aspect of the speculation that happened. 22 The second one is when you have very, 3222 1 very little capital and you're taking very large 2 risks of this variety, that's definitely 3 speculative. 4 Q. Sir, isn't the case that the only 5 situation that you've identified as 6 non-speculative is the one in which you lose 7 whichever way interest rates move? 8 A. No. I didn't say that. 9 Q. Let me ask you some questions about 10 the -- about this value trading. 11 What kind of speculation is there in 12 moving from, let's say, Fannie Mae 9s to Freddie 13 Mac 9s? 14 A. Well, assuming there was no other good 15 reason to do it, one might be doing that in order 16 to judge that the market had mispriced one of 17 those securities relative to the other. 18 Q. And how can that possibly be 19 speculative? 20 A. Well, when one is trading with Goldman 21 Sachs, Morgan Stanley, and Salomon Brothers, and 22 one is doing this on a repeated basis -- that is, 3223 1 going in, coming out with very high volumes of 2 trade -- I would have a hard time with classifying 3 it as anything other than speculating. One is 4 speculating that one has somehow beaten the 5 market; that is, one is able to choose securities 6 in a manner that's better than one's 7 counter-parties. 8 Q. Again, you haven't talked to Mr. Smith 9 about this? 10 A. You're speaking of Mr. Terry Smith now? 11 Q. Yes, sir. 12 A. No, I haven't. 13 Q. So, you would say that trading Freddie 14 Mac 9s for Fannie Mae 9s and posit that you get an 15 increased yield from the trade is speculative? 16 A. Now, when you're referring to yields, 17 are you just counting the cash flows at projected 18 prepayments are you adjusting for the options in 19 those yields. 20 Q. Whatever you want to use. You have to 21 use a prepayment assumption, don't you? 22 A. You would have to use a prepayment 3224 1 assumption. If one used as USAT was doing static 2 prepayment assumptions, simply knowing that one of 3 those securities was estimated at projected 4 prepayments to have a higher yield than another 5 wouldn't mean, for example, that one had found an 6 under or overpriced security. 7 Q. Well, if you're using the best tools 8 you have and you found that situation, wouldn't it 9 be imprudent not to take advantage of it? 10 A. Depends on the circumstances. But in 11 this case, I would say it would have been 12 imprudent to attempt to take advantage of it if 13 one is doing it repeatedly and on a large scale 14 such as USAT. 15 Q. Well, let's try to focus in, if we can, 16 Dr. Duffie. Single situation. Wouldn't it be 17 imprudent if, using the best tools available, you 18 had identified such a situation not to take 19 advantage of it? 20 A. And you're asserting that these 21 securities other than the fact that one has one 22 agency label, the other has another agency label, 3225 1 their actual economic characteristics are 2 otherwise identical? 3 Q. Yes, sir. That's what we say when we 4 say Freddie Mac 9s and Fannie Mae 9s? 5 A. I would call that a speculative trade. 6 Q. Haven't you been published on the 7 subject of bargains in the marketplace? 8 A. I don't recall using that word 9 specifically with regard to speculation; but if 10 you want to refresh my memory about something I've 11 written... 12 Q. Well, about the existence because of 13 special-demand situations for -- I don't know 14 whether -- I don't recall your term -- but that 15 there are situations that occur in the market for 16 special value? 17 A. Yes. But unfortunately, in order to 18 get into those situations, one must pay an entry 19 fee which is the price of the security in 20 question. 21 Q. Now, did I hear you correctly to say 22 that you have not measured the losses that you 3226 1 claim occurred from value trades? 2 A. At United MBS? 3 Q. Yes, sir. 4 A. No. All I have available are the total 5 embedded losses in the portfolio. I didn't 6 attempt to split out what those losses would have 7 been had value trading not occurred. 8 Q. Well, now, if you haven't calculated 9 them, how do you know there were losses? 10 A. Well, I can see the losses. That is 11 referring to the total. I can see that there are 12 significant losses on the order of $100 million. 13 Q. How do you know that's attributable to 14 so-called value trades? 15 A. I don't. 16 Q. If you haven't -- 17 A. Again, I didn't assert what fraction of 18 these losses are associated with value trading. 19 What I did assert was that value trading was with 20 a speculative activity. 21 Q. In fact, they could have gained from 22 their value trades and had other losses that 3227 1 overwhelmed them. Right? 2 A. I guess anything is possible. 3 Q. Well, what would have been the result 4 of no trading? 5 A. No trading at all, holding the 6 portfolio, no hedging, no rebalancing associated 7 with changes. 8 Q. You've criticized value trading, and 9 I'm asking you -- and you said you haven't 10 calculated it. 11 A. I haven't calculated the fraction of 12 these losses that were associated with value 13 trading or speculative trading. 14 Q. If fact, you've told us that you're not 15 even able to tell us that there were losses. 16 Right? 17 A. That's correct. I've characterized the 18 trading as speculative. I haven't calculated the 19 losses. 20 Q. And it could have been gains? 21 A. It's possible. 22 Q. Now, what would have been the result, 3228 1 if you can tell us, of no trading, no -- none of 2 this speculative trading that you have identified? 3 A. Numerically, I can't tell you because 4 they didn't adopt a no trading stance. They 5 actually did what they did. 6 Q. Well, can you tell us whether it would 7 have been a gain or a loss from the situation that 8 resulted? 9 A. Not without doing the computations. 10 Q. Let me ask you some questions about the 11 roll-down. Have we agreed that that is a proper 12 strategy? 13 A. What's the set of strategies that 14 you're drawing from? 15 Q. That is that interest rates have moved 16 down to such a degree that the portfolio manager 17 has -- believes that he's now at an imbalance 18 situation on his durations. 19 A. Okay. And the set of choices when 20 one's describing a proper strategy, are you 21 including, for example, getting out of the 22 position altogether if its risk is too large 3229 1 relative to your capital? 2 Q. Well, you indicated there are several 3 strategies, but I thought that you had said that 4 the roll-down was at least an appropriate 5 response. I misinterpreted what you said? 6 A. I believe I qualified that by saying if 7 one is committed to hanging on to the swaps, then 8 one should roll down the mortgages as one goes in 9 order to mitigate the interest rate risk exposure. 10 Q. Let me ask you to look at an article by 11 Mr. Smith called "Rebalancing." And it's 12 Exhibit B1737. 13 MR. GUIDO: What's the number again? 14 MR. BLANKENSTEIN: 230, Ken. 15 A. I have two copies. Is that okay? 16 Q. (BY MR. NICKENS) While we're at it, 17 why don't we look at Exhibit B2055? Have you 18 reviewed those articles before? 19 A. If I have, I don't recall. At least 20 not in this form. I have reviewed an article by 21 Mr. Smith. 22 Q. Which one was that? 3230 1 A. It was something that was put out from 2 the Federal Home Loan Bank of Dallas. I believe 3 he co-authored it with somebody. I could be wrong 4 about that, but I don't believe it was this 5 particular article. 6 Q. Is it this article called "Managing the 7 Risk in Risk-controlled Arbitrage"? 8 A. That's consistent with my memory. 9 Q. I believe this is a reprint or version 10 of 2055. 11 A. I'll be happy if everybody else is 12 happy to take that as fact. 13 Q. Well, I'll be happy -- you can have 14 that if you want to refer to that. But you said 15 earlier that you had seen in some article by 16 Mr. Smith an indication that a portfolio manager 17 should rebalance with movements of interest rates 18 of 25 basis points or slightly more. 19 Do you recall that testimony? 20 A. Yeah. I believe that I was somewhat 21 vague as to what the exact increments would be but 22 that I suggested that Mr. Smith had talked about 3231 1 rebalancing at smaller than -- smaller increments 2 than were chosen in the case of USAT. 3 Q. Can you point that out to us? We've 4 been unable to find it. 5 A. Well, let me see if I can review the 6 document. 7 Q. I can tell you, Dr. Duffie, that 8 Mr. Smith was unable to find any such reference 9 when he was here yesterday. Maybe you could send 10 in your answer. 11 A. Perhaps I'm mistaken in my memory, but 12 I definitely recall -- perhaps it was in 13 discussions with counsel -- the works of Terry 14 Smith and perhaps I've mischaracterized this 15 particular article. 16 Q. Yes, sir. Perhaps it was in 17 discussions with counsel that it was suggested 18 that 25 basis points might be an appropriate 19 number to mention? 20 A. No. In fact, again, I was rather 21 vague. I didn't remember exactly what I 22 understood that Mr. Smith had said about that. 3232 1 Q. And were you shown the contemporaneous 2 documents there are at least two specific 3 references we discussed yesterday in which 100 4 basis points was identified as a trigger point? 5 A. No. I can't point that out. 6 Q. And that wasn't discussed with you? 7 A. No, it wasn't. 8 Q. And if you look at the article by 9 Mr. Smith on rebalancing -- it's this one. It 10 indicates that this article -- this is at the 11 bottom of the first paragraph -- after discussing 12 the fact this was a third article that this 13 article focuses on measures that can indicate when 14 rebalancing is necessary. 15 A. I've just reviewed the article. Is 16 there a question pending? I'm sorry. 17 Q. Well, do you see there anywhere there 18 is a mention of any kind of numerical standard for 19 rebalancing? 20 A. I don't notice any numerical standard 21 that's offered in this article. 22 Q. Wouldn't you expect if you were 3233 1 publishing an article on rebalancing and the 2 indications of when you should do so that if there 3 was a widely-accepted standard, that you would 4 discuss it? 5 A. Well, first of all -- 6 Q. I mean, if you could answer the 7 question and then put any explanation that you 8 want on it. 9 A. Okay. If there was a widely-accepted 10 standard and you were writing in an attempt to 11 summarize what those standards were, then why not 12 mention it? But I don't think it would be a 13 necessary condition to have a widely-accepted 14 standard in order to propose strategies that 15 involved timely rebalancing of a portfolio and 16 that wouldn't necessarily be in a fixed number of 17 basis points. 18 Q. With regard to the broader rebalancing 19 issue, let me ask -- refer you back to Mr. Smith's 20 article, B2055, the one with his picture on it. 21 The fourth page. 22 MR. GUIDO: B2055? Is that what you 3234 1 said? 2 MR. NICKENS: Yes, sir. 3 Q. (BY MR. NICKENS) Do you see in the -- 4 it's the last paragraph in the lower -- or the 5 last two paragraphs in the lower right-hand 6 column. 7 A. Of the front page? 8 Q. Of the fourth page. I'm going to start 9 with the second sentence. "Should rates move down 10 and a mortgage balance begin to prepay, the 11 mortgage securities should be rolled into lower 12 coupons. The alternative is to partially unwind 13 swaps to reach the desired short funding. The 14 latter rebalancing strategy is much more costly 15 than rolling down coupons because a payment will 16 be required to unwind the swap." 17 Now, is that consistent or inconsistent 18 with the advice that you gave us this afternoon? 19 A. Well, I don't believe it's much more 20 costly. The losses and the swaps are there 21 whether you offset them and recognize the loss now 22 or whether you leave them in the swap portfolio. 3235 1 So -- 2 Q. You just disagree with Mr. Smith? 3 A. I disagree with the portion of this 4 that deals with whether or not the latter strategy 5 is more costly. 6 Q. Now, you indicated that you should -- 7 and I want to make sure I got this right -- that 8 you should monitor the prepayments daily or 9 perhaps even -- did you say intradaily? 10 A. In some cases, but not in the case 11 of -- I wouldn't guess that it would be necessary 12 in the case of an institution that had relatively 13 generic positions and lots of capital. I wasn't 14 speaking very, very broadly there. 15 Q. So, you'd watch prepayments -- you're 16 talking about actual prepayments? 17 A. Well, you don't see them coming across 18 your screen, but you become aware as prices change 19 of instruments that things are happening and that 20 you would try to rebalance the portfolio as they 21 do. 22 Q. Well, I'm trying to figure out what 3236 1 you're looking at on this daily or more frequently 2 than daily basis. Are you looking at the 3 prepayments or are you looking at prices? 4 A. Well, again, you do not get intraday 5 prepayment information. Well, not to my knowledge 6 anyway. I think the information is not provided 7 intraday. At least I don't believe it was 8 provided. I couldn't testify to that as a fact. 9 Q. Well, how often was it provided back in 10 1985 and '86? 11 A. One had monthly mortality rates and one 12 also had projections of prepayments which changed 13 day by day. 14 Q. So, if you were watching on a daily 15 basis for the prepayment information or maybe even 16 more frequently than that, you'd be looking at the 17 same number for 29 days in a row? 18 A. Not true. 19 Q. Well, it only is published once a 20 month. Right? 21 A. The historical occurrence of prepayment 22 for the previous month is published at the end of 3237 1 the month. That doesn't mean that projections of 2 prepayments are going to stay the same through the 3 month, nor does it mean that the price of the 4 mortgage-backed security will not change to 5 reflect those changes in prepayment projection. 6 Q. So, you wouldn't be looking at 7 prepayments. No one would be looking at their 8 screen there for 29 days in a row at the same 9 number. Right? 10 A. If by "prepayments" here you're meaning 11 the recorded historical incidence of prepayments 12 during that month, no. 13 Q. So, what you meant to say was that 14 you'd be looking at prices? 15 A. Well, not only did I mean to say that. 16 I had said you'd be looking at prices as well as 17 changes in projected prepayment rates. 18 Q. And who published those? 19 A. I didn't say that they were published. 20 Q. Well, who made them available? 21 A. Well, in 1986, presumably, traders kept 22 track of what they were projecting for prepayments 3238 1 on various instruments. 2 Q. Now, these are those same traders that 3 have the competitive advantage over USAT and the 4 rest of the thrift institutions? 5 A. Among others. 6 Q. And so, you'd have to depend upon them 7 to give you the prepayment information that you 8 should be monitoring more than once a day? 9 A. Depends on your contacts. You 10 mentioned earlier that one might have contacts in 11 the industry. One could call them and say, "What 12 do you think prepayments are going to be on this 13 issue this month?" 14 Q. You see any problem with that? 15 A. No, not at all. 16 Q. Now, when did the roll-down begin? 17 A. I don't think I can give you an exact 18 date. 19 Q. Well, what assumption have you made as 20 to when it began? 21 A. Well, I didn't need to make an 22 assumption to reach any of the conclusions that 3239 1 I've reached. 2 Q. Well, I thought that you gave the 3 opinion to Judge Shipe that interest rates had 4 moved 200 basis points before there was any sale 5 or any roll-down. 6 A. I'm sorry. I thought you were speaking 7 chronologically. Yes. 8 Q. To do that, you'd have to make some 9 assumption as to when the roll-down began, didn't 10 you? 11 A. No. In order to reach the conclusion 12 that the roll-down strategy happened too late, all 13 I would need to know is that it happened after 14 interest rates had already gone down 200 basis 15 points. 16 Q. Well, that's what I'm asking you. When 17 did it begin? 18 A. Well, you can let me refresh my memory 19 by checking to see when interest rates came down 20 significantly. It would be sometime in the middle 21 of 1986; but if you want it closer to a day or a 22 week, I might have to consult with any documents 3240 1 that are available that would give me a more 2 precise date. 3 Q. Let me see if I understand what's going 4 on here. In order to determine whether or not 5 the -- when the roll-down began, you're going to 6 check on when interest rates fell 200 basis 7 points? 8 A. Well, there might be a faster way. If 9 you're only interested in finding -- 10 Q. It seems like the conclusion is sort of 11 foregone if the opinion is that it began after 12 interest rates fell 200 basis points and the way 13 you're going to find out is to look at when 14 interest rates fell 200 basis points. Have I 15 misunderstood this? 16 A. I'm sorry. I didn't know that there 17 was any controversy over the issue that the 18 roll-down began substantially after interest rates 19 had fallen 200 basis points. If that point is 20 still at issue, then I'll try to go back into the 21 documentary evidence and find out exactly when 22 those transactions occurred. But if you need an 3241 1 exact date at this time as I sit here, I can't 2 give you one unless, for example, I'm allowed to 3 rely on this fact which I had taken to be assumed 4 that the roll-down began after a 200 basis point 5 drop in the interest rates. 6 Q. Well, who gave you that fact to be 7 assumed? 8 A. I can't tell you where I first heard 9 it. I know I discussed it with counsel. 10 Q. Can you give us any other source for 11 this assumption? 12 A. No. No. 13 Q. What were the rates at which -- what 14 were the rates at when the roll-down began? Sir, 15 if you don't know when the roll-down began, you 16 can't answer that be question. 17 A. Oh, that's not necessarily true. 18 Q. Okay. Explain to be me how you can. 19 A. I can go backwards by the same logic 20 that we just did. I can see when rates went down 21 200 basis points and go back up and then find out 22 what rates were at that point in time. But I 3242 1 can't tell you if you're just asking me to sit 2 here today what interest rates were on a specific 3 date in 1986. 4 Q. Well -- 5 A. At least not without refreshing my 6 memory. 7 Q. What was the coupon on the securities 8 that they sold? 9 A. Well, they started with high coupon 10 MBSs on the order of 11 to 12, 12 and a half 11 percent. So, those are the ones that rolled off 12 first. I presumed those are the ones that they 13 replaced first. 14 Q. Have you looked at what happened to the 15 amount of invested proceeds from the beginning to 16 the end of the roll-down? 17 A. I believe Mr. Guido asked me that 18 question this morning and I was not able to 19 identify what happened exactly to those proceeds, 20 whether they were entirely plowed back into 21 mortgage-backed securities or not. My assumption 22 is that they were not, but I don't know that. 3243 1 Q. Well, you've given the opinion that 2 they did not invest their proceeds. 3 A. That's from the fact that they lost 4 substantial amounts from the mismatch of the swaps 5 portfolio with the mortgage-backed portfolio. If 6 they had reinvested the proceeds as they arrived, 7 there would have been a smaller loss. 8 Q. So, from the fact that there was a 9 loss, you have inferred that they didn't reinvest 10 the proceeds? 11 A. Well, the swaps lost on the order of 12 120 million. The mortgage portfolio gained on the 13 order of 79 million. So, that 40 million gap is 14 presumably associated with the fact that the 15 proceeds of those sales were not reinvested in 16 additional mortgage-backed securities so as to 17 maintain a duration match between the swaps 18 portfolio and the mortgage-backed securities 19 portfolio, at least a large fraction of it. 20 Q. Now, have you been told that what they 21 were doing is that they were rolling the sale of 22 the old MBS into the basis for the substitute MBS? 3244 1 A. What do you mean by "the basis"? 2 Q. They were reducing their cost basis of 3 the substitute purchases by the amounts of their 4 gains. Have you been told that? 5 A. I don't even understand the question 6 yet. Maybe -- 7 Q. Then I gather you haven't been told 8 that? 9 A. Well, it's possible that I've been told 10 that in other terms, but I didn't understand the 11 terms in which you placed the question. 12 Q. Now, if we check the records and see 13 that the -- in fact, the amount of the invested 14 proceeds stays about the same throughout this 15 period, what conclusion should we draw with regard 16 to your opinions? 17 A. When you say "invested proceeds stays 18 the same," do you mean the amount of market value 19 worth of mortgage-backed securities stayed the 20 same through this period? 21 Q. I'm talking about the face amount. 22 A. Okay. From that, you wouldn't 3245 1 necessarily conclude that much because the face 2 amount is not the same as market value, 3 particularly when you're comparing high coupon to 4 low coupon MBS. 5 Q. You would expect it to grow? 6 A. Depends on whether all the proceeds 7 were invested or not. 8 Q. Now, you said -- you told the Court 9 that in addition to monitoring daily the 10 prepayments that when -- and let me make sure I 11 heard this correctly. You said that what you 12 ought to do is you ought to watch it and when you 13 see a prepayment, you ought to buy substitute MBS. 14 Right? 15 A. Well, just to be very clear that this 16 wasn't my first choice for the strategy. My first 17 choice was not to try to stick with the same 18 static swap portfolio and try to keep adjusting 19 the mortgage portfolio. But I said if one is 20 committed to that strategy, then one wouldn't be 21 able, as we agreed a few moments ago, to see the 22 prepayments on a daily basis but one would get 3246 1 prepayment information on a daily basis and 2 attempt to adjust the mortgage-backed security 3 portfolio in light of that information. 4 Q. Surely it wouldn't make any sense for 5 you to wait until you received the prepayment, 6 correct? 7 A. Well, certainly if interest rates go 8 down through the month substantially and you were 9 getting a lot of news about the fact that 10 prepayment rates are growing rapidly, no, it would 11 not make sense to wait until the end of the month. 12 Q. Until the prepayment was actually made 13 because you get the prepayment at par. Right? 14 A. Well, if someone prepays their mortgage 15 in entirety, it's not par value in -- well, it's 16 par value in the usual sense. It's a sinking fund 17 bond, that's true. 18 Q. One of the purposes of trying to roll 19 down is to capture the premium associated in the 20 price with those premium coupon MBS, isn't it? 21 A. When you say "capture," what do you 22 mean? 3247 1 Q. Well, as interest rates go down, the 2 value of the premium coupon goes up. Right? 3 A. In this case, yes. 4 Q. But if you wait till it's prepaid, then 5 you get your money at par. And so, you have lost 6 the opportunity of that higher price attributable 7 to it prior to the prepayment? 8 A. Well, that's not necessarily -- I mean, 9 putting aside the issue of what's the right 10 strategy from the point of view of interest rate 11 risk, you haven't necessarily lost anything by 12 hanging on to the security. You're hanging on to 13 something whose market value is presumably going 14 to be judged fairly in the market. So, 15 delaying -- 16 Q. You would expect the marketplace to 17 value those things efficiently? 18 A. I wouldn't expect that I would be able 19 to make a choice sitting, let's say, as to whether 20 the market had under or overpriced it. So, on 21 grounds of whether I would profit from trading now 22 or profit from trading at the end of the month 3248 1 when the prepayment had occurred, I would 2 certainly see no economic advantages in pursuing 3 taking those profits intramonth. Again, putting 4 aside the question of interest rate risk, I just 5 testified a moment ago that on grounds of interest 6 rate risk, you would, indeed, want to take action 7 intramonth. 8 Q. Well, it is plain that you wouldn't 9 want to wait until you actually had the 10 prepayment? 11 A. With regard to interest rate risk, no, 12 absolutely not. 13 MR. NICKENS: That's all I have, Your 14 Honor. 15 THE COURT: Do you have some redirect? 16 MR. GUIDO: Yes, Your Honor. 17 Q. (BY MR. GUIDO) The -- 18 THE COURT: If you'd like to move the 19 podium, you may do so. 20 MR. GUIDO: I would, Your Honor. 21 THE COURT: Let me ask a question 22 before the redirect. I understood you to say that 3249 1 the United MBS securities were not hedged? 2 A. No, Your Honor. They weren't. They 3 were partially hedged, but the remaining risk was 4 very, very large. 5 THE COURT: So, if interest rates went 6 down, then they would stand to gain from that; is 7 that correct? 8 THE WITNESS: That's correct. At least 9 if they went down in a nice, regular fashion. 10 THE COURT: But that would not be the 11 same for the USAT securities; is that correct? 12 THE WITNESS: If you consider USAT as a 13 whole, that institution would also stand to profit 14 from a reduction in interest rate risk -- pardon 15 me -- from a reduction in interest rates, although 16 in relative terms, it wasn't quite as severe as it 17 was for the United MBS portfolio. 18 THE COURT: Was that because of the old 19 mortgages that they had outstanding? 20 THE WITNESS: It was because the United 21 MBS portfolio did not have very much liabilities 22 of long maturities, in other words, things that 3250 1 would offset the changes in the mortgage 2 portfolio, nor did it have any substantial amount 3 of hedges in it whereas the institution as a 4 whole, while it was very exposed relative to its 5 tiny capital, did have a lot of different things 6 in it that caused this 5 to 7 billion-dollar 7 portfolio not to lose such a big fraction of its 8 market value as would have occurred with the 9 United MBS portfolio. 10 I don't know if I said that in clear 11 terms. I can repeat my answer in other terms if 12 you'd like, Your Honor. 13 THE COURT: But I was under the 14 impression that the United MBS portfolio would 15 stand to gain by a reduction in rates; is that 16 correct? 17 THE WITNESS: As would the entire 18 institution, USAT itself, that's correct, Your 19 Honor. 20 THE COURT: Would it be correct, 21 though, as to only the mortgage-backed securities 22 that were funded through the swaps and hedged in 3251 1 that procedure. 2 THE WITNESS: Well, we can actually see 3 that in Table 9.1. If you just look at the 4 mortgage-backed securities in USAT, including the 5 effects of swaps, it was also -- it also stood to 6 gain from a reduction in interest rates and, 7 conversely, to lose from an increase in interest 8 rates. 9 THE COURT: USAT's portfolio? 10 THE WITNESS: USAT's portfolio, as 11 well. So, both portfolios were positioned the 12 same way. Both stood to drop in value with an 13 increase in interest rates and to increase in 14 value with a reduction in interest rates. 15 THE COURT: All right. Thank you. 16 That was my question. Excuse me, Mr. Guido. You 17 may continue. 18 MR. GUIDO: That's all right, Your 19 Honor. 20 21 REDIRECT-EXAMINATION 22 3252 1 (4:07 p.m.) 2 Q. (BY MR. GUIDO) Now, when you talk 3 about the entire portfolio, are you talking about 4 the table on Page 21 as opposed to your discussion 5 of 9.1 which you refer to as Joe's portfolio? 6 A. Right. A moment ago in response to a 7 question to the Court, I was referring to Table 8 9.1 which is all of the mortgage-backed securities 9 held in USAT along with its swap hedges. 10 Q. I thought 9.1 was mortgage-backed 11 securities. 12 A. Yes. But the net effect of the 13 mortgage-backed securities as well as the swap 14 hedges are all recorded, at least according to 15 USAT's estimates in Table 9.1, where as in my 16 report, that information -- that portfolio at 17 least is included in the entire institution 18 counting all the other liabilities and hedges that 19 were in place. 20 Q. And your chart on Page 21 includes both 21 the information for the chart of 9.1 and also the 22 chart on 8.1 as well as all the rest of the 3253 1 interest rate sensitive instruments in the 2 portfolio? 3 A. One has to be careful. It includes the 4 information in Chart 9.1 because that same 5 information shows up on -- in my report. With 6 regard to the information on Chart 8.1 which is 7 the -- unless I have reversed them -- is the USAT 8 portfolio. I reversed it. So, let me start the 9 answer again. Just inverted those. 10 In my report on Page 21, I show the 11 sensitivity of the United MBS portfolio as 12 recorded in Table 8.1 -- this is data provided by 13 USAT -- with regard to the USAT mortgage-backed 14 security portfolio which is illustrated in Table 15 9.1, the same portfolio is included in my 16 analysis. But I did not use these numbers because 17 I had no way to break them out separately. 18 Q. I understand. But the assumption is 19 that it's included in the table on Page 21? 20 A. That's correct. 21 Q. That's all I wanted to clarify, Your 22 Honor. This is -- there are two subsets and then 3254 1 on Page 21, there is the entire institution which 2 includes those two subsets plus other interest 3 rate sensitive instruments? 4 A. That's correct. 5 Q. Now, the -- you were asked a number of 6 questions about the assumption about when the 7 roll-down should occur and you were asked some 8 questions about Terry Smith's testimony. And will 9 you look at B2055 again, please? 10 A. Yes. 11 Q. And look at the -- on Page 73 of that 12 document, the upper right-hand corner, the second 13 paragraph, the second sentence. It starts "If no 14 rebalancing is completed." Do you see that 15 sentence? 16 A. Yes. 17 Q. It says "If no rebalancing is 18 completed" -- 19 MR. NICKENS: I'm sorry. I apologize. 20 I'm not -- I don't have the place. 21 MR. GUIDO: On Page 73, the top 22 right-hand corner. 3255 1 MR. NICKENS: Okay. 2 Q. (BY MR. GUIDO) And the paragraph 3 begins "The results of the 200 basis point." Do 4 you see that, Mr. Duffie? 5 A. Yes, I do. 6 Q. Now, you go down to the middle of that 7 paragraph and it says "If no rebalancing is 8 completed by the time rates have moved 200 basis 9 points, the chances are that any rebalancing will 10 be too little too late." 11 Do you agree or disagree with that 12 statement, Dr. Duffie? 13 A. I agree with that statement. 14 Q. Now, you were asked questions about 15 what -- how you would ascertain whether or not a 16 portfolio was speculative or not speculative. And 17 the Judge asked you a question about United MBS's 18 portfolio and although we haven't gotten to that 19 point yet in the case to put in the specifics of 20 the portfolio, when you looked at the table on 8.1 21 which was the table of the portfolio of United 22 MBS, you answered two questions that the Judge 3256 1 presented to you. And I want to clarify something 2 about your answers. 3 Was there a period of time when that 4 portfolio was put on that it was substantially 5 unhedged? Looking at this chart, can you -- 6 A. Well, if you want to be really 7 technical, it was substantially unhedged from very 8 early on given that the institution had not enough 9 capital to support even these very small risks. 10 But on -- 11 Q. Absolute basis. 12 A. On an absolute basis, it became very 13 substantially unhedged beginning, as I think I 14 mentioned this morning, about -- about June of 15 1987 when you can see that the exposures to an 16 increase in interest rates become very large 17 relative to the exposures to a reduction in 18 interest rates. 19 Q. Let me rephrase the question. Can you 20 tell from this table whether or not this portfolio 21 had a large set of mortgage-backed securities in 22 the portfolio with no hedges? 3257 1 A. Well, let me say from the outset that I 2 don't know if these hedges which, for example, 3 includes caps, whether they are allocated in the 4 accounts of USAT one to one with mortgage-backed 5 securities. The accounting issue is not 6 particularly important here. 7 Q. I'm not asking accounting issues. I'm 8 just asking you whether or not based on this chart 9 you can ascertain -- because I think the Judge 10 asked you the question -- is at any period of 11 time -- I'm not talking about a mismatch. But at 12 any period of time, was there a substantial amount 13 of mortgage-backed securities that were bought in 14 this portfolio prior to the time that hedges were 15 bought? Can you ascertain that from this chart? 16 A. Yes. I mean, at least in terms of 17 small amounts of hedges, they were available from 18 the beginning. But in terms of substantial 19 amounts of hedges, there was never a substantial 20 amount relative to the amount of mortgage-backed 21 securities. For example, this portfolio did not 22 include a big swap portfolio as the other 3258 1 portfolio did. So, it was basically naked to 2 general rises in interest rates by mid to late 3 1987. The hedges that were in place were 4 relatively ineffectual with regard to protecting 5 it from interest rate risk. 6 Q. So that the portfolio when it was 7 structured was positioned to benefit from a 8 movement one direction or the other. Is that what 9 you're testifying to? 10 A. It was structured to benefit from a 11 movement in one direction in particular, yeah. A 12 reduction in interest rates. 13 Q. The -- you were given a document that 14 was marked Exhibit B2802 and you were asked a 15 number of questions before and after that about 16 when an OAS model was used. And apparently 17 counsel overlooked a few questions about that 18 document? 19 MR. NICKENS: Your Honor, I object -- I 20 object to those comments. I may have or I may 21 have decided not to ask some questions. 22 THE COURT: Let's leave that aside and 3259 1 go to your point, Mr. Guido. 2 MR. GUIDO: I'm sorry, Mr. Nickens. 3 Q. (BY MR. GUIDO) Look at Page 2 of 4 Exhibit B2802. It says in the first full 5 paragraph "Unfortunately, the static cash flow 6 yield analysis suffers from two weaknesses." 7 Is that basically the methodology that 8 was used at USAT, the static cash flow yield 9 analysis? 10 A. From the general description given 11 here, I would say yes. 12 Q. Okay. Now, it says "The first defect 13 is the static yield spread to comparable 14 treasuries not only includes a premium for default 15 risk and liquidity risk but compensation for the 16 short call position of the security holder as 17 well." 18 Do you see that sentence? 19 A. Yes, I do. 20 Q. Do you agree with that sentence? I 21 guess the first question is: What do you 22 interpret that sentence to mean? 3260 1 A. Well, first of all, when you know the 2 yield spread as measured at a static prepayment 3 schedule, then essentially you've just backed out 4 what the price of the instrument is. And in 5 knowing the price, some of that price is to 6 compensate the investor for default risk if there 7 is a credit, liquidity risk if that's an issue, 8 and some of it is to compensate the investor for 9 the fact that the homeowner has an option which 10 can be exercised against the investor. 11 So, that price is basically written 12 down for those three disadvantages of owning this 13 kind of a bond. 14 Q. Let's look at the next sentence. It 15 says "This equation" -- referring to the equation 16 that explains the sentence -- "clearly illustrates 17 why the static cash flow yield is inadequate for 18 comparing the relative value of securities." 19 Do you see that? 20 A. Okay. 21 Q. Do you agree with that sentence? 22 A. Well, first of all, if you're quoting 3261 1 these instruments at a yield and you don't adjust 2 the yield for the options embedded in the 3 portfolio, then that yield is going to mislead you 4 about these relative values. That is, the value 5 associated with, for example, the prepayment 6 option. 7 Q. Now, look -- 8 A. If you just -- 9 Q. I'm sorry. 10 A. -- know the price and you from that 11 back out the associated yield, whether you do it 12 on a static model or not, provided you know how 13 that yield was calculated, you could go back and 14 forth and find out whether the price was 15 appropriate or not. 16 Q. Okay. Now, look at the next paragraph. 17 It says "It follows that the cash flow yield 18 spread does not provide sufficient information for 19 investor choice since an investor cannot 20 distinguish whether the difference in his cash 21 flow yield spreads among securities are due to 22 credit risks, liquidity risks, or compensation for 3262 1 the short call positions." 2 A. Yeah. I agree with that. 3 Q. Now, is that why the static cash flow 4 analysis was not useful to engage in the value 5 trading that Sandy Lorenson described she was 6 doing in the United MBS portfolio? 7 A. Yes. And the principal problem was 8 associated with the prepayment aspect of it. That 9 is, that just doing a static yield analysis would 10 not highlight whether a security was, in fact, 11 fairly priced in the market because one would not 12 have been able with that static analysis to be 13 able to properly account for the prepayment 14 options. That is, relative to one's trading 15 partners which is what matters when one is trading 16 for value trading purposes. 17 Q. So, the differences in the market 18 prices could be based upon real changes in market 19 conditions and not just some anomaly in the 20 market? 21 A. I'm sorry. Could you repeat the 22 question? I didn't understand it. 3263 1 Q. So, the differences between the 2 relative values of two different securities that 3 Sandy Lorenson was saying she was trading against 4 each other could be due to real differences in 5 market changes of those values and not temporary 6 anomalies? 7 A. Well, yeah. To be more specific, if 8 you were trading a Fannie Mae and a Ginnie Mae 9 with the same coupon rate and you see one has a 10 higher yield spread as measured by a static model 11 than the other, you could be misled into thinking 12 that one is offering somehow a better investment 13 than the other when, in fact, it may be the case 14 that one is offered that higher yield because that 15 particular security is subject to greater 16 prepayment risk and that risk might only be 17 apparent through more careful analysis of the 18 options embedded in those mortgages; that is, for 19 example, through an option adjusted spread model. 20 I think that was a long-winded way of saying 21 "yes." 22 Q. Now, you were asked whether it was 3264 1 appropriate to measure the risk that was imposed 2 on this institution measuring its capital based on 3 the market value of its stock. 4 Do you recall those questions? 5 A. Yes. 6 Q. Now, was your opinion about the risks 7 that the mortgage-backed securities portfolio 8 imposed upon this institution based on its capital 9 levels only based on your analysis of the 10 relationship to the market value of the stock, or 11 did it include the other two forms of value that 12 we talked about in your direct testimony? 13 A. Fine. Well, certainly, the stock 14 market valuation was -- was the most accurate 15 signal to me of the actual capital available. But 16 where, for example, I'm not able -- where I'm not 17 able to see the stock market valuation, for 18 example, were this institution not to be publicly 19 traded and I would only have balance sheet 20 information and I looked at the United balance 21 sheet, my conclusions wouldn't change. These 22 risks were still extremely large relative to even 3265 1 the balance sheet accounting-based information on 2 its capital. 3 Q. And does that include its balance sheet 4 including goodwill? 5 A. Even if you folded in the goodwill -- 6 and I wouldn't do that, as I testified this 7 morning -- these risks are still very large 8 relative to capital counting the goodwill. 9 Q. You were asked a number of questions in 10 cross-examination that -- whether or not you were 11 engaging in self-fulfilling prophecy when you used 12 historical -- actual historical data as opposed to 13 anticipated prepayment data for the calculations 14 on your Chart 21. And you indicated that you had 15 used historical data there. But with regard to 16 the information that's available for the two 17 portfolios that we're discussing here, the one in 18 8.1, United MBS, and the one in 9.1, the one that 19 you refer to as Joe's portfolio or we refer to as 20 the USAT portfolio, were you using historic data 21 or were you using the projections that the 22 institution had at hand at that time? 3266 1 A. Well, in fact, we know from the fact 2 that the investment committee minutes were dated 3 that those were the estimates that USAT had, given 4 its own information, of the interest rate 5 sensitivity contemporaneously. That is, with its 6 own -- with its own data at that time of what 7 prepayments were expected to be at that time. 8 Q. So, this -- these two charts, 8.1 and 9 9.1, are their own data generated at that time by 10 them and reflected in their investment committee 11 minutes? 12 A. That's correct. 13 MR. NICKENS: Objection. Leading. 14 THE COURT: I think it summarizes the 15 statement. It's all right. 16 MR. GUIDO: Thank you, Your Honor. 17 Q. (BY MR. GUIDO) Now, you were asked a 18 question at the outset about the preparation of 19 your report. 20 A. Yes. 21 Q. And you were asked a number of 22 questions about what you relied upon to prepare 3267 1 your report. And you failed to mention that you 2 had relied upon the report that you had previously 3 used in the Far West case. 4 Do you recall that testimony? 5 A. Yes, I do. 6 Q. And you indicated that you thought when 7 you were asked the question because you had been 8 looking at the chart that I had that the question 9 was what USAT materials you had had. Is your 10 answer -- was your answer occasioned by the fact 11 that you misunderstood the question? 12 A. Well, had I listened to the question 13 very carefully, I would have seen the distinction 14 at one point in the question between reference to 15 the USAT documents and another question which 16 referred to documents generally. So, it was 17 simply an oversight on my part. I should have 18 understood the question, but I simply did not 19 recognize as the scope of that question changed 20 that it had changed. 21 Q. Okay. You were not intending to 22 mislead the Court? 3268 1 A. Oh, definitely not. 2 Q. Now, you were also given Exhibit B3773 3 that was the district court opinion in which there 4 were some statements made about your testimony by 5 the district court judge. 6 MR. GUIDO: And I'd just like to put 7 into the record the citation to the appellate 8 court decision which I didn't have at that time, 9 Your Honor, that overturned that opinion and what 10 the appellate court said about that opinion is 11 reflected in this citation. And the citation is 12 Franklin Savings versus Director Office of Thrift 13 Supervision. It's 934 Fed 2nd 1127, the Tenth 14 Circuit, 1991. 15 I have no further questions, Your 16 Honor. Thank you, Dr. Duffie. 17 MR. NICKENS: Your Honor, I have no 18 objection to your consideration. My understanding 19 of that appellate decision was that it was 20 reversed on the basis that the district court had 21 considered evidence other than the OTS record. 22 But whatever it may reflect, I have no objection 3269 1 to your consideration for that opinion for 2 whatever purpose you find it appropriate. 3 THE COURT: All right. Thank you. Do 4 you have any redirect? 5 MR. NICKENS: I have no recross. 6 THE COURT: Do any of the other 7 respondents have questions? 8 MS. CLARK: No, Your Honor. 9 MR. BLANKENSTEIN: No questions. 10 MR. KEETON: No, Your Honor. 11 MR. EISENHART: No questions, Your 12 Honor. 13 THE COURT: All right. Thank you, 14 Mr. Duffie. You may step down. 15 THE WITNESS: Thank you, Your Honor. 16 THE COURT: Dr. Duffie. I'm sorry. We 17 will adjourn until 9:00 o'clock tomorrow. 18 19 20 21 22