Free Memorandum - District Court of Federal Claims - federal

Case 1:93-cv-00531-LAS

Document 226-3

Filed 11/07/2007

Page 1 of 16

EXHIBIT II

Case 1:93-cv-00531-LAS

Document 226-3

Filed 11/07/2007

Page 2 of 16

UNITED STATES COURT OF FEDERAL CLAIMS

__________________________________________ ) AMBASE CORPORATION and ) CARTERET BANCORP, INC., ) ) Plaintiffs, ) ) and ) ) FEDERAL DEPOSIT INSURANCE ) CORPORATION, ) ) Plaintiff-Intervenor, ) ) v. ) ) THE UNITED STATES OF AMERICA, ) ) Defendant. ) __________________________________________)

No. 93-531 Judge Loren Smith

SUPPLEMENTAL DECLARATION OF CHARLES W. CALOMIRIS OF JULY 12, 2007

Introduction I. II. III. Carteret's Cost of Deposits Increased as a Result of the Breach, Reducing Its Capital Carteret's Capital Position But-For the Breach Would Have Improved as a Result of Changes in Its Mix of Assets Carteret's Reserves Were Almost Sufficient to Cover the Difference Between the Actual Sales Value of Its Commercial Real Estate Assets and the Principal Balance of Those Assets

CONFIDENTIAL MATERIALS ­ SUBJECT TO PROTECTIVE ORDER

Case 1:93-cv-00531-LAS

Document 226-3
-2-

Filed 11/07/2007

Page 3 of 16

IV. IV.

Appraisals of Properties in Carteret's Commercial Real Estate Portfolio Show that Full Sales Value Was Not Achieved by the RTC Correction of Table 19 from Previous Declaration

Appendix A: Materials Relied Upon

INTRODUCTION 1. This report supplements my report of May 19, 2007, which considered the

damages to AmBase Corporation ("AmBase") and Carteret Bancorp ("Carteret") that resulted from the elimination of supervisory goodwill from Carteret's regulatory capital requirements.1 In this report, I provide additional details regarding certain items that were included in my previous report and make minor corrections to some of the calculations included in that report. These items include the increase in Carteret's cost of deposits resulting from the breach, Carteret's capital position but-for the breach in 1992, and Carteret's commercial real estate portfolio. I. CARTERET'S COST OF DEPOSITS INCREASED AS A RESULT OF THE BREACH, REDUCING ITS CAPITAL 2. In my previous report, I showed that Carteret's interest rate on six-month CDs

was 105.5 basis points lower than the average rate on six-month negotiable CDs in the secondary market on April 3, 1989.2 After the breach, Carteret was forced to pay higher interest rates on deposits in order to keep its depositors. By the end of 1990, Carteret's interest rate on six-month CDs was 1 basis point higher than the average rate on six-month negotiable CDs in the secondary market.3 In 1991 and 1992, Carteret's rates on six-month CDs were consistently higher than the comparable rates offered on CDs in the secondary market.4

1. 2. 3. 4.

Declaration of Charles W. Calomiris of May 19, 2007. Id. at 54. Id. Id.

CONFIDENTIAL MATERIALS ­ SUBJECT TO PROTECTIVE ORDER

Case 1:93-cv-00531-LAS

Document 226-3
-3-

Filed 11/07/2007

Page 4 of 16

3.

As discussed in my previous declaration, precisely calculating how much extra

interest Carteret was forced to pay depositors as a result of the breach is challenging. One must account for the changes in Carteret's mix of deposits, how that mix was affected by the breach, the interest rates offered to depositors on the different classes of deposits by Carteret, and lags in repricing of deposits. I performed a straightforward calculation in my previous declaration based on the assumption that 1991 and 1992 deposits in CDs with durations between 6 and 23 months at origination carried interest rates that were 105.5 basis points higher than their interest rates but-for the breach.5 In this section, I perform a slightly more detailed analysis of the interest overpayments. 4. Estimating the extra interest paid by Carteret as a result of the breach requires

assumptions about (1) the timing of rate changes, and (2) the timing of deposit repricing. One possible assumption about the timing of rate changes would be that the market immediately incorporated the full effects of the breach at the time of the breach. Given my conclusion that the spread between the actual and but-for interest rates was at least 105.5 basis points at the end of 1990, then this more aggressive approach would also conclude that the spread between the actual and but-for interest rates was 105.5 basis points in August 1989--the time of the breach. A more moderate assumption would be that the spread between actual and but-for interest rates gradually increased over time, such that the spread was zero at the time of the breach and gradually increased to 105.5 basis points by the end of 1990. The most conservative assumption about the timing of interest rate changes would be that the spread between actual and but-for interest rates was zero until the end of 1990, at which point it immediately changed to 105.5 basis points.

5. The calculation in my original report contained an error. The calculation in Table 18 of my previous declaration did not account for the conservative assumption that only half of the 6-23 month CD deposits in 1991 and 1992 carried breach-affected rates. Instead, the calculation shown in Table 18 represented the assumption that all 6-23 month CD deposits in 1991 and 1992 carried breach-affected rates. Id. at 56-57.

CONFIDENTIAL MATERIALS ­ SUBJECT TO PROTECTIVE ORDER

Case 1:93-cv-00531-LAS

Document 226-3
-4-

Filed 11/07/2007

Page 5 of 16

Although the more moderate assumption of a gradual change in the spread is the most reasonable assumption, in the analysis in the remainder of this supplemental report I use the most conservative assumption that the spread between actual and but-for interest rates does not change until the end of 1990. 5. In my previous report, I focused on CDs with durations at origination of between

6 months and 23 months. Table 1 shows the year-end deposit balances in CDs with those durations from 1989 to 1992. TABLE 1: BALANCE OF DEPOSITS IN CARTERET CDS WITH DURATIONS OF 6-23 MONTHS AT ORIGINATION, 1988 - 1992
6-11 Month CDs 12-23 month CDs Total 6-23 Month CDs (Million $) (Million $) (Million $) Sep. 1988 577.0 801.3 1,378.3 Dec. 1989 634.4 1,196.7 1,831.1 Dec. 1990 545.6 860.6 1,406.1 Dec. 1991 583.8 869.0 1,452.8 Dec. 1992 427.1 602.9 1,030.0 Sources: Carteret Bancorp Inc. and Subsidiary, FHLB Docket #02-4702, Consolidated Financial Statements, Sept. 30, 1988 (AMB013960-AMB013995, at AMB013979); Carteret Bancorp Inc., 1989 Annual Report, Dec. 31, 1989 (C-AM-A-0394527); Carteret Savings Bank, FA & Subsidiaries, Consolidated Financial Statements, Dec. 31, 1990, (AMB014070); Carteret Savings Bank, FA and Subsidiaries, 1992 Annual Report, at 17 (CAM 5070016). Date

6.

To estimate the timing of deposit repricing, I take the fiscal year-end deposit

balances for 6-11 month CDs and 12-23 month CDs shown in Table 1. For each of these two categories at the end of a given fiscal year, I assume that durations of the CDs at origination are evenly distributed--that is, one-sixth of the balance in 6-11 month CDs is in 6-month CDs, onesixth is in 7-month CDs, and so on. I assume one-twelfth of the balance in 12-23 month CDs is in 12-month CDs, one-twelfth is in 13-month CDs, and so on. 7. Next, I assume that the months of origination for each given CD duration are

evenly distributed. Thus, one-sixth of the December 1992 balance in 6-month CDs was originated in December 1992, one-sixth was originated in November 1992, and so on. One-

CONFIDENTIAL MATERIALS ­ SUBJECT TO PROTECTIVE ORDER

Case 1:93-cv-00531-LAS

Document 226-3
-5-

Filed 11/07/2007

Page 6 of 16

twelfth of the December 1992 balance of 12-month CDs was originated in December 1992, onetwelfth was originated in November 1992, and so on. 8. This results in a time series of originations for each type of CD at the end of each

fiscal year. In some cases, there is overlap between the time series for a given type of CD across fiscal years. For example, the estimated amount of 18-month CDs originated in December 1990 when using the December 1991 balance of 12-23 month CDs is different than the estimated amount of 18-month CDs originated in December 1990 using the December 1992 balance of 1223 month CDs. In these cases, I use the average of the two estimates to arrive at the originations estimate in a given month for a particular type of CD. 9. For some CD durations in some months, the method described above does not

result in an estimate of the amount originated in the type of CD in a given month. For example, the method described above does not give an estimate of the amount of 6-month CDs originated in January 1991 because all the 6-month CDs in existence at the end of December 1991 were originated in July 1991 or later. In these cases, I use the average of the origination estimate from the latest month before the given month and the origination estimate from the earliest month after the given month. For example, I estimate the originations of 6-month CDs in January 1991 to be equal to the average of the estimate of the December 1990 6-month CD originations and the July 1991 6-month CD originations. 10. Given this time series of CD originations, I can estimate Carteret's monthly CD

balance for every month going back to the date of the breach. Given this monthly time series, and using the most conservative estimate that the difference between Carteret's actual interest rates and but-for interest rates increased from zero to 105.5 basis points at the end of 1990, I can estimate the amount of interest overpayments that were made each month from January 1991

CONFIDENTIAL MATERIALS ­ SUBJECT TO PROTECTIVE ORDER

Case 1:93-cv-00531-LAS

Document 226-3
-6-

Filed 11/07/2007

Page 7 of 16

through December 1992. Each month's overpayment is equal to the estimated balance in 6-23 month CDs that month that were originated in January 1991 or later, multiplied by 105.5 basis points, divided by 12. This calculation results in interest overpayments of $19.8 million from January 1991 through December 1992, and $17.0 million from January 1991 through September 1992. When averaging across all months in 1991, 50 percent of the balance in Carteret's 6-23 month CDs was originated at breach-affected rates (assuming rates were not affected by the breach until the end of 1990). When averaging across all months in 1992, 95 percent of the balance in Carteret's 6-23 month CDs was originated at breach-affected rates (assuming rates were not affected by the breach until the end of 1990). This calculation is conservative because (1) I am using the most conservative assumption about the timing of the interest rate changes, (2) the difference between actual interest rates and but-for interest rates was more than 105.5 basis points in 1991 and 1992, as Table 17 of my previous report showed, and (3) I am only examining interest payments on 6-23 month CD deposits even though the costs of other deposits were affected by the breach. II. CARTERET'S CAPITAL POSITION BUT-FOR THE BREACH WOULD HAVE IMPROVED AS A RESULT OF CHANGES IN ITS MIX OF ASSETS 11. In Table 16 of my previous declaration, I gave a conservative estimate of a $36.9

million deficit in risk-based capital, with a surplus in tangible capital and core capital.6 As discussed at length during my deposition, Carteret could have improved its risk-based capital position by securitizing its mortgage assets. Carteret had $1.561 billion in single-family residential mortgage loans as of September 30, 1992 that were given 50 percent risk-weight.7 Carteret could have securitized these loans into mortgage-backed securities (MBS), which

6. 7.

Id. at 48-49. Carteret Savings Bank, FA, Thrift Financial Report, Sept. 30, 1992, at 36 (CAM5350309).

CONFIDENTIAL MATERIALS ­ SUBJECT TO PROTECTIVE ORDER

Case 1:93-cv-00531-LAS

Document 226-3
-7-

Filed 11/07/2007

Page 8 of 16

carried a 20 percent risk-weight. I assume that Carteret would have incurred $3.9 million in underwriting fees--equal to 0.25 percent of the principal balance of the loans being securitized.8 The combination of changing this mix of assets and substituting $17.0 million in deposit interest expense overpayment for the $28.2 million estimate from my previous declaration reduces Carteret's but-for risk-based capital deficit in September 1992 from $36.9 million to $18.2 million, decreases Carteret's but-for tangible capital surplus from $78.5 million to $63.6 million, and reduces Carteret's but-for core capital from a $3.8 million surplus to a $10.8 million deficit. Table 2 reflects Carteret's but-for capital position as of September 1992 given these changes.

8. My assumption of underwriting fees equal to 0.25 percent of the principal balance of the securitized loans is based on other mortgage securitizations from the 1990s. See, e.g., New Century Financial Corp., SEC Form 10K405, Dec. 31, 1997, at Exhibit 10.27.

CONFIDENTIAL MATERIALS ­ SUBJECT TO PROTECTIVE ORDER

Case 1:93-cv-00531-LAS

Document 226-3
-8-

Filed 11/07/2007

Page 9 of 16

TABLE 2: CARTERET'S BUT-FOR CAPITAL PROFILE AS OF SEPTEMBER 1992 (UPDATED TABLE 16 FROM DECLARATION OF CHARLES W. CALOMIRIS OF MAY 19, 2007) (MILLION DOLLARS)
Tangible Capital Core Capital Adj. Tangible Assets 4,765.1 Risk-Based Capital Fully Capitalized Items 12.7

Actual Level Adjustments: ­ Qualifying Supervisory Goodwill + But-For Total Supervisory Goodwill Balance ­ Premiums from 1990 Branch Sales + Transaction Costs from Branch Sales + Deposit Interest Expense Overpayment Resulting from Breach + Supplementary Capital + Profits Lost from Asset Shrinkage + Profits Lost from Deposit Runoff + Cash Infusion from AmBase + Reduction in Asset Writedowns ­ 50% Risk-Weight on Single-Family Residential Mortgage Loans + 20% Risk-Weight on MBS ­ MBS Transaction Costs But-For Level Actual Requirement Actual Surplus (Deficit) But-For Requirement But-For Surplus (Deficit)

Tangible Capital (61.7)

Tangible Assets 4,765.1

Core Capital (14.0)

Risk-based Capital (15.2)

Risk-Weight Assets 2,882.5

0.0

0.0

47.7

0.0

47.7

47.7

0.0

139.1

139.1

139.1

139.1

139.1

139.1

0.0

40.7

40.7

40.7

40.7

40.7

0.0

0.0

2.4

2.4

2.4

2.4

2.4

0.0

0.0

17.0

17.0

17.0

17.0

17.0

0.0

0.0

0.0 Not calculated Not calculated 36.0

0.0

0.0 Not calculated Not calculated 36.0

0.0 Not calculated Not calculated 36.0

43.3 Not calculated Not calculated 36.0

0.0

0.0 Not calculated Not calculated 0.0

Not calculated

Not calculated

Not calculated

Not calculated

36.0

0.0

50.0

50.0

50.0

50.0

50.0

75.0

0.0

-

-

-

-

1,561.4

780.7

0.0

-

-

-

-

1,561.4

312.3

0.0

3.9 138.1 71.5 (133.2) 74.5 63.6

3.9 4,964.9

3.9 138.1 143.0 (157.0) 148.9 (10.8)

3.9 4,964.9

3.9 180.2 220.2 (235.4) 198.5 (18.2)

0.0 2,580.6

0.0 12.7

III. CARTERET'S RESERVES WERE ALMOST SUFFICIENT TO COVER THE DIFFERENCE BETWEEN THE ACTUAL SALES VALUE OF ITS COMMERCIAL REAL ESTATE ASSETS AND THE PRINCIPAL BALANCE OF THOSE ASSETS 12. I have reviewed documents related to the sales of every property in Carteret's

commercial real estate portfolio at the time it was seized by the government at the end of 1992.

CONFIDENTIAL MATERIALS ­ SUBJECT TO PROTECTIVE ORDER

Case 1:93-cv-00531-LAS

Document 226-3
-9-

Filed 11/07/2007

Page 10 of 16

Based on the RTC's and FDIC's experience in disposing of Carteret's commercial real estate assets, I estimate that the reserves were almost sufficient to cover the losses from the sales. 13. Carteret's commercial real estate assets during conservatorship/receivership fall

into two categories: non-securitized assets and securitized assets. Based on the pertinent documents, Carteret's non-securitized assets, which included whole loans and REOs, were sold by the government for a total of $325.9 million.9 These assets had a principal balance at the time they were sold of $443.7 million.10 14. For the securitized assets, the RTC pooled together the remaining loans and

created securities based on the cash flows generated by these assets. For each pool of securitized assets, the RTC created three categories of securities: Class A certificates, Class B certificates, and bonds.11 Both the Class A certificates and bonds were auctioned, while the RTC held the Class B certificates.12 The Class A certificates represented a 49 percent claim on the residual equity position of the securitization and granted the buyer the right to actively manage the portfolio of real estate investments underlying the securitization.13 The Class B certificates represented a claim on the remaining 51 percent of the residual equity, but conferred an entirely passive position on the RTC.14 Furthermore, as the underlying real estate assets contained in the securitization returned principal payments over the course of time, valuation of the portfolio

9. Arriving at this figure involved aggregating the sales proceeds indicated for each non-securitized loan and REO from documents listed in the Real Estate Sales section of the Materials Relied Upon in Appendix B in my previous declaration. See Declaration of Charles W. Calomiris of May 19, 2007, at 117-118. 10. Arriving at this figure involved aggregating the principal balance at disposition indicated for each nonsecuritized loan and REO. See Id. 11. Executive Summary, Mortgage Securitization Case, RTC 1993-S1 Case (C-AM-A-0360076 ­ 83 at C-AMA-0360076); RTC Multiple Investor Fund Case (RTC Mortgage Trust Series 1993-N3) (Aug. 17, 1993) (C-AM-A0360029 ­ 36 at C-AM-A-0360030). 12. Id. at C-AM-A-0360077; C-AM-A-0360031. 13. Id. 14. Id.

CONFIDENTIAL MATERIALS ­ SUBJECT TO PROTECTIVE ORDER

Case 1:93-cv-00531-LAS

Document 226-3
- 10 -

Filed 11/07/2007

Page 11 of 16

must also account for these "equity adjustments." Thus, valuation of the securitizations involves deriving the value of all of these potential cash flows. 15. As of November 30, 1992, Carteret had booked approximately $189.1 million in

reserves relating to its commercial real estate assets.15 The principal balance of both the nonsecuritized assets and the securitized assets amounted to approximately $691.6 million at the time the assets were sold.16 The RTC and FDIC obtained proceeds of approximately $452.2 million in disposing of the assets.17 Furthermore, from what I can determine from the available records, at least $15.0 million18 of additional principal was returned in the form of equity adjustments for the securitized assets. Accounting for all of these factors, the most conservative estimate of the portfolio value suggests that Carteret was under-reserved by no more than $35.3 million. Additional proceeds may have been generated from the disposition of Carteret's commercial real estate assets, but the estimate of $35.3 million represents my best understanding of the documents I have reviewed at this time. TABLE 3: POTENTIAL LOSSES TO CARTERET COMMERCIAL REAL ESTATE PORTFOLIO
[1] [2] [3] [4] [5]=[1]-[2]-[3]-[4] Principal Balance of all Assets at Disposition Reserve Balance at November 1992 Sales Proceeds of Assets Equity Adjustments Amount Under-Reserved Before Accounting for Remaining Principal Balance of Loans Dollars in Millions 691.7 189.1 452.2 15.0 35.3

15. Commercial Asset Portfolio Analysis (Nov. 30, 1992), at B-2 (CAM3862680). 16. Arriving at this figure involved aggregating the principal balance indicated for each loan, REO, and securitization from documents listed in the Real Estate Sales section of the Materials Relied Upon in Appendix B in my previous declaration. See Declaration of Charles W. Calomiris of May 19, 2007, at 117-118 17. Arriving at this figure involved aggregating the sales proceeds indicated for each loan, REO, and securitization. See Id. 18. Id.; Financial Data Warehouse ­ RRS Balance Sheet Drill Down Report for Period Ending Dec. 31, 1995, Account 071190.

CONFIDENTIAL MATERIALS ­ SUBJECT TO PROTECTIVE ORDER

Case 1:93-cv-00531-LAS

Document 226-3
- 11 -

Filed 11/07/2007

Page 12 of 16

16.

This estimate of the adequacy of Carteret's reserve levels given the actual

experience of its asset sales does not contradict the statement by Angelo Vigna, the top OTS regulator in the Northeast region, that Carteret's write-downs were "ultraconservative."19 For reasons discussed below and in my previous declaration, Carteret would have obtained better results in disposing of these assets than did the RTC/FDIC after Carteret was seized. For example, regulators have testified that a financial institution's assets lose up to twenty percent of their value when taken over by the government for disposition.20 But even assuming that Carteret would have done no better than the receiver in disposing of these assets and thus was "underreserved" by between $9.8 million to $35.3 million as of the time of seizure, additional losses of that magnitude would not, in the no-breach world, have been sufficient to lead to Carteret's failure. IV. APPRAISALS OF PROPERTIES IN CARTERET'S COMMERCIAL REAL ESTATE PORTFOLIO SHOW THAT FULL SALES VALUE WAS NOT ACHIEVED BY THE RTC 17. I have reviewed several appraisals of the assets in Carteret's commercial real

estate portfolio from 1991 or later.21 All the appraisal reports follow proper appraisal conventions. Additionally, all appraisal reports sufficiently support their appraisal decisions with substantiated real estate data.22 Real estate appraisers typically use three approaches to valuing property: the cost approach,23 the sales comparison approach,24 and the income capitalization

19. Declaration of Charles W. Calomiris of May 19, 2007, at 76. 20. Id. at 78-79. 21. The documents containing the appraisals I have reviewed were found in the Real Estate Appraisals section of my Materials Relied Upon in Appendix B in my previous declaration. See Declaration of Charles W. Calomiris of May 19, 2007, at 118-120. I have reviewed one other appraisal that was not included in the Materials Relied Upon from my previous declaration. See CAM918 1908 ­ 2014. 22. See APPRAISAL INSTITUTE, THE APPRAISAL OF REAL ESTATE 50 (12th ed. 2001). 23. Id. ("In the cost approach, value is estimated as the current cost of reproducing or replacing the improvements (including an appropriate entrepreneurial incentive or profit) minus the loss in value from depreciation plus land or site value.") 24. Id. ("In the sales comparison approach, value is indicated by recent sales of comparable properties in the market.")

CONFIDENTIAL MATERIALS ­ SUBJECT TO PROTECTIVE ORDER

Case 1:93-cv-00531-LAS

Document 226-3
- 12 -

Filed 11/07/2007

Page 13 of 16

approach.25 The appraisal reports of Carteret's properties that I have reviewed all use at least one of these approaches and provide sufficient support for their valuations. 18. Based on all the appraisals I have reviewed, Carteret's properties were sold by the

RTC on average for below their appraisal values. I examined the difference between the RTC's sale price and the pre-sale appraisal value of the assets in Carteret's commercial real estate portfolio by calculating the percentage a property sold for above or below its appraisal value. I calculated this by subtracting the appraisal value from the sales price, and then divided this result by the appraisal value. When multiple appraisal values after the beginning of 1992 were available, I used the lowest appraisal value. I excluded assets that lacked sales data or appraisal data, assets sold in securitizations, participations,26 and properties with a sale price of zero. TABLE 4: RELATIONSHIP BETWEEN SALES PROCEEDS & APPRAISAL VALUES
Equal-Weighted % Difference Between Sales Proceeds & Appraisal Value Appraisal Value Weighted % Difference Between Sales Proceeds & Appraisal Value

Number of Properties Relation of Sale Price to Appraisal(s) Properties Where Sale Price Exceeded 11 22.4% 25.2% All Appraisals Properties Where Sale Price Was Below 28 ­24.2% ­26.9% All Appraisals Total 39 ­11.1% ­12.8% Note: All appraisals included in this table were performed in 1992 or later. I exclude the appraisals for assets that were sold in securitizations and assets that I identified as being participations. Sources: Documents listed in Appendix B under Real Estate Sales and Real Estate Appraisals from the Declaration of Charles W. Calomiris of May 19, 2007; CAM918 1908 ­ 2014.

As Table 4 shows, I found that far more properties sold below their appraisal value than for above their appraisal value. In addition, the sales proceeds for Carteret properties were, by and large, significantly below their appraisal values. Even though a few properties sold at prices above the appraisal price, the average percentage spread for properties sold below appraisal

25. Id. ("In the income capitalization approach, value is indicated by a property's earning power, based on the capitalization of income.") 26. I identified participations based on asset descriptions in sales documents, appraisal documents, and a document listing Carteret's loan participations. See Loan Participations (C-AM-A-0177423 ­ 480).

CONFIDENTIAL MATERIALS ­ SUBJECT TO PROTECTIVE ORDER

Case 1:93-cv-00531-LAS

Document 226-3
- 13 -

Filed 11/07/2007

Page 14 of 16

value was significantly greater than the average percentage spread for properties sold above appraisal value. Sales prices ranged from 50 percent above appraisal price to 71 percent below appraisal price. The average difference between sales price and appraisal value for properties sold below appraisal value was 24.2 percent, whereas the average difference between sales price and appraisal value for properties sold above appraisal value was 22.4 percent. On average, the assets in Carteret's commercial real estate portfolio that were not securitized or part of participations sold for 11.1 percent below their lowest appraisal values from 1992 or later. IV. 19. CORRECTION OF TABLE 19 FROM PREVIOUS DECLARATION

Table 19 of my previous declaration included an arithmetic error in the estimated

value of Carteret's servicing portfolio.27 In that report I stated that "[c]ontemporaneous independent appraisals suggest a value of approximately $101.6 million for Carteret's $8.0 billion servicing rights alone as of September 30, 1991, a value of $63.0 million to $79.4 million for Carteret's purchased loan servicing rights of $5 billion as of September 1991, and a value of $46.1 million for servicing rights on Carteret's originated loans only as of June 1992."28 The sum of the larger estimates of Carteret's purchased and originated loan servicing rights is $125.5 million (equal to $79.4 million plus $46.1 million), whereas Table 19 in my previous report reported a value of $135.5 million. Table 5 is a correction of Table 19 from my previous report. TABLE 5: POTENTIAL SOURCES OF CAPITAL FOR CARTERET IN 1992 (CORRECTED TABLE 19 FROM DECLARATION OF CHARLES W. CALOMIRIS OF MAY 19, 2007)
Source Carteret Mortgage Co. Carteret's Servicing Portfolio Raising Funds in the Capital Markets Estimated Amount (Million $) $18.4 - $57.8 $101.6 - $125.5 N/A

27. Declaration of Charles W. Calomiris of May 19, 2007, at 63. 28. Declaration of Charles W. Calomiris of May 19, 2007, at 63-64.

CONFIDENTIAL MATERIALS ­ SUBJECT TO PROTECTIVE ORDER

Case 1:93-cv-00531-LAS

Document 226-3
- 14 -

Filed 11/07/2007

Page 15 of 16

I declare that the forgoing opinion is to the best of my knowledge a fully true and correct one. I reserve the right to update opinions in the report based on continuing research.

Dated: July 12, 2007

CONFIDENTIAL MATERIALS ­ SUBJECT TO PROTECTIVE ORDER

Case 1:93-cv-00531-LAS

Document 226-3
- 15 -

Filed 11/07/2007

Page 16 of 16

APPENDIX A: MATERIALS RELIED UPON
Declaration of Charles W. Calomiris of May 19, 2007 (and all documents relied upon). New Century Financial Corp., SEC Form 10-K405, Dec. 31, 1997. Executive Summary, Mortgage Securitization Case, RTC 1993-S1 Case (C-AM-A-0360076 ­ 83) RTC Multiple Investor Fund Case (RTC Mortgage Trust Series 1993-N3) (Aug. 17, 1993) (C-AM-A-0360029 ­ 36). Financial Data Warehouse ­ RRS Balance Sheet Drill Down Report for Period Ending Dec. 31, 1995, Account 071190. CAM918 1908 ­ 2014. Loan Participations (C-AM-A-0177423 ­ 480).

CONFIDENTIAL MATERIALS ­ SUBJECT TO PROTECTIVE ORDER