This is the accessible text file for GAO report number GAO-04-116 
entitled 'Farmer Mac: Some Progress Made, but Greater Attention to Risk 
Management, Mission, and Corporate Governance Is Needed' which was 
released on October 16, 2003.

This text file was formatted by the U.S. General Accounting Office 
(GAO) to be accessible to users with visual impairments, as part of a 
longer term project to improve GAO products' accessibility. Every 
attempt has been made to maintain the structural and data integrity of 
the original printed product. Accessibility features, such as text 
descriptions of tables, consecutively numbered footnotes placed at the 
end of the file, and the text of agency comment letters, are provided 
but may not exactly duplicate the presentation or format of the printed 
version. The portable document format (PDF) file is an exact electronic 
replica of the printed version. We welcome your feedback. Please E-mail 
your comments regarding the contents or accessibility features of this 
document to Webmaster@gao.gov.

This is a work of the U.S. government and is not subject to copyright 
protection in the United States. It may be reproduced and distributed 
in its entirety without further permission from GAO. Because this work 
may contain copyrighted images or other material, permission from the 
copyright holder may be necessary if you wish to reproduce this 
material separately.

Report to Congressional Requesters:

October 2003:

FARMER MAC:

Some Progress Made, but Greater Attention to Risk Management, Mission, 
and Corporate Governance Is Needed:

GAO-04-116:

GAO Highlights:

Highlights of GAO-04-116, a report to congressional requesters

Why GAO Did This Study:

In the late 1990s, GAO found that the Federal Agricultural Mortgage 
Corporation (Farmer Mac), a federal government-sponsored enterprise, 
had significant assets in nonmission investments and analyzed its long-
term viability. Recently, Congress asked GAO to report on Farmer Mac’s 
(1) financial condition, (2) mission, (3) corporate governance, and 
(4) oversight provided by the Farm Credit Administration (FCA).

What GAO Found:

Farmer Mac’s income has increased since 1999, and its capital 
continues to exceed required levels. At the same time, we identified 
trends that indicated a more complex risk profile. For example, its 
off-balance sheet standby agreements have grown 350 percent in 3 years 
and comprise nearly 50 percent of Farmer Mac’s total loans. To date, 
the underlying loans have been performing better than the on-balance 
sheet loans. However, if rapid growth in standby agreements continues, 
and Farmer Mac were to undergo stressful economic conditions, it could 
face substantial funding liquidity risk. Farmer Mac has risk 
management systems in place, but certain aspects of its risk 
management capacity have not kept pace with its increasingly complex 
portfolio. For example, the loans used in the loss estimation model 
have characteristics that differ from Farmer Mac’s portfolio both with 
respect to geographic distribution and interest rate terms. In 
addition, although Farmer Mac has maintained sufficient liquidity to 
support its loan purchase and guarantee activity, it has lacked a 
formal contingency plan to address potential liquidity needs under 
stressful agricultural economic conditions.

Since our 1999 report, Farmer Mac has significantly reduced the ratio 
of nonmission investments and correspondingly increased its mission 
activities—providing long-term credit to farmers and ranchers at 
stable interest rates. These activities include loan purchases, 
guarantees, and commitments related to agricultural mortgages. 
However, there is geographic and lender concentration in the loan and 
guarantee portfolio, and the overall impact of the activities on the 
agricultural real estate market is unclear. Farmer Mac’s enabling 
legislation lacks specific or measurable mission-related criteria that 
would allow for a meaningful assessment of its mission achievement. In 
addition, the depth and liquidity of the current market for 
agricultural mortgage backed securities (AMBS) is unknown because 
Farmer Mac’s strategy of holding AMBS has been a contributing factor 
in limiting the development of a liquid, secondary market for these 
securities.  

The Sarbanes-Oxley Act of 2002 and the proposed New York Stock 
Exchange (NYSE) listing standards are both applicable to Farmer Mac 
because its securities are publicly traded and listed on the NYSE. 
However, Farmer Mac’s efforts to meet the new standards regarding an 
independent board could be limited by its statutory board structure. 
Under its statute, two-thirds of the board’s directors are elected by 
institutions that have a business relationship with Farmer Mac and own 
the only two classes of voting stock. Since 2002, FCA enhanced 
oversight of Farmer Mac by performing a more thorough annual safety 
and soundness examination, and by proposing liquidity standards and 
regulatory limits for nonmission investments. However, FCA still faces 
challenges, including limitations in its tools to analyze capital and 
credit risk, as well as the lack of criteria and procedures to assess 
and report on Farmer Mac’s mission achievement.

What GAO Recommends:

GAO recommends that Farmer Mac improve its risk management and 
corporate governance practices, including improve its loan loss 
estimation model, and develop a contingency funding liquidity plan.

GAO also recommends that FCA improve the model that analyzes Farmer 
Mac’s credit risk and assess Farmer Mac’s impact on the agricultural 
real estate market.

GAO suggests that Congress consider legislative changes that would 
establish clearer, measurable mission goals for Farmer Mac; amend 
Farmer Mac’s board structure; and allow FCA to adjust capital 
standards for Farmer Mac.

Farmer Mac agreed with some GAO findings and recommendations and did 
not address others. FCA agreed with GAO’s findings and 
recommendations.

www.gao.gov/cgi-bin/getrpt?GAO-04-116 

To view the full product, including the scope and methodology, click 
on the link above. For more information, contact Ms. Davi M. 
D’Agostino or Ms. Jeanette Franzel at (202) 512-8678.

[End of section]

Contents:

Letter: 

Background: 

Results in Brief: 

Farmer Mac's Financial Condition Has Improved, but Risk Management 
Practices Have Not Kept Pace with Its More Complex Risk Profile: 

Mission-Related Activities Have Increased, but Impact of Activities on 
Agricultural Real Estate Market Is Unclear: 

Farmer Mac's Statutory Governance Structure Does Not Reflect Interests 
of All Shareholders and Some Corporate Governance Practices Need to Be 
Updated: 

FCA Has Taken Steps to Enhance Oversight of Farmer Mac, but Faces 
Challenges That Could Limit the Effectiveness of Its Oversight: 

Conclusions: 

Recommendations: 

Matters for Congressional Consideration: 

Agency Comments and Our Evaluation: 

Appendixes:

Appendix I: Objectives, Scope, and Methodology: 

Appendix II: Farmer Mac's Programs and Products: 

Appendix III: Financial Trends and Comparisons with Other Entities: 

Revenue Has Increased, but Some Financial Performance Indicators Lag 
Comparative Entities: 

Appendix IV: Farmer Mac's Underwriting Standards: 

Appendix V: Interest Rate Risk: 

Asset-Liability Management: 

Prepayment Model: 

Farmer Mac's IRR Measurement Process: 

Appendix VI: Farm Credit Administration Credit Risk Model: 

Data Limitations: 

Model Limitations: 

Appendix VII: Comments from the Federal Agricultural Mortgage 
Corporation: 

GAO Comments: 

Appendix VIII: Comments from the Farm Credit Administration: 

GAO Comments: 

Appendix IX: GAO Contacts and Staff Acknowledgments: 

GAO Contacts: 

Acknowledgments: 

Glossary of Terms:

Tables: 

Table 1: Loans Covered by Standby Agreements: 

Table 2: Comparison of Total Minimum Capital Levels Per $100 of Loans: 

Table 3: Annual Compensation and Options Granted for CEO's of Farmer Mac 
and Housing GSEs, 2002: 

Figures: 

Figure 1: Percentage of Outstanding Balance of Loans, AMBS and Standby 
Agreements, as of December 31, 2002: 

Figure 2: Long-term Interest Rates on Loans Secured by Agricultural Real 
Estate: 

Figure 3: Securitization Status of Farmer Mac I Portfolio, as of 
December 31, 2002: 

Figure 4: Farmer Mac Portfolio Exposure by Loan Origination Type: 

Figure 5: Farmer Mac I Geographic Concentration of Exposure by Region, 
as of December 31, 2002: 

Figure 6: Farmer Mac's Impaired Loans from 1997 to 2002: 

Figure 7: Farmer Mac's Nonperforming to Total Loans Compared to Other 
Entities, as of December 31, 2002: 

Figure 8: Income by Program Assets: 

Figure 9: Farmer Mac's ROA Compared to Other Entities: 

Figure 10: Farmer Mac's ROE Compared to Other Entities: 

Figure 11: Farmer Mac's Capital to Asset Ratios Compared to Other 
Entities: 

Abbreviations: 

ACA: Agricultural Credit Association:

AMBS: agricultural mortgage-backed securities:

CEO: chief executive officer:

Fannie Mae: Federal National Mortgage Association:

Farmer Mac: Federal Agricultural Mortgage Corporation:

FCA: Farm Credit Administration:

FCBT: Farm Credit Bank of Texas:

FCS: Farm Credit System:

FHA: Federal Housing Administration:

FHESSA: Federal Housing Enterprises Financial Safety and: Soundness 
Act:

FHFB: Federal Housing Finance Board:

FLCA: Federal Land Credit Association:

Freddie Mac: Federal Home Loan Mortgage Corporation:

GAAP: generally accepted accounting principles:

GSE: government-sponsored enterprise:

HUD: Housing and Urban Development:

IRR: interest rate risk:

LIBOR: London Interbank Offered Rate:

LTV: loan-to-value:

MTN: medium term-notes:

MVE: market value of equity:

NII: net interest income:

NYSE: New York Stock Exchange:

OFHEO: Office of Federal Housing Enterprises Oversight:

OSMO: Office of Secondary Market Oversight:

QRM: Quantitative Risk Management:

ROA: return on assets:

ROE: return on equity:

SAB: staff accounting bulletin:

SEC: Security Exchange Commission:

SFAS: Statement of Financial Accounting Standard:

SPI: subordinated participation interest:

USDA:

Letter October 16, 2003:

The Honorable Thad Cochran 
Chairman 
The Honorable Tom Harkin 
Ranking Minority Member 
Committee on Agriculture, Nutrition, and Forestry 
United States Senate:

The Honorable Richard G. Lugar 
United States Senate:

Farmer Mac is a government-sponsored enterprise (GSE)[Footnote 1] 
established by Congress to create a secondary market in agricultural 
real estate and rural housing loans, and improve the availability of 
agricultural mortgage credit. In 1998 and 1999, we found that a 
significant amount of Farmer Mac's assets were in nonmission 
investments and we discussed issues surrounding the long-term viability 
of Farmer Mac.[Footnote 2] Recently, you asked us to conduct a 
comprehensive review of Farmer Mac. This report discusses (1) Farmer 
Mac's current financial condition and risk management practices; (2) 
the extent to which Farmer Mac has achieved its statutory mission; (3) 
Farmer Mac's corporate governance as it pertains to board structure and 
oversight, and executive compensation; and (4) the Farm Credit 
Administration's (FCA) oversight of Farmer Mac.

To address these objectives, we analyzed trends in Farmer Mac's key 
indicators of financial performance and condition for fiscal year 2002-
-including measures of earnings and profitability, capital, liquidity, 
and its asset and liability mix--and determined how Farmer Mac has 
managed and measured the risks it faces--credit, liquidity, and 
interest rate risk.[Footnote 3] We reviewed documents and interviewed 
representatives from Farmer Mac, FCA, other market participants, and 
individuals with expertise in the agricultural real estate market. We 
analyzed Farmer Mac's loan portfolio growth. We obtained and reviewed 
FCA's previous examinations and its most recent examination of Farmer 
Mac and other consultants' studies related to Farmer Mac. We did not 
report specific details of Farmer Mac's investment and loan portfolio 
nor details of reports of auditors, consultants, and examiners because 
of the proprietary nature of such information.

We conducted our work in California, Indiana, New York, Virginia, and 
Washington, D.C., between August 2002 and May 2003 in accordance with 
generally accepted government auditing standards. Appendix I contains a 
detailed description of the scope and methodology of our work.

Background:

Farmer Mac, a GSE, was chartered by Congress in the Farm Credit Act of 
1971, as amended by the Agricultural Credit Act of 1987 (the 1987 
Act).[Footnote 4] It is a federally chartered and privately operated 
corporation that is publicly traded on the New York Stock Exchange. 
Farmer Mac is also an independent entity within the Farm Credit System 
(FCS), which is another GSE. As an FCS institution, Farmer Mac is 
subject to the regulatory authority of FCA. FCA, through its Office of 
Secondary Market Oversight (OSMO), has general regulatory and 
enforcement authority over Farmer Mac, including the authority to 
promulgate rules and regulations governing the activities of Farmer Mac 
and to apply its general enforcement powers to Farmer Mac and its 
activities. According to the 1987 Act, Farmer Mac, in extreme 
circumstances, may borrow up to $1.5 billion from the U.S. Treasury to 
guarantee timely payment of any guarantee obligations of the 
corporation.[Footnote 5]

Under the 1987 Act, Farmer Mac's mission is to provide for a secondary 
marketing arrangement for agricultural real estate and rural housing 
loans subject to its underwriting standards. A secondary market is a 
financial market for buying and selling loans, individually or by 
securitizing them. When loans are securitized, they are repackaged into 
a "pool" by a trust in order to be sold to investors. By returning cash 
to primary lenders in exchange for their loans, theoretically, a 
secondary market would generate additional funds for the lenders to 
lend and enhance the lenders' ability to manage credit and interest 
rate risk. Ideally, a Farmer Mac-sponsored secondary market would 
increase liquidity to lenders by providing the lenders access to 
national capital markets. This in turn would reduce regional imbalances 
in loanable funds and possibly increase the overall availability of 
credit to the primary agricultural real estate market and lower 
interest rates for borrowers.

To relieve structural impediments that had limited Farmer Mac's ability 
to function efficiently, Congress passed the Farm Credit System Reform 
Act of 1996 (the 1996 Act), which significantly revised Farmer Mac's 
statutory authority and had significant impact on Farmer Mac's 
operations.[Footnote 6] Among other things, the 1996 Act allowed Farmer 
Mac to (1) purchase agricultural mortgage loans directly from lenders, 
"pool" the loans, and issue and sell securities that are backed by 
these pools to investors and (2) eliminate the mandatory requirement 
for loan originators and poolers to retain 10 percent, first-loss 
subordinated participation interest (SPI) with each securitized loan 
pool.[Footnote 7]

Farmer Mac operates a cash window program where Farmer Mac purchases 
mortgages directly from lenders for cash and purchases bonds from 
agricultural lenders. (See app. II for Farmer Mac's programs and 
products.) Periodically, Farmer Mac transfers its purchased loans into 
trusts that it uses as vehicles for the securitization of those loans. 
Securitization is the transfer of assets (in this case, loans) to a 
third party or trust. In turn, the third party or trust issues 
certificates to investors. Farmer Mac refers to the certificates sold 
to investors as "guaranteed securities" or as agricultural mortgage-
backed securities (AMBS). The cash flow from the transferred loans 
supports repayment of the AMBS. Farmer Mac guarantees timely payments 
to investors holding the certificates, regardless of whether the trust 
has actually received such scheduled loan payments.

Farmer Mac loan programs are divided into two main groups referred to 
as Farmer Mac I and Farmer Mac II. Farmer Mac I consists of 
agricultural and rural housing mortgage loans that do not contain 
federally provided primary mortgage insurance. For loans underlying 
pre-1996 Act Farmer Mac I AMBS, 10-percent first-loss subordinated 
interests mitigate Farmer Mac's credit risk exposure. Before Farmer Mac 
incurs a credit loss, losses are first absorbed by the poolers' or 
originators' subordinated interest. As of December 31, 2002, Farmer Mac 
had not experienced any credit losses related to the pre-1996 Act 
Farmer Mac I AMBS, and the first-loss subordinated interests are 
expected to exceed the estimated credit losses on those loans. Current 
risks in Farmer Mac's loan and guarantee portfolio, such as those 
discussed later in this report, are generated primarily by post-1996 
guaranteed securities.

Farmer Mac receives an annual guarantee fee from the third party or 
trust involved based on the outstanding balance of the Farmer Mac I 
post-1996 guaranteed securities. During 2002, all AMBS sold were to 
Zions Bank, a related party of Farmer Mac, and totaled $47.7 million. 
Guarantee fees earned from Zions Bank were $1.0 million in 
2002.[Footnote 8]

Farmer Mac II consists of agricultural mortgage loans containing 
primary mortgage insurance provided by the U.S. Department of 
Agriculture (USDA). USDA-guaranteed loans collateralizing Farmer Mac II 
AMBS are backed by the full faith and credit of the United States. 
Similar to the pre-1996 Act securities, as of December 31, 2002, Farmer 
Mac had experienced no credit losses on any Farmer Mac II AMBS and did 
not expect to incur any such losses in the future.

Farmer Mac's long-term standby purchase commitments (standby 
agreements), introduced in 1999, represent a commitment by Farmer Mac 
to purchase eligible loans from financial institutions at an 
undetermined future date when a specific event occurs. This commitment 
represents a potential obligation of Farmer Mac that does not have to 
be funded until such time as Farmer Mac is required to purchase a loan. 
The specific events or circumstances that would require Farmer Mac to 
purchase loans under a standby agreement include when (1) an 
institution determines it will sell some or all of the loans under the 
agreement to Farmer Mac or (2) a borrower fails to make installment 
payments for 120 days on a loan covered by a standby agreement. 
Financial institutions effectively transfer the credit risk on the 
loans covered by a standby agreement to Farmer Mac. Consequently, these 
institutions' regulatory capital requirements and loss reserve 
requirements would then be reduced. To date, FCS institutions have been 
the only participants in standby agreements. In exchange for Farmer 
Mac's commitment under the standby agreement, Farmer Mac receives an 
annual commitment fee from institutions entering into these agreements, 
based on the outstanding balance of the loans covered by the standby 
agreement. In 2002, these fees represent a significant portion of 
Farmer Mac's total revenues.

Farmer Mac funds its loan purchases primarily by issuing debt 
obligations of various maturities. As of December 31, 2002, Farmer Mac 
had outstanding $2.9 billion of short-term discount notes and $1.0 
billion of medium-term notes. To the extent the proceeds of the debt 
issuances exceed Farmer Mac's need to fund program assets, those 
proceeds are used to purchase assets for the nonmission investment 
portfolio.

As of December 31, 2002, loans held by Farmer Mac and loans that either 
back Farmer Mac AMBS or are subject to standby agreements totaled $5.5 
billion. Nearly $3 billion of the $5.5 billion loan and guarantee 
portfolio is not on Farmer Mac's balance sheet. See figure 1 for a 
breakdown of the $5.5 billion loan and guarantee portfolio. As of 
December 31, 2002, Farmer Mac employed 33 persons.

Figure 1: Percentage of Outstanding Balance of Loans, AMBS and Standby 
Agreements, as of December 31, 2002:

[See PDF for image]

[End of figure]

Like any other private financial firm, Farmer Mac faces credit, 
liquidity, interest rate, and operations risks when conducting its 
secondary market operations. Farmer Mac is exposed to the following 
risks:

* Credit risk--the possibility of financial loss resulting from default 
by borrowers on farming assets that have lost value or other parties' 
failing to meet their obligations. Credit risk occurs when Farmer Mac 
holds mortgages in portfolio and when it guarantees principal and 
interest payment to investors in the AMBS it issues. Farmer Mac is also 
exposed to credit risk for the approximately $2.7 billion of loans 
under Farmer Mac standby agreements, which represent unconditional 
commitments to purchase performing loans at a market price, and to 
purchase120 day delinquent loans at par.

* Liquidity risk--the possibility or the perception that Farmer Mac 
will be unable to meet its obligations as they come due because of an 
inability to liquidate assets or obtain adequate funding (referred to 
as "funding liquidity risk") or will not be able to easily unwind or 
offset specific exposures without significantly lowering market prices 
because of inadequate market depth or market disruptions ("market 
liquidity risk").

* Interest rate risk--the potential that changes in prevailing interest 
rates will adversely affect on-balance sheet assets, liabilities, 
capital, income or expenses at different times in different amounts.

* Operations risk--the possibility of financial loss resulting from 
inadequate or failed internal processes, people and systems, or from 
external events.

As a GSE, the structure of Farmer Mac's board of directors was 
congressionally established. Its 15-member board of directors includes 
5 members elected by Class A stockholders that are banks, insurance 
companies, and other financial institutions, 5 members elected by Class 
B stockholders that are FCS institutions, and 5 members appointed by 
the president of the United States. Farmer Mac has a third class of 
common stock that is held by the general public, Class C, which does 
not have voting rights.

The federal government's creation and continued relationship with 
Farmer Mac has created the perception in financial markets that the 
government will not allow the GSE to default on its debt and AMBS 
obligations, although no such legal requirement exists. As a result, 
Farmer Mac can borrow money in the capital markets at lower interest 
rates than comparably creditworthy private corporations that do not 
enjoy federal sponsorship. During the 1980s, the federal government did 
provide limited regulatory and financial relief to Fannie Mae when the 
GSE was experiencing financial difficulties; and in 1987, Congress 
provided financial assistance to FCS.

Results in Brief:

Since 1999, Farmer Mac's financial condition has improved, but its risk 
management practices have not kept pace with its more complex risk 
profile. Farmer Mac's net income has steadily increased from $4.6 
million in 1997 to $22.8 million in 2002, for a total increase of 392 
percent. On the other hand, Farmer Mac's off-balance sheet standby 
agreements, which are commitments to purchase loans under specific 
circumstances, such as when a loan becomes 120 days delinquent, have 
grown 350 percent in 3 years to $2.7 billion and represent nearly 50 
percent of the total loans included in Farmer Mac's programs. Regarding 
the credit quality of the loans underlying current standby agreements, 
those loans have been performing better than the loans on Farmer Mac's 
balance sheet. While these standby agreements have fueled revenue 
growth, going forward, if this rapid growth continues, standby 
agreements could generate substantial funding liquidity risk under 
stressful economic conditions. Further, nonperforming, or impaired, 
loans have been increasing for Farmer Mac and totaled $75.3 million at 
the end of 2002 as compared to zero at the end of 1997. While Farmer 
Mac has substantially increased its allowance for loan losses and 
reserve for losses (loan loss allowance), the ratio of its allowance to 
its impaired loans has gone down by over 50 percent since December 31, 
1998. This indicates that Farmer Mac's impaired loans have increased at 
a faster rate than the increases in its loan loss allowance and may 
also indicate increasing credit risk. Nevertheless, forensic 
accountants retained by Farmer Mac Board's outside counsel concurred 
with Farmer Mac's methodology for estimating loan loss allowance.

Farmer Mac has risk management systems in place, such as underwriting 
standards for purchasing and guaranteeing loans (including loans 
underlying standby agreements), and has generally sound processes in 
place for estimating credit losses. However, Farmer Mac has not (1) 
consistently well documented the exceptions made to its loan 
underwriting and servicing procedures, (2) included the current 
characteristics of its loan portfolio in the loan loss estimation 
model, and (3) adequately documented the results of the model compared 
to actual portfolio and economic conditions, resulting in the increased 
possibility that management's objectives of minimizing credit risk have 
not been met. We make recommendations to Farmer Mac designed to enhance 
its loan loss estimation model and to improve its documentation of 
policies and procedures, and management's actions that relate to 
reducing credit risk. Regarding liquidity risk, Farmer Mac has 
maintained sufficient liquidity to support its loan purchase and 
guarantee activity through continued access to the capital markets. 
However, Farmer Mac lacks a formalized contingency plan to address its 
potential liquidity needs that could potentially be created by the 
standby agreements under stressful agricultural economic conditions. 
Although Farmer Mac issues debt securities for liquidity purposes, it 
is not required and it has decided not to obtain a credit rating from a 
nationally recognized statistical rating agency. As for interest rate 
risk, the methods employed by Farmer Mac to measure interest rate 
sensitivity appeared reasonable but we identified limitations with some 
elements of its prepayment methodology. In terms of capital, Farmer Mac 
exceeded the capital levels required by its statute and regulator but 
could improve its plan for capital adequacy. Specifically, it lacked a 
test for sufficiency in assessing its capital adequacy, other than its 
stated goal of meeting its statutory minimum and regulatory risk-based 
capital requirements. We make recommendations to Farmer Mac to develop 
a contingency funding liquidity plan, improve the quality of its 
prepayment model, and enhance its analysis of capital adequacy. 
Finally, Farmer Mac also faces some uncertainty involving its line of 
credit with the Department of the Treasury (Treasury). Specifically, 
while the legal opinion of Farmer Mac's outside counsel disagrees, 
Treasury has taken the position that it is not obligated to cover 
losses on AMBS held in Farmer Mac's portfolio.

Farmer Mac has increased its agricultural mortgage loan purchase and 
guarantee activity since our 1999 report, and has reduced the relative 
size of its nonmission portfolio. Nevertheless, its enabling 
legislation contains broad mission purpose statements and lacks 
specific or measurable mission-related criteria that would allow for a 
meaningful assessment of whether Farmer Mac had achieved its public 
policy goals. Farmer Mac's strategy of holding AMBS for profitability 
reasons has been a contributing factor in limiting the development of a 
liquid secondary market for these securities. As a result, the depth 
and liquidity of the demand for AMBS in the current market are unknown. 
Farmer Mac introduced the standby agreement program to provide greater 
lending capacity for agricultural real estate lenders. However, FCS 
institutions' increased use of standby agreements potentially reduces 
the sum of capital required to be held by FCS and Farmer Mac. Such a 
reduction in capital could be consistent with a reduction in risk if 
there were diversification at the secondary market level. However, as 
of December 31, 2002, 10 financial institutions generated 90 percent of 
Farmer Mac's business, and over 70 percent of the outstanding balance 
of Farmer Mac's loan portfolio was located in the Southwest and 
Northwest. Finally, the size of Farmer Mac's nonmission investment 
portfolio has decreased as a percentage of its total on-and off-balance 
sheet portfolio. Still, the composition and criteria for nonmission 
investments could potentially lead to investments that are excessive in 
relation to Farmer Mac's financial operating needs or otherwise would 
be inappropriate to the statutory purpose of Farmer Mac. We make 
recommendations to Farmer Mac to reevaluate its current strategy of 
holding AMBS in its portfolio and issuing debt to obtain funding. We 
also suggest that Congress consider legislative changes to establish 
clearer mission goals for Farmer Mac.

Like other publicly traded companies, Farmer Mac is in the process of 
taking actions to ensure that it complies with provisions of the 
Sarbanes-Oxley Act of 2002 (Sarbanes-Oxley) requirements, the Security 
Exchange Commission (SEC) rules, and proposed changes in the New York 
Stock Exchange (NYSE) listing standards. In accordance with these new 
requirements, Farmer Mac has reaffirmed its audit committee charter and 
has hired internal and external auditors that are from different firms. 
Sarbanes-Oxley requires that members of audit committees of listed 
companies be independent and proposed NYSE listing standards require 
that a majority of the board of directors of listed companies be 
independent. Since Farmer Mac's Class A and Class C stock are listed on 
the NYSE, Farmer Mac is currently subject to the auditor independence 
requirements of Sarbanes-Oxley and, unless waived, the listing 
standards. Farmer Mac has taken steps to update its corporate 
governance practices, but its statutory board structure, which is set 
by law, could make it difficult to comply with the board independence 
requirements proposed in the NYSE listing standards. Moreover, since 
Farmer Mac shareholders include both institutions that utilize its 
services (Class A and Class B common stock) and public investors (Class 
C common stock), and because all members of the board of directors are 
chosen by the cooperative investors or by the President of the United 
States, the board may face difficulties in representing the interests 
of all shareholders. Additionally, since Farmer Mac is a publicly 
traded company, the nonvoting structure of Farmer Mac's Class C common 
stock may not be appropriate in today's corporate governance 
environment. In most respects, Farmer Mac's board policies and 
processes appear reasonable, but the process to identify and select 
nominees, director training, and succession planning could be further 
developed and formalized. Finally, Farmer Mac's total executive 
compensation was within its consultants' recommended parameters; 
however, its vesting program appears more generous than industry 
practices, given Farmer Mac's maturity. We make recommendations to 
Farmer Mac designed to provide more transparency to the nomination 
process and succession planning and more consistency in training for 
directors. We also recommend that Farmer Mac reevaluate the vesting 
period for stock options. We further suggest that Congress consider 
legislative changes to amend the structure of the Farmer Mac board and 
the structure of Farmer Mac's Class C common stock.

Since 2002, FCA took several steps to enhance supervisory oversight of 
Farmer Mac but faces significant challenges that could limit the 
effectiveness of its oversight. FCA's June 2002 annual safety and 
soundness examination was more comprehensive than previous 
examinations. FCA also has taken some actions to improve its regulatory 
framework for Farmer Mac by developing proposed regulations regarding 
liquidity standards and nonmission investments. Although FCA has 
increased its efforts to help oversee and examine Farmer Mac's 
operations, our review identified weaknesses in FCA's off-site 
monitoring process regarding call reports. As it continues to oversee 
Farmer Mac, FCA faces five significant challenges related to Farmer 
Mac's risk-based capital model as well as regulatory management. First, 
limitations exist in FCA's model used to estimate Farmer Mac's credit 
risk for calculation of the risk-based capital requirement. 
Individually, each limitation may under or over estimate the risk-based 
capital for Farmer Mac's credit risk, but overall, the relative 
magnitude of these effects is unclear. Second, FCA's risk-based capital 
regulation does not capture credit risk on Farmer Mac's liquidity 
investments, AgVantage bonds, and counterparty risk on derivatives. 
Third, FCA's market risk and income models may understate estimated 
levels of required risk-based capital. Fourth, FCA does not have 
criteria and procedures to assess and report on the relationship of 
Farmer Mac's activities to the achievement of its mission. Finally, 
being the single regulator for both FCS institutions and Farmer Mac 
could cause potential conflicts of interest because FCA may, in times 
of stress, attempt to support one type of participant at the expense of 
the other. We make recommendations to FCA designed to enhance the risk-
based capital model, improve off-site monitoring of Farmer Mac, and 
assess and report on how Farmer Mac is achieving its mission. We also 
suggest that Congress consider a legislative change to allow FCA more 
flexibility in setting minimum capital requirements for Farmer Mac.

We provided a draft of this report to the heads or their designees of 
the Farmer Mac, FCA, SEC, and Treasury. We received written comments 
from Farmer Mac and FCA that are reprinted in appendixes VII and VIII 
respectively. SEC did not provide comments. Farmer Mac, FCA, and 
Treasury also provided technical comments that we have incorporated as 
appropriate. Farmer Mac stated that it agreed with the report's 
findings and conclusions on Farmer Mac's risk management practices and 
has taken a number of steps toward implementing the majority of the 
recommendations. While Farmer Mac seemed to agree with the report's 
recommendations to improve its analysis of capital adequacy, develop a 
contingency funding plan, and improve documentation of management 
exceptions to its eight major underwriting standards, it did not 
address the rest of our recommendations. Farmer Mac commented that the 
uncertainty regarding the Treasury line of credit is a moot point 
because a legal opinion by its outside counsel stated that the Treasury 
line of credit would be available in the circumstances noted. Our 
position is that this issue may remain unresolved until Farmer Mac 
approaches Treasury for assistance. Farmer Mac appeared to disagree 
with our concern about funding liquidity risk that might arise from 
standby agreements. However, we noted that Farmer Mac seems to believe 
that liquidity funding risk is captured and accounted for in the risk-
based capital model, whereas it is not. Moreover, Farmer Mac has not 
fully recognized all loan amounts that could be presented to Farmer Mac 
for funding as part of its liquidity funding needs. FCA generally 
concurred with the report's findings and conclusions that are focused 
on FCA's work to oversee the safety and soundness of Farmer Mac and 
agreed to implement the report's recommendations. FCA does not agree 
that additional data and modeling would add value to the risk-based 
capital model, although FCA stated that it is studying the possibility 
of updating the data used in the model. We discuss Farmer Mac and FCA's 
comments and our response in greater detail at the end of this letter.

Farmer Mac's Financial Condition Has Improved, but Risk Management 
Practices Have Not Kept Pace with Its More Complex Risk Profile:

Farmer Mac's net income has steadily increased from $4.6 million in 
1997 to $22.8 million in 2002, for a total increase of 392 percent. At 
the same time, we identified trends that increased the complexity of 
Farmer Mac's risk profile, such as rapid growth in its standby 
agreement program, certain weaknesses in its risk management practices, 
and an uncertainty involving Treasury's line of credit.

Farmer Mac's Income Has Increased, and Risk Profile Has Become More 
Complex:

Two primary revenue sources contributed to the growth of Farmer Mac's 
net income--interest income and commitment fees. Interest income earned 
on Farmer Mac's portfolio of loans, guaranteed securities, and 
investments has more than doubled due to substantial growth in Farmer 
Mac's portfolios over the same period. Interest income was Farmer Mac's 
principal source of revenue in 1997. But recently, a new source of 
revenue--commitment fees earned on standby agreements--has grown since 
the product's inception in early 1999, amounting to over 25 percent of 
Farmer Mac's total revenues for 2002. See appendix III for further 
discussion of trends and comparisons of Farmer Mac's financial 
condition.

High Growth in Standby Agreements Fuels Revenue, but Could Generate 
Funding Liquidity Risk Under Stressful Conditions:

Although Farmer Mac's net income has been increasing since 1997, there 
could be indicators of increasing funding liquidity risk due to high 
levels of growth in Farmer Mac's standby agreement program. Farmer 
Mac's earnings growth has principally been driven by its off-balance 
sheet standby agreements. First offered in early 1999, the standby 
agreement program grew rapidly. As shown in table 1, the balance of 
loans covered by standby agreements grew from zero in 1998 to $2.7 
billion at December 31, 2002, with high rates of growth in recent 
years.

Table 1: Loans Covered by Standby Agreements:

Loans covered by standby agreements; 1998: $0; 1999: $0.6 billion; 
2000: $0.9 billion; 2001: $1.9 billion; 2002: $2.7 billion.

Percentage increase from previous year; 1998: N/A; 1999: N/A; 2000: 
50%; 2001: 111%; 2002: 42%.

Source: Farmer Mac 2002 SEC 10K filing.

[End of table]

Corresponding with the growth in loans covered under the standby 
agreement program is an increase in Farmer Mac's revenues. For 
instance, revenues from commitment fees were $1.6 million in 1999 or 7 
percent of Farmer Mac's total revenues that year. By 2002, revenues 
from commitment fees had increased to $11.0 million, representing 22 
percent of total revenues. During economic times when agricultural land 
values have been rising and interest rates have been relatively low, 
Farmer Mac has, in 2002, purchased about $3.3 million of the eligible 
agricultural mortgage loans placed under standby agreements.

Farmer Mac stated that since the program began in 1999, the relatively 
few defaulted loans they have had to purchase reflect the credit 
quality of the loans underlying standby agreements. Further, those 
loans are underwritten and required to be serviced to the same 
standards used for all other loans backing Farmer Mac's AMBS. Between 
1999 and 2002, the standby agreements had no net credit losses. At 
December 31, 2002, loans that were at least 90 days delinquent under 
standby agreements were $3.5 million or 0.13 percent of the total 
amount of loans under standby agreements. This was well below the $54.7 
million or 2.21 percent of Farmer Mac's on-balance-sheet loans and 
guarantees. The lower delinquencies and losses under the standby 
agreement program indicate that the program through December 31, 2002, 
experienced lower credit risk than Farmer Mac's other programs.

However, guidance from financial regulators indicates rapid growth of 
programs or assets is thought to be an increased risk factor. Many 
financial institution failures of past decades were blamed, in part, on 
unchecked growth particularly in new and innovative products with 
complicated risk characteristics. The rapid growth of the standby 
agreements could result in increased funding liquidity risk to Farmer 
Mac because Farmer Mac's commitment under these agreements differ from 
its off-balance sheet AMBS. For AMBS, if an underlying loan becomes 90 
days delinquent, Farmer Mac has the option of purchasing the loan, or 
just making installment payments. Under its standby agreements, if an 
underlying loan is 120 days delinquent, the lender can require Farmer 
Mac to buy the loan. Therefore, standby agreements represent a 
potential future obligation of Farmer Mac, which does not have to be 
funded until such time as Farmer Mac is required to purchase an 
impaired loan.[Footnote 9] In other words, going forward, if the rapid 
growth of standby agreements continue, at a time when either the 
agricultural sector is severely depressed or interest rates are 
adversely changing, Farmer Mac could be required to purchase large 
amounts of impaired or defaulted loans under the standby agreements, 
thus subjecting Farmer Mac to increased funding liquidity risks and the 
potential for reduced earnings. (See liquidity section for further 
discussion.):

Additionally, because of its rapid growth and recent implementation, 
there is limited historical information to project the number of loans 
covered by standby agreements that Farmer Mac may need to purchase in 
the future. As a result, management has limited quantitative data on 
which to base risk management and other operating decisions.

Similar to Farmer Mac's other guaranteed obligations, when Farmer Mac 
is required to purchase an impaired or defaulted loan under its standby 
agreement obligation, it may adversely affect its earnings in four 
ways: (1) it requires an earning asset to be sold or a liability to be 
incurred in exchange for an asset that might not be an earning asset; 
(2) it increases administrative expenses for monitoring, collection, 
and recovery efforts; (3) the annual commitment fees Farmer Mac 
receives on the loan would cease; and (4) under economically stressful 
conditions, Farmer Mac could incur losses on the disposal of impaired 
loans it is required to purchase under the standby agreements if the 
net proceeds from the sale of collateral on the loan is insufficient.

Increase in Impaired Loans and Charge-offs May Indicate Increasing 
Credit Risk:

Farmer Mac has established an allowance for loan losses and reserve for 
losses (a loan loss allowance) to cover estimated, probable loan losses 
for its current portfolio of loans, commitments, and guarantees. The 
loan loss allowance has increased substantially from $1.6 million at 
December 31, 1997, to $19.4 million at December 31, 2002. On the other 
hand, the ratio of loan loss allowance to impaired loans has decreased 
from a high of 59.1 percent at December 31, 1998, to 27.8 percent at 
December 31, 2002. This ratio, a primary credit risk indicator, shows 
that the balance of Farmer Mac's impaired assets has increased at a 
faster rate than the increases in its loan loss allowance. (We further 
discuss management's monitoring and assessment of credit risk in a 
later section of this report.):

Farmer Mac's increase in impaired loans, real estate owned, and write 
offs of bad loans as well as the growth in its on-and off-balance sheet 
loans, guarantees, and standby agreements is indicative of increasing 
credit risk. Impaired loans totaled $75.3 million at December 31, 2002, 
compared to zero at December 31, 1997.[Footnote 10] Since Farmer Mac 
only began purchasing loans after the 1996 Act was passed, the loan and 
guarantee portfolio is relatively new. As loan portfolios age, 
delinquencies typically increase, eventually peak, and then taper off, 
establishing a track record of performance often referred to in the 
industry as a "seasoned" portfolio. According to Farmer Mac, its loans 
are just becoming seasoned, so the losses and delinquencies are 
increasing. Farmer Mac's write offs of impaired loans have been limited 
to date but delinquencies are increasing. During 2002, Farmer Mac wrote 
off $4.1 million of bad loans, or 8 basis points[Footnote 11] of post-
1996 Act loans and guarantees, which was a significant increase over 
the $2.2 million, or 6 basis points, written off in 2001.[Footnote 12]

Farmer Mac's Controls Over Credit Risk Were Generally Sound but Had 
Certain Weaknesses:

Although Farmer Mac has underwriting standards for purchasing and 
guaranteeing loans (including loans underlying standby agreements), and 
has processes for estimating credit losses, Farmer Mac's implementation 
of its standards and its processes need improvement to enhance its 
overall controls over credit risk. One of its underwriting standards 
permits management to override one or more of the other eight standards 
when, in management's opinion, other factors compensate for certain 
loan weaknesses. Farmer Mac has made use of this provision without 
consistently and thoroughly documenting the basis for the exceptions 
made. For loans it has purchased, including loans under standby 
agreements, Farmer Mac's process for estimating credit losses is 
generally sound but has certain weaknesses. In estimating losses, 
Farmer Mac uses a risk model based on loans that differ from those in 
its own portfolios and under its standby agreements with respect to 
geographic distribution and interest rate terms. This lack of 
comparability and other limitations of the model may affect the 
reasonableness and accuracy of Farmer Mac's estimated losses resulting 
from credit risk either upward or downward. In estimating the credit 
risk of loans under standby agreements, a complicating factor is that 
Farmer Mac lacks the historical experience with the long term standby 
agreements needed to accurately estimate the types of loans and amount 
of loans it may ultimately be obligated to purchase, along with any 
associated losses. In addition, for estimating and allocating loan 
losses, Farmer Mac reverses the order of the methods called for in 
accounting guidance and does so without quantifying the effects of its 
approach. Finally, recent reviews have shown weak documentation 
describing Farmer Mac's use of its loan loss estimation model, its 
quantification process, management's judgment and key decisions, and 
the summary results of the loss estimation process.

Farmer Mac's Loan Underwriting and Servicing Procedures Were Clear, but 
Exceptions Were Not Consistently Well Documented:

Farmer Mac uses underwriting standards and processes for monitoring the 
loans it purchases and guarantees (including those loans under its 
standby agreements). It also has standards for "sellers" and loan 
"servicers." Farmer Mac's underwriting process includes identifying and 
analyzing potential risks of loss associated with its loan purchases 
and guarantees prior to entering into such agreements. A key element of 
Farmer Mac's system of internal control in underwriting is the use of 
established, written standards (for both internal use and for external 
loan sellers and servicers) that require analysis of numerous 
qualitative and quantitative borrower and property characteristics for 
loans, prior to purchase or prior to inclusion in a standby agreement. 
These standards help streamline the process for buying and guaranteeing 
loans, lower transaction costs, and increase efficiency while providing 
criteria and controls over the process of accepting loans for purchase 
or for inclusion in standby agreements. For example, Farmer Mac has 
underwriting standards as documented in its Seller/Servicer Guide (the 
guide) to (1) assess whether a borrower has sufficient income and a 
good credit history and (2) set a maximum loan-to-value ratio (LTV) 
limit. Farmer Mac monitors its credit risk through periodic monitoring 
of the borrower's and seller/servicer's performance by reviewing the 
payment history, visiting borrower and servicers' facilities, and in 
the case of seriously delinquent loans with expected loss in collateral 
value, obtaining updated property inspections and valuations.

As shown in appendix IV, Farmer Mac has nine underwriting standards to 
which all loans must conform, in order for Farmer Mac to purchase or 
guarantee the loans. Underwriting standard number nine allows 
management to override one or more specific underwriting criteria when, 
in management's opinion, other factors compensate for certain loan 
weaknesses. For example, in cases when the borrower's debt-to-asset 
ratio may not meet standards but the LTV ratio is better than 
requirements, then credit risk could be balanced by the LTV ratio. As 
of December 31, 2002, a total of $1.4 billion (30 percent) of the 
outstanding balance of loans held and loans underlying standby 
agreements and post-1996 Farmer Mac I Guaranteed Securities were 
approved based upon compensating strengths. Further, during 2002, 
$327.7 million (28 percent) of the loans purchased or added under 
standby agreements were approved based upon compensating strengths.

However, recent reviews noted that management's assessments supporting 
the override of underwriting criteria, including quantification and 
evaluation of compensating risk factors, was often not well-documented. 
Without consistently well-documented reasons for exceptions to the 
underwriting standards, Farmer Mac increases the risk that management's 
objectives of balancing risk have not been met. During 2003, Farmer Mac 
has begun gathering related data, but has not yet developed a process 
for fully utilizing the data in its management decision process for 
making future overrides and for estimating credit risk and allowance 
for losses on those specific loans.

Weaknesses Exist in Farmer Mac's Monitoring and Assessment of Changes 
in Credit Risk:

As part of the financial monitoring and reporting process, Farmer Mac's 
management is responsible for assessing the current level of risk 
associated with individual loans and loan portfolios that have been 
purchased or guaranteed by Farmer Mac, including loans under its 
standby agreements that are off-balance sheet, and estimating credit 
losses on those loans for financial reporting purposes. Farmer Mac 
records its estimated losses on loans held in an "allowance for loan 
losses" account, which serves to reduce the balance of Farmer Mac's 
loans. Farmer Mac estimates credit losses on loans backing its 
guaranteed securities and loans covered by its standby agreements and 
records those losses in "reserve for losses," which appears as a 
liability on Farmer Mac's balance sheet. When Farmer Mac records 
estimated losses in the allowance for loan loss and reserve for losses 
accounts, Farmer Mac's pretax income, and therefore its core capital, 
is reduced.

Farmer Mac uses a credit risk modeling tool called the Loan Pool 
Simulation and Guarantee Fee Model (the model) as a basis for 
estimating loan losses each quarter. This model, developed by an 
outside consultant, uses equations to estimate the probability, 
amounts, and distribution of losses over a period of time based upon 
loss experience from the Farm Credit Bank of Texas (FCBT) from 1979 to 
1992. Because Farmer Mac does not have adequate historical experience 
from its own portfolio for estimating losses on loans, data from FCBT 
are used as a proxy. According to Farmer Mac management, this was the 
best data available for estimating Farmer Mac's future losses on loans. 
The resulting projections of losses were reviewed by Farmer Mac 
management prior to being recorded to the financial statements. While 
this was the best data available, we did find a number of limitations 
in Farmer Mac's loan loss estimation model, its data, and application 
of the results to estimate losses, which may impact the reasonableness 
of the allowance and reserve amounts, and related losses recorded in 
the company's financial statements. We further discuss the data 
limitation in the FCA oversight section of this report since FCA also 
used FCBT data in its model to estimate Farmer Mac's credit risk.

The model used by Farmer Mac to estimate credit risk has some 
limitations. The primary limitation of the model is that Farmer Mac's 
loan and guarantee portfolios and the loans included under standby 
agreements have different characteristics from the loan characteristics 
of FCBT loans used in the model. Although the loans used in the model 
have similar characteristics with respect to key underwriting 
variables, they differ from Farmer Mac's portfolio both with respect to 
geographic distribution and interest rate terms. Specifically, the data 
supporting Farmer Mac's loan loss estimation process include loans 
issued in the 1970s and 1980s by FCBT, which were adjustable-rate 
mortgages, tied to a farm credit cost of funds index that changed 
slowly over time. In contrast, the loans now held and guaranteed by 
Farmer Mac are either rapidly changing adjustable-rate mortgages, or 
fixed-rate mortgages with financial penalties to the borrowers that 
eliminate the incentive to refinance when interest rates drop. 
Additionally, the FCBT loans were limited to Texas, while Farmer Mac 
may purchase loans in any state.

There are other complicating factors. First, Farmer Mac's current 
portfolio has a high geographic concentration in the Western part of 
the United States and is dominated by three lenders. Moreover, Farmer 
Mac's estimation of credit risk for the loans under standby agreements 
is limited by Farmer Mac's lack of historical experience for estimating 
the amount of loans it may ultimately be obligated to purchase under 
the standby agreements.

Farmer Mac has not quantified the impact of its current approach for 
estimating and allocating loan losses versus the approach set forth in 
accounting standards as the preferred methodology. SEC Staff Accounting 
Bulletin (SAB) No. 102 (July 6, 2001), states that "A registrant's loan 
loss allowance methodology generally should…identify loans to be 
evaluated for impairment on an individual basis under SFAS No. 114 and 
segment the remainder of the portfolio into groups of loans with 
similar risk characteristics for evaluation and analysis under SFAS No. 
5."[Footnote 13] This same approach is also set forth in a current 
American Institute of Certified Public Accountants proposed Statement 
of Position on Allowance for Credit Losses dated June 19, 2003.

Farmer Mac's calculation of its estimated loan loss allowances uses the 
reverse order of the approach set forth in the accounting standards as 
clarified in SAB 102. Farmer Mac's model calculates an overall loss 
result, from which management allocates portions to the allowance for 
losses (related to loans held by Farmer Mac) and the reserve for losses 
(related to loans guaranteed by Farmer Mac and included in its standby 
agreements). From the overall loss amounts calculated, Farmer Mac 
deducts specifically identified loan loss estimates and considers the 
remaining amount to be sufficient to cover the remainder of the 
portfolio. Farmer Mac's management stated that its methodology does not 
result in a materially different loss estimate than if it followed the 
preferred methodology of the accounting standards. However, Farmer Mac 
has not quantified the effects of using this methodology.

Documentation on the Loan Loss Estimation Model Was Weak:

Reviews of Farmer Mac conducted in 2002 concluded that Farmer Mac had 
weak documentation describing (1) how its loan loss estimation model 
works, (2) its quantification process, (3) management's judgment and 
key decisions, and (4) the summary results of the loss estimation 
process. Although Farmer Mac received an unqualified ("clean") opinion 
on its 2002 annual financial statements, Farmer Mac received several 
recommendations as a result of recent reviews to improve the loan loss 
estimation process, such as applying the model's results consistently 
with management's policies and improving documentation. During 2002, 
management took a number of actions in response to these 
recommendations to improve the data used for estimating losses as well 
as the disclosure of the risks inherent in its portfolio. In addition, 
to assess the reliability of Farmer Mac's estimated losses on loans, 
the Board of Directors' outside counsel retained a forensic accounting 
firm in 2002 to review management's processes and controls for 
estimating these losses. Nevertheless, it suggested improvements for 
Farmer Mac's SEC annual and quarterly filings and for internal 
documentation. Similarly, reports of recent reviews noted that 
management should document (1) the similarities and differences of 
using the model for both loans and guarantees recorded on the balance 
sheet as well as standby agreements that were not recorded on the 
balance sheet; (2) management's reconciliation of the model's loss 
projections to actual amounts recorded in the financial statements; and 
(3) the results of updated collateral evaluations and reviews of 
impaired loans, and the results' effect on the recorded allowance and 
reserve amounts.

Farmer Mac Maintained Access to Capital Markets, Its Primary Source of 
Liquidity, but It Lacked a Formal Liquidity Contingency Plan:

Farmer Mac maintained access to the capital markets, which are its 
primary source of liquidity, to support its loan purchase and guarantee 
activity, despite the lack of a credit rating that would make Farmer 
Mac's debt more comparable to other firms' debt issuances. Farmer Mac's 
reserve of liquid assets was a secondary source of liquidity, which as 
of September 30, 2002, was adequate to pay off current on-balance-sheet 
liabilities for close to 30 days.[Footnote 14] However, Farmer Mac 
lacked a formal contingency plan for potential liquidity funding needs 
under stressful agricultural economic conditions, including unexpected 
demands for additional liquidity that the standby agreements may 
create.

Farmer Mac Issued Debt Securities for Liquidity, but Has Not Pursued a 
Credit Rating To Date:

Our analysis indicated that Farmer Mac had been able to maintain 
access, at stable interest rates, to the discount note market, even 
during several periods of market stress and company exposure to public 
criticism in 2001 and 2002.[Footnote 15] However, these events 
temporarily affected the interest rates on medium-term notes.[Footnote 
16] Farmer Mac obtained cash for its loan purchase activities primarily 
through periodic sales of debt securities at varying maturities. 
Referring to publicly traded firms, Moody's Investors Service (Moody's) 
said that Farmer Mac was the largest issuer of unrated debt in the 
United States.[Footnote 17] Yet, Farmer Mac has issued discount notes 
at virtually the same interest rates as Fannie Mae, which obtains an 
annual "risk to the government" or financial strength rating from a 
nationally recognized rating agency.[Footnote 18] Broker-dealers who 
trade agency securities said that a cause was that (1) Farmer Mac has a 
GSE charter just as Fannie Mae and Freddie Mac do and, therefore, 
investors tend to conclude that they have a similar risk profile and 
(2) investors purchase Farmer Mac's discount notes to diversify 
portfolios that also held Fannie Mae and Freddie Mac short-term 
debt.[Footnote 19] Farmer Mac officials noted that the spreads on debt 
issuances are driven by the relatively small size of Farmer Mac 
issuances relative to the other GSEs, and at this time, the financial 
and human resources required to obtain a rating would not be 
justifiable. While having a credit rating may not have an effect on the 
interest rates on Farmer Mac's debt, such a rating would provide 
investors and creditors with information to assess Farmer Mac's 
financial soundness without government backing. This would facilitate 
investors and creditors comparing Farmer Mac with other entities and 
might also broaden the population of potential purchasers of Farmer 
Mac's debt securities, in particular municipalities, who purchase debt 
securities, due to internal policies that prohibit purchasing unrated 
financial instruments.

Farmer Mac Has Maintained Close to 30 Days of Liquidity:

Farmer Mac's liquidity investment portfolio was a secondary source of 
liquidity and provided for close to 30 days of funds should access to 
capital markets be temporarily impaired. As a comparison, Farmer Mac's 
reserve was larger than FCA's requirement that FCS institutions 
maintain a liquidity reserve of at least 15 days, although FCA 
officials said that they were evaluating the adequacy of a 15-day 
liquidity reserve.[Footnote 20] On the other hand, Farmer Mac's 
liquidity reserve of 15 days is considerably less than the stated 
liquidity goals of Fannie Mae, which maintained 3 months of liquidity 
to ensure that it could meet all of its obligations in any period of 
time in which it did not have access to the capital markets.[Footnote 
21] As of September 30, 2002, Farmer Mac's liquidity portfolio was 
worth $1.4 billion and consisted primarily of high-quality, short-term 
investments. However, according to our review of SEC filings, the range 
of permissible investments set by the board has expanded to include 
investments that do not have characteristics of traditional liquidity 
investments. For example, Farmer Mac's investment in a significant 
amount of unrated preferred stock of two FCS institutions represents 
fixed-rate investments that carry the potential for increased return, 
but also increased risk.

Farmer Mac Lacked a Formal Contingency Plan for Liquidity:

Farmer Mac does not yet have a formal contingency plan to maintain 
liquidity should its access to the capital markets be impaired, 
although as previously discussed, it does maintain a large liquidity 
portfolio to temporarily meet liquidity needs. In addition, management 
has standard written repurchase agreements with large investment banks, 
which it could use to pledge or sell its assets as a temporary source 
of liquidity.[Footnote 22] As of early 2003, Farmer Mac was in the 
process of developing a liquidity policy. Because Farmer Mac primarily 
relies on external sources of funds, Farmer Mac is exposed to funding 
liquidity risk and its access to these external funds could potentially 
be impaired by external or internal events.[Footnote 23] For example, 
in 2002, Farmer Mac increasingly relied on issuing discount notes for 
liquidity, as discount notes in combination with interest rate swaps 
would provide the lowest interest costs.[Footnote 24] According to 
financial regulatory guidance, for safety and soundness purposes, an 
effective plan for managing liquidity risk should not necessarily 
employ the cheapest source of funding. In addition, each institution's 
liquidity policy should include a contingency plan for liquidity, which 
would address alternative funding sources if initial projections of 
funding sources and uses were incorrect. The contingency plan would 
clearly identify any back-up facilities (lines of credit), and note the 
conditions where they might be used.

Off-balance Sheet Standby Agreements Can Potentially Create Unexpected 
Demands for Additional Funding Liquidity:

In addition to meeting liquidity demands from expected obligations, 
Farmer Mac may face unexpected demands on funding liquidity should 
lenders that participate in the standby agreements exercise their 
contracts. To date, Farmer Mac has not experienced material demands for 
additional liquidity that might arise from standby agreements and under 
current circumstances, Farmer Mac appears to have adequate liquidity to 
fund purchases of those underlying loans. However, the risk exists that 
if standby agreements continue to grow and their risks are not closely 
managed, during an economic downturn, Farmer Mac could experience a 
large and sudden increase in the exercise of standby agreements by 
lenders. In the event that Farmer Mac would be required to purchase 
large amounts of impaired or defaulted loans underlying the standby 
agreements, Farmer Mac management said that its strategy would be to 
rely on the capital markets for additional cash by either issuing more 
debt or selling its AMBS. However, since Farmer Mac did not sell AMBS 
to independent third party investors in 2002, the depth and liquidity 
of the demand for these securities in the current market are 
unknown.[Footnote 25] Broker-dealers with whom we spoke, stated that a 
Farmer Mac entrance into the debt markets to sell a significant amount 
of debt (in addition to what they currently issue) would require 
substantial investor education by Farmer Mac to generate additional 
interest in their debt securities.

Farmer Mac Managed Its Interest Rate Risk, but Elements of Its 
Prepayment Model Have Limitations:

Our discussions with Farmer Mac officials, reviews of Farmer Mac and 
FCA documents, and analysis of data from SEC filings indicated that as 
of December 31, 2002, Farmer Mac effectively managed its interest rate 
risk through a combination of yield maintenance clauses in loan 
contracts and through asset-liability matching; however, we found that 
prepayment model limitations could affect Farmer Mac's interest rate 
risk measurement.[Footnote 26] We observed that Farmer Mac has placed 
reliance on its ability to issue discount notes matched to interest 
rate swap transactions. Because discount notes are short-term 
liabilities and the majority of Farmer Mac's assets are longer term, a 
potential mismatch of interest rates could occur. Moreover, the 
retained portfolio strategy has increased the amount of interest rate 
risk that Farmer Mac must manage.[Footnote 27] By holding AMBS on its 
balance sheet, Farmer Mac retains and therefore must manage the 
interest rate risk in addition to the credit risk associated with AMBS. 
If Farmer Mac sold the AMBS to investors, it would only retain and have 
to manage the credit risk associated with AMBS. However, much of the 
concern relating to interest rate risk is mitigated through Farmer 
Mac's use of callable debt and interest rate swaps, which have the 
effect of adjusting the net interest payments to closely match the 
interest characteristics of Farmer Mac's assets. (See appendix V for 
further discussion.):

Farmer Mac measured and reported interest rate risk based on parameters 
set by board policy as follows. Farmer Mac's principal metrics for 
analyzing interest rate risk are:

* market value of equity (MVE)-at-risk calculation, which represents 
the current economic value of the firm;[Footnote 28]

* net interest income (NII) forecast, which represents the change in 
earnings relative to changes in interest rates; and:

* duration gap calculation, which measures the interest rate mismatch 
between Farmer Mac' assets and liabilities.[Footnote 29]

For further discussion of Farmer Mac's interest rate risk measurement 
process, see appendix V.

During 2002 Farmer Mac managed its MVE within board-approved limits, 
with one exception. NII was also managed within the board-approved 
range. The duration gap, which is measured in months, widened from--0.8 
months in December 2001 to--3.6 months in December 2002, as loan 
prepayments increased as interest rates declined, and these figures 
were still within the range of the board-approved parameters.

We found that Farmer Mac had a reasonable process and tools to measure 
interest rate risk, but the quality of its risk measurement is 
potentially limited by elements of its prepayment model. Prepayment 
models are an important component of interest rate risk measurement. 
Since Farmer Mac has prepayment penalties or yield maintenance terms on 
57 percent of its outstanding balance of loans and guarantees, 
including 91 percent of its loans with fixed interest rates, Farmer 
Mac's exposure to interest rate risk stemming from prepayments is 
limited. But, Farmer Mac does hold some loans that are subject to 
interest rate risk caused by prepayments, such as fixed-rate loans with 
less than full yield maintenance acquired through bulk purchase 
transactions, or Part Time Farm loans, which generally allow prepayment 
without penalty. Farmer Mac's prepayment risk model was developed 
internally based on models that predict prepayment behavior for 
residential (housing) mortgage borrowers. But, agricultural real estate 
borrowers may behave differently than residential mortgage borrowers. 
Farmer Mac management said that they followed this approach due to the 
unavailability of external data on agricultural mortgage prepayments. 
They also said that Farmer Mac backtests, that is, compares its 
prepayment model's prediction to the prepayment rates actually observed 
in the recent past, and finds a close correspondence between the 
model's predictions and the experience of its portfolio. A consultant 
to Farmer Mac has indicated that Farmer Mac's current practice of 
incorporating proportional adjustment factors in single family 
prepayment models is consistent with practices at other agricultural 
lenders. Farmer Mac has begun the process of estimating prepayment 
functions based directly on agricultural real estate mortgages. Farmer 
Mac management noted that they are currently working with the 
consultant to develop an agriculture mortgage prepayment model so that 
it can better model prepayment risk. For further information regarding 
prepayment risk, see appendix V.

Farmer Mac Exceeded Statutory and Regulatory Capital Requirements, but 
Could Improve Its Planning for Capital Adequacy:

As of December 31, 2002, Farmer Mac had capital in excess of its 
statutory and regulatory requirements. Its core capital was $184 
million, exceeding its statutory minimum capital requirement of $137.1 
million. Its regulatory capital was $204 million, compared to the 
regulatory risk-based capital requirement of $73.4 million. Although 
Farmer Mac met statutory and regulatory capital requirements, Farmer 
Mac's analysis of capital adequacy could be improved.[Footnote 30]

Pursuant to Farmer Mac's risk-based capital regulation, it is the 
responsibility of the Farmer Mac board to ensure that Farmer Mac 
maintains total capital at a level sufficient for continued financial 
viability and to provide for growth, in addition to ensuring sufficient 
capital to meet statutory and regulatory capital requirements.[Footnote 
31] In projecting Farmer Mac's capital needs in the 2002 Business Plan, 
the Farmer Mac board established a capital goal, based on Farmer Mac's 
current circumstances and needs, at a certain fixed amount above the 
higher of the statutory leverage minimum capital requirement or the 
required risk-based capital level. In doing so, Farmer Mac has not 
performed a test of sufficiency for financial viability and growth 
other than exceeding the statutory and regulatory requirements. Farmer 
Mac officials said that, in their view, FCA's regulatory risk-based 
capital requirement was set at a very conservative level and noted that 
the statutory minimum is higher than the risk-based capital 
requirement. However, regulatory requirements are only minimums and 
financial institutions often find it prudent to keep capital in excess 
of minimum requirements. Moreover, Farmer Mac's minimum statutory 
capital requirement,[Footnote 32] which is not risk-based, is set in 
law and may not be sufficiently responsive to Farmer Mac's emerging 
risks to serve as a proxy for capital sufficiency. In particular, the 
statutory minimum requirement of 0.75 percent capital for off-balance-
sheet obligations applies to Farmer Mac's $2.7 billion of standby 
agreements, a program that did not exist when the statute was enacted. 
Whenever Farmer Mac is obligated under a standby agreement to purchase 
a delinquent loan, it must also increase the capital held against the 
loan from 0.75 to 2.75 percent, nearly a 270 percent increase. As noted 
in our discussion of liquidity risk, Farmer Mac's potential problem is 
that multiple loans would likely be sold to Farmer Mac during times of 
agricultural economic stress or under other adverse conditions. 
Bringing these loans onto Farmer Mac's balance sheet would increase 
Farmer Mac's required capital level, and in the current environment, 
Farmer Mac's current capital is able to absorb this increase. However, 
if standby agreements or off-balance-sheet assets continue to grow, 
Farmer Mac may need to raise capital to withstand such a shock under 
stressful economic conditions. By comparison, for capital requirement 
purposes, bank regulators' risk-based capital standards treat similarly 
structured, off-balance-sheet financial standby arrangements, such as 
guarantees, financial letters of credit, and other direct credit 
substitutes, as if they were on the balance sheet.

Moreover, Farmer Mac's annual filings with SEC illustrate the 
limitations of using the regulatory and/or statutory minimum capital as 
a proxy for having an internal capital adequacy standard. According to 
Farmer Mac's 2002 annual filing with SEC, based on the minimum capital 
requirements, Farmer Mac's current capital surplus of $46.9 million 
could ultimately allow Farmer Mac to carry the risk of an additional 
$15.2 billion of off-balance-sheet guarantees through a combination of 
selling on-balance sheet program assets and adding guarantees.

Disagreements about the Extent of Coverage of Treasury's Line of Credit 
Could Generate Uncertainty:

We identified an issue involving Farmer Mac's $1.5 billion line of 
credit with Treasury that could impact Farmer Mac's long-term financial 
condition. Treasury has expressed serious questions about whether 
Treasury is required to purchase Farmer Mac obligations to meet Farmer 
Mac-guaranteed liabilities on AMBS that Farmer Mac or its affiliates 
hold.[Footnote 33] On the other hand, a legal opinion from Farmer Mac's 
outside counsel states that Treasury would be required to purchase the 
debt obligations whether the obligations are held by a subsidiary of 
Farmer Mac or by an unrelated third party. This disagreement could 
create uncertainty as to whether Treasury would purchase obligations 
held in Farmer Mac's portfolio in times of economic stress. This 
uncertainty also relates to statements made by Farmer Mac to investors 
concerning Treasury's obligation to Farmer Mac, which in turn, could 
affect Farmer Mac's ability to issue debt at favorable rates. 
Ultimately, this uncertainty could impact its long-term financial 
condition.

Farmer Mac's subsidiary, Farmer Mac Mortgage Securities Corporation, 
holds the majority of AMBS that Farmer Mac issued. Farmer Mac's charter 
(the 1987 Act) gives it the authority to issue obligations to the 
Secretary of the Treasury to fulfill its guarantee obligations. 
According to the 1987 Act, the Secretary of the Treasury may purchase 
Farmer Mac's obligations only if Farmer Mac certifies that (1) its 
reserves against losses arising out of its guarantee activities have 
been exhausted and (2) the proceeds of the obligations are needed to 
fulfill Farmer Mac's obligations under any of its guarantees.[Footnote 
34] In addition, Treasury is required to purchase obligations issued by 
Farmer Mac in an amount determined by Farmer Mac to be sufficient to 
meet its guarantee liabilities not later than 10 business days after 
receipt of the certification. However, Treasury has indicated that the 
requirement to purchase Farmer Mac obligations may extend only to 
obligations issued and sold to outside investors.

In a comment letter dated June 13, 1997, and submitted to FCA in 
connection with a proposed regulation on conservatorship and 
receivership for Farmer Mac (1997 Treasury letter),[Footnote 35] 
Treasury stated "…we have 'serious questions' as to whether the 
Treasury would be obligated to make advances to Farmer Mac to allow it 
to perform on its guarantee with respect to securities held in its own 
portfolio---that is, where the Farmer Mac guarantee essentially runs to 
Farmer Mac itself." The 1997 Treasury letter indicated that if the 
purchase of obligations extended to guaranteed securities held by 
Farmer Mac this would belie the fact that the securities are not backed 
by the full faith and credit of the United States, since a loan to 
Farmer Mac to fulfill the guarantee would benefit holders of Farmer 
Mac's general debt obligations. The 1997 Treasury letter stated 
"Treasury's obligation extends to Farmer Mac only in the prescribed 
circumstances, and is not a blanket guarantee protecting Farmer Mac's 
guaranteed securities holders from loss. Nor is the purpose of the 
Treasury's obligation to protect Farmer Mac shareholders or general 
creditors." According to Treasury, the 1997 letter remains its position 
concerning Farmer Mac's line of credit.

Meanwhile, the opinion of Farmer Mac's outside counsel is that the 
guarantee is enforceable whether AMBS are held by a subsidiary of 
Farmer Mac or by an unrelated third party. Farmer Mac's legal opinion 
also states that Treasury could not decline to purchase the debt 
obligations issued by Farmer Mac merely because the proceeds of the 
obligations are to be used to satisfy Farmer Mac's guarantee with 
respect to AMBS held by a subsidiary. According to Farmer Mac, if the 
conditions set forth in the 1987 Act are met--required certification 
and a limitation on the amount of obligations of $1.5 billion--then 
there is no exception in the 1997 Act that authorizes Treasury to 
decline to purchase the obligations. Farmer Mac states that 
discriminating among Farmer Mac guaranteed securities based on the 
identity of the holder in determining whether Farmer Mac could fulfill 
its guarantee obligations would lead to an anomalous situation in the 
marketplace and thereby hinder the achievement of Congress' mandate to 
establish a secondary market for agricultural loans.

Mission-Related Activities Have Increased, but Impact of Activities on 
Agricultural Real Estate Market Is Unclear:

Our analysis of Farmer Mac's impact on the agricultural real estate 
loan market indicated that Farmer Mac has increased its agricultural 
mortgage loan purchase and guarantee activity since our last report in 
1999. At the same time, its enabling legislation contains broad mission 
purpose statements and lacks specific or measurable mission-related 
criteria that would allow for a meaningful assessment of whether Farmer 
Mac had achieved its public policy goals. For example, the statute does 
not contain specific mission criteria for Farmer Mac to make credit 
available for specific clientele such as small, beginning, and 
disadvantaged farmers. In assessing whether Farmer Mac has made 
available long-term credit to farmers and ranchers at stable interest 
rates, we found that its long-term interest rates were similar to the 
rates of agricultural real estate lenders. In addition, Farmer Mac's 
strategy of holding AMBS to lower funding costs and increase 
profitability may have limited the development of a secondary market 
for these securities. Farmer Mac introduced the standby agreement 
program to provide greater lending capacity for agricultural real 
estate lenders, but growth in standby agreements, as with other 
guarantee obligations, could potentially result in reducing the sum of 
capital required to be held by the Farm Credit System and Farmer Mac 
without corresponding mitigating factors such as lender and geographic 
diversification. We found that Farmer Mac's business activities are 
largely concentrated among a small number of business partners and its 
portfolio is concentrated in the West. Finally, the size of Farmer 
Mac's nonmission investment portfolio has decreased as a percentage of 
its total on-and off-balance sheet portfolio. Still the composition and 
criteria for nonmission investment could potentially lead to 
investments that are excessive in relation to Farmer Mac's financial 
operating needs or otherwise be inappropriate to the statutory purpose 
of Farmer Mac.

Farmer Mac Has Continued to Grow, but Mission Criteria Are Lacking:

Farmer Mac's loan and guarantee portfolio has continued to grow since 
1999, but purchase activity notwithstanding, the extent to which Farmer 
Mac has met its public policy mission is difficult to measure. Farmer 
Mac's enabling legislation contains only broad mission related 
guidance; therefore, measurable criteria are not available. The 1987 
Act stated that Farmer Mac was to provide for a secondary marketing 
arrangement for agricultural real estate mortgages in order to (1) 
increase the availability of long-term credit to farmers and ranchers 
at stable interest rates; (2) provide greater liquidity and lending 
capacity in extending credit to farmers and ranchers; and (3) provide 
an arrangement for new lending to facilitate capital market investments 
in providing long-term agricultural funding, including funds at fixed 
rates of interest.

Farmer Mac stated that as a secondary market institution, it faced 
significantly lower economic risks than primary lenders, such as FCS 
institutions and commercial banks, given its ability to attain 
geographic and commodity diversification, access to national and 
international capital markets, and the ability to borrow at lower costs 
due to its agency status. It also noted that the lower capital 
requirements provided to primary lenders through the Farmer Mac I 
program created the potential for increased lending capacity, higher 
profitability, and potentially lower interest rates for farmers and 
ranchers. Notwithstanding these claims, and with respect to the mission 
related guidance, over the past 2 years, the long-term interest rates 
that Farmer Mac offered to agricultural real estate lenders, through 
the Farmer Mac I program have decreased along with the rates of the 
primary agricultural real estate lenders (see fig. 2). We found that 
agricultural mortgage yields have not declined over time relative to 
10-year Treasury securities and that long-term fixed interest rates on 
Farmer Mac I loans were similar to those offered by commercial banks 
and FCS institutions (see fig. 2).

Figure 2: Long-term Interest Rates on Loans Secured by Agricultural 
Real Estate:

[See PDF for image]

[End of figure]

Farmer Mac's Strategy of Retaining AMBS Has Been a Contributing Factor 
in Limiting the Development of a Liquid Secondary Market for AMBS:

Farmer Mac's strategy of holding the loans it purchases and securitizes 
as AMBS has been a contributing factor in limiting the development of a 
liquid secondary market for AMBS. This retained portfolio strategy was 
initially announced in Farmer Mac's 1998 third quarter filing with SEC. 
The explanation given at the time for retaining AMBS was that market 
volatility resulted in lower rates on Treasury securities but wider 
spreads on AMBS. These conditions lowered potential gains on issuance 
of AMBS but facilitated Farmer Mac's retention of AMBS at favorable 
spreads; therefore, Farmer Mac would hold the AMBS until market 
conditions changed.[Footnote 36] According to USDA, holding AMBS has 
typically been more profitable but also more risky than selling AMBS to 
investors. During 2002, Farmer Mac did not make any sales of AMBS to 
unrelated parties.[Footnote 37] Farmer Mac noted that the economics of 
retention have proven superior, and Farmer Mac's growth, profitability, 
and greater capital market presence should facilitate future AMBS 
sales. As of December 31, 2002, Farmer Mac had securitized and sold 7 
percent of its entire Farmer Mac I portfolio (see fig. 3).

Figure 3: Securitization Status of Farmer Mac I Portfolio, as of 
December 31, 2002:

[See PDF for image]

[End of figure]

Farmer Mac's Business Activities Are Largely Concentrated:

With the development of the standby agreement program, Farmer Mac has 
continued to provide products to facilitate capital market investments 
in order to provide long-term agricultural funding, which in turn, 
could result in additional agricultural lending. This is consistent 
with its mission to provide an arrangement for new lending to 
facilitate capital market investments in providing long-term 
agricultural funding, including funds at fixed rates of interest. The 
additional lending would be a result of the lower amount of capital 
that lending institutions would be required to hold, provided their 
products were guaranteed or in a standby agreement with Farmer Mac. The 
risks associated with lower capital requirements would be in part 
mitigated through sufficient diversification relating to participating 
lenders and geography. However, Farmer Mac's activities have been 
largely concentrated in a small number of financial institutions. 
According to Farmer Mac's 2002 annual filing with SEC, Farmer Mac 
purchased eligible loans from 63 financial institutions, and provided 
standby agreements to 16 entities. During 2002, 10 institutions 
generated 90 percent of Farmer Mac's business, and 3 FCS institutions 
represented 47 percent of the outstanding balance of the standby 
agreement program as of December 31, 2002.

Moreover, Farmer Mac's portfolio does not represent the nationwide 
distribution of general farm-related real estate indebtedness across 
commercial banks and FCS institutions. As shown in figure 4, FCS 
institutions were the source for approximately 2 percent of Farmer Mac 
I program loans in 1996, but by December 2002, they accounted for more 
than 55 percent. In contrast, commercial banks participation rate has 
dropped from 80 percent of Farmer Mac I program loans in 1996 to 22 
percent as of December 2002. This compares to FCS institutions holding 
36 percent and commercial banks holding 32 percent of nationwide farm-
related real estate debt, as of 2002. Representatives from USDA and a 
bank association noted that the banking industry strongly supported the 
creation of Farmer Mac in 1987 because they viewed Farmer Mac as a new 
source of competitively priced funding. While commercial banks' 
relative share of Farmer Mac's business has been falling, bank-held 
farm mortgage volume has doubled since Farmer Mac was created. Farmer 
Mac management said that the decline in the commercial banks' 
participation in Farmer Mac's programs was due to the falling interest 
rate environment and a general desire of the commercial banks' to 
retain loans in portfolio. Management anticipated that when interest 
rates begin to rise in the near future, as is forecasted by USDA, 
commercial banks and mortgage brokers will begin to take advantage of 
Farmer Mac's longer-term products.

Figure 4: Farmer Mac Portfolio Exposure by Loan Origination Type:

[See PDF for image]

[End of figure]

By shifting credit risk exposure from FCS institutions to Farmer Mac, 
standby agreements, as with other guarantee obligations, potentially 
lower the overall capital required to be held by FCS (see table 2). 
Whereas the total capital for an unguaranteed loan is $7, the total 
capital for a loan under a standby agreement or swap is only $2.15.

Table 2: Comparison of Total Minimum Capital Levels Per $100 of Loans:

Transaction: FCS loan; FCS institution: $7.00; Farmer Mac: N/A; Total 
capital: $7.00.

Transaction: Standby agreement; FCS institution: $1.40; Farmer Mac: 
$0.75; Total capital: $2.15.

Transaction: Loan sale; FCS institution: $0.00; Farmer Mac: $2.75[A]; 
Total capital: $2.75.

Transaction: Swap for AMBS; FCS institution: $1.40; Farmer Mac: $0.75; 
Total capital: $2.15.

Source: Farm Credit System.

[A] This assumes that Farmer Mac retains the loan or securitizes the 
loan and holds the AMBS on its balance sheet.

[End of table]

Farmer Mac's absorption of FCS institutions' credit risk through the 
standby agreement program might be consistent with a lower capital 
requirement if concentration of credit risk was reduced by geographic 
diversification. However, Farmer Mac's risk exposure is concentrated in 
the western part of the United States. As of December 31, 2002, over 70 
percent of the outstanding balance of Farmer Mac's loan portfolio was 
located in the Southwest and Northwest (see fig. 5). For comparative 
purposes, the corresponding percentage of general farm debt in those 
regions was only 31 percent of nationwide farm debt.[Footnote 38] 
Greater geographic diversification of Farmer Mac loans would lower 
risks of concentration and mitigate risks associated with the lower 
capital requirements.

Figure 5: Farmer Mac I Geographic Concentration of Exposure by Region, 
as of December 31, 2002:

[See PDF for image]

[End of figure]

Proportion of Nonmission Investments Has Declined, but Issues Remain 
about Composition and Potential Growth:

When analyzing Farmer Mac from a mission perspective, an excessively 
large nonmission investment portfolio in relation to Farmer Mac's 
business needs could potentially lead to charges that Farmer Mac is 
misusing its status as a GSE. As of December 31, 2002, the nonmission 
investments that have $830.4 million combined with $723.8 million in 
cash and cash equivalents equaled 37 percent of total balance sheet 
assets at Farmer Mac. This figure is down from 66 percent in 1997, when 
we last reported on GSE nonmission investments, reflecting an increase 
in Farmer Mac's assets resulting from loan purchase and guarantee 
activity since that time.[Footnote 39] Included in Farmer Mac's 
nonmission investments, as reported in SEC filings as of December 31, 
2002, were $93 million in unrated, preferred stock of CoBank, which is 
an FCS institution. This investment is one of Farmer Mac's top five 
holdings of nonmission assets and represents 6 percent of its liquidity 
portfolio. Also in 2002, Farmer Mac's board approved a change to the 
limit of its nonmission investment portfolio. As an alternative to the 
fixed-dollar amount, the Board approved a percentage limit of 30 
percent of the total portfolio, including on-balance sheet assets and 
off-balance sheet commitments. This is the same maximum that FCA allows 
its institutions (other than Farmer Mac). The effect of this change was 
to remove absolute limits on the size of the nonmission investment 
portfolio. FCA officials with whom we spoke said that FCA neither 
endorsed nor objected to the policy change. FCA officials noted that 
they would monitor the portfolio growth to ensure that the potential 
incentive to growing the nonmission investment portfolio was balanced 
with appropriate growth in the loan and guarantee portfolio.

Farmer Mac's Statutory Governance Structure Does Not Reflect Interests 
of All Shareholders and Some Corporate Governance Practices Need to Be 
Updated:

Similar to other publicly traded companies, Farmer Mac is in the 
process of taking actions to ensure that it complies with recent 
legislative and regulatory requirements and proposed changes in NYSE 
listing standards. In accordance with the new requirements, Farmer Mac 
has reaffirmed its audit committee charter and has recently hired 
internal and external auditors who are from different firms. The 
Sarbanes-Oxley Act requires that members of audit committees of listed 
companies be independent and requires that SEC issue and adopt rules 
directing the national securities exchanges to prohibit listing any 
securities of a company that is not in compliance with the audit 
committee requirements. Proposed NYSE listing standards stress the 
oversight role of boards of directors and the independence of the 
directors. Since Farmer Mac's securities are registered with SEC and 
Farmer Mac's Class A and Class C stock are listed on NYSE, Farmer Mac 
is currently subject to the requirements of Sarbanes-Oxley and 
implementing SEC rules, and, absent a waiver, the proposed listing 
standards as they become effective. We noted that Farmer Mac's board 
has taken steps to update its corporate governance practices, but its 
board structure, which is set by law, could make it difficult to comply 
with the board independence requirements proposed in NYSE listing 
standards. Moreover, Farmer Mac's governance structure contains 
elements of a cooperative and elements of an investor-owned, publicly 
traded corporation. Because Farmer Mac shareholders include both 
institutions that utilize its services and public investors, and 
because all members of the board of directors are chosen by the 
cooperative investors or by the President of the United States, the 
board may face difficulties in representing the interests of all 
shareholders. The interests and loyalties of directors of publicly 
traded corporations, including publicly traded GSEs, should be clearly 
focused on serving the interests of all shareholders. However, we found 
that the statutory structure of Farmer Mac's board and the voting 
structure of its common stock hamper Farmer Mac's ability to have such 
a focus. In addition, although discussed to some degree in its proxy 
statement, we found from our discussions with Farmer Mac's 15 board 
members that (1) Farmer Mac's process for identifying and selecting 
board nominees was not transparent to them, (2) training for directors 
was inconsistent, and (3) executive management succession planning was 
not well documented. When assessing Farmer Mac's compensation for its 
executive management, we found that Farmer Mac's total executive 
compensation was within its consultants' recommended parameters; 
however, its vesting program appears more generous than industry 
practices, given Farmer Mac's maturity.

Farmer Mac's Governance Structure Contains Elements of a Cooperative 
and Elements of an Investor-Owned Corporation:

Like other Farm Credit System institutions, Farmer Mac resembles a 
cooperative controlled by institutions that utilize its services. Under 
the 1987 Act, Farmer Mac has three classes of common stock. Class A 
voting common stock is owned by banks, insurance companies, and other 
financial institutions. Class B voting common stock is owned by FCS 
institutions, but ownership of Class C nonvoting common stock is not 
restricted. According to the background of Farmer Mac's charter act, 
Class C nonvoting common stock was created as a means for Farmer Mac to 
raise capital and to preserve equal distribution of voting stock 
between Farm Credit System and non-Farm Credit System 
Institutions.[Footnote 40] However, unlike ownership interests in the 
other FCS institutions, but like the common stock of Fannie Mae and 
Freddie Mac, Farmer Mac's Class A and Class C stock is publicly traded 
on the NYSE. Farmer Mac, through the sale of the stock and the issuance 
of debt securities, depends on the capital markets for funding. Unlike 
the other GSEs, including Fannie Mae and Freddie Mac, Farmer Mac is 
subject to the securities laws, and files disclosure documents with 
respect to its securities issuances. In compliance with the 
requirements of the securities laws, Farmer Mac files quarterly and 
annual reports, proxy statements, and other documents that provide 
information to investors about financial condition and management.

Farmer Mac's board of directors is not elected by all of its 
shareholders. Under the 1987 Act, Farmer Mac's board of directors 
consists of 15 members, 5 of whom are to be elected by holders of the 
Class A voting common stock, 5 are to be elected by holders of Class B 
voting common stock, and 5 are appointed by the President of the United 
States, with the advice and consent of the Senate. The five members 
appointed by the President (1) could not be, or have been officers and 
directors of any financial institutions or entities and (2) were to be 
representatives of the general public--not more than three of whom 
could be members of the same political party and at least two were to 
be experienced in farming or ranching. According to statements made at 
the time of consideration of the 1987 Act, this structure was to 
protect the interests of both the Farm Credit System and commercial 
lenders by providing for equal representation on the board by FCS, 
commercial lenders, and the public sector.[Footnote 41]

Compliance with the disclosure requirements of the 1934 Act provides 
investors with information about Farmer Mac, including information that 
enables investors to compare Farmer Mac with other publicly traded 
companies that participate in the capital markets. However, unlike most 
other publicly traded corporations, Farmer Mac is controlled not by 
investors but by institutions that have a business relationship with 
Farmer Mac. Farmer Mac's board of directors has a fiduciary 
responsibility to act in the best interests of the institution and its 
shareholders; Farmer Mac shareholders included businesses that are 
users of Farmer Mac's financial services and investors in nonvoting 
Class C stock. This structure requires that directors act in the best 
interests of shareholders that may have widely divergent interests. 
Class A and Class B shareholders are concerned with the use of Farmer 
Mac services, while Class C shareholders are generally investors 
concerned with maximizing their profits. Good corporate governance 
requires that the incentives and loyalties of the board of directors of 
publicly traded companies reflect the fact that the directors are to 
serve the interests of all the shareholders. Shareholders of public 
companies can contribute to the governance of corporate conduct with a 
view to enhancing corporate responsibility. Shareholders who exercise 
the power to elect and remove directors can influence corporate policy 
through governance proposals and nominations to the boards of 
directors.

Class C Common Stock Does Not Have Voting Rights:

Farmer Mac's Class C shareholders cannot vote on significant matters 
that generally require shareholders' votes--such as nominating the 
board of directors, executive compensation policies, and the selection 
of the independent auditor. We explained Farmer Mac's nonvoting 
structure to some shareholder advocacy groups, who stated that 
shareholders should be able to vote and voice their opinion on 
governance and management issues. These investor groups advocate "one 
share, one vote." According to Farmer Mac management, the provisions of 
Farmer Mac's charter intended that the agricultural lending industry 
control the board and stockholder voting issues while the company 
developed, which is a process that they believe is still under way. 
Further, they said that holders of Class C common stock acquire the 
stock with that information clearly disclosed to them and implicitly 
accept the representation of their interests by the board and, to a 
large degree, by Class A and Class B stockholders as surrogates 
representing their economic interest, since all classes have the same 
dividend and liquidation rights. However, given Farmer Mac's rapid 
growth and today's corporate governance environment, this nonvoting 
structure may no longer be appropriate.

Eliminating statutory control of the Farmer Mac's board by Class A and 
Class B shareholders and providing an equal voice to Class C 
shareholders, as well as eliminating the statutory requirement that the 
President appoint members of Farmer Mac's board would provide for a 
board elected by all Farmer Mac shareholders. We note however, that 
holders of Class A and Class B stock also hold a significant proportion 
of the Class C shares. According to Farmer Mac's 2003 proxy statement, 
the company's executive officers and directors are the "beneficial 
owners" of 29.8 percent of Farmer Mac's outstanding nonvoting common 
stock, as defined by SEC rules. Almost half this amount is shares owned 
by Zion's Bancorporation, one of whose officers is on Farmer Mac's 
board of directors. SEC's beneficial owners definition includes stock 
options that are exercisable within 60 days; in Farmer Mac's case, 
unexercised options comprise most of the executive officers' and 
directors' beneficially held shares. Consequently, even if Class C 
shareholders were allowed to vote, the Farmer Mac board of directors 
would be elected by many shareholders that currently hold the right to 
vote. In contrast, the executive management and directors of Fannie Mae 
and Freddie Mac have combined beneficial ownership of less than 1 
percent of their respective companies' outstanding common stock.

Farmer Mac Is Subject to NYSE Listing Standards on Corporate 
Governance:

Farmer Mac is subject to NYSE listing requirements, and will be subject 
to proposed listing standards on corporate governance, as well as 
statutory and regulatory requirements. Recent reforms have prompted 
Farmer Mac's board to reassess its oversight role of Farmer Mac and 
take actions to comply with new requirements within the bounds set by 
its statute. Based on our interviews with Farmer Mac's 15 board 
directors, its board committees are taking actions to comply with the 
provisions of the Sarbanes-Oxley Act of 2002 (Sarbanes-Oxley), SEC 
rules, and the proposed NYSE listing standards. For example, Farmer 
Mac's board has revised its audit committee's written charter to 
include the committee's responsibilities, and has recently hired 
internal and external auditors who are not from the same firm. However, 
because Farmer Mac's board structure is established by its charter act, 
it may encounter difficulties in complying with the new standards, 
which require that a majority of the board be independent and that key 
committees (audit, nominating, and compensation) consist entirely of 
independent directors.

In response to recent corporate scandals, corporate governance 
policymakers have focused on the importance of an independent board of 
directors who act in the best interest of the corporation. The 
Sarbanes-Oxley Act contains new requirements concerning the composition 
and duties of the audit committee, including a requirement that all 
audit committee members be independent, which means that the committee 
member cannot accept any consulting, advisory, or other compensatory 
fees from the company (other than compensation for serving as 
director), or be affiliated with the company or any of its 
subsidiaries. Sarbanes-Oxley also requires that SEC adopt rules 
requiring national securities exchanges to prohibit listing any company 
that does not satisfy these requirements.

The NYSE has submitted proposed corporate governance listing standards 
to SEC. To increase the quality of board oversight and lessen the 
potential for conflicts of interest, the proposed listing standards 
require that a majority of the board of directors of listed companies 
be independent. No director qualifies as independent unless the board 
of directors affirmatively determines that the director has "no 
material relationship" with the listed company, either directly or as 
an officer or director of an organization that has a relationship with 
a company. Material relationships can include commercial, banking, 
consulting, legal, and accounting relationships.[Footnote 42] It is not 
clear, however, whether Farmer Mac directors' business relationships 
with Farmer Mac would prevent these individuals from serving as 
independent directors under the NYSE proposed rules. Farmer Mac's 2002 
annual proxy statements indicated that 6 of 15 directors were listed as 
having certain relationships or having conducted related transactions 
with Farmer Mac. In comparison, in their 2002 annual proxy statements, 
Fannie Mae reported that 4 of their 18 directors and Freddie Mac 
reported that 3 of their 18 directors as having business relationships. 
Because the Class A and Class B directors are from institutions that 
have financial relationships of varying degrees with Farmer Mac, they 
may not be independent, thus the statutory structure of Farmer Mac's 
board could make it difficult for Farmer Mac to adopt corporate 
governance practices and policies that may be required or recommended 
by authorities on corporate governance issues. When commenting on our 
report, Farmer Mac officials stated that Farmer Mac was in compliance 
with existing and proposed NYSE standards. Further, they said that 12 
out of the 15 Farmer Mac directors were "independent" in the opinion of 
the board's corporate governance consultant.

Consistency and Transparency of Some Board Processes Could Be Improved:

Regarding board processes, we found that Farmer Mac's board nomination 
process, director training, and management succession planning were not 
as concise, formal, or well documented as best practices would suggest. 
For example, during our interviews with existing directors, we received 
inconsistent responses regarding Farmer Mac's criteria for identifying 
and selecting directors and the process for nominating directors, 
raising concerns about consistency and transparency in the nomination 
process. To further demonstrate the significance of having a 
transparent process for nominating directors, SEC has proposed new 
rules requiring expanded disclosure of companies' nomination process 
and specific disclosure of procedures by which shareholders may 
communicate with directors. The new rules are to enable shareholders to 
evaluate a company's board of directors and nominating committee. The 
proposals include disclosure of the nominating committee's process for 
identifying and considering nominees, including criteria used to screen 
nominees and including the minimum qualifications and standards the 
nominating committee believes company directors should have.

Regarding the training for directors, from our interviews with the 
directors, we found that some directors were provided with in-depth 
training, while others were given a brief orientation to Farmer Mac's 
operations. Finally, at the time of our review, most directors informed 
us that they were uncertain if Farmer Mac had an executive management 
succession plan. Farmer Mac's corporate governance consultant confirmed 
that an executive management succession plan did exist, but had not 
been communicated to the entire board. According to Farmer Mac 
officials, an executive management succession plan was presented and 
approved at the June 2003 annual board meeting.

Farmer Mac's Total Executive Compensation Was Within Consultants' 
Recommended Parameters, but Its Vesting Program Appears Generous:

Farmer Mac's total executive compensation package was within the 
parameters provided by two compensation consultants, although Farmer 
Mac is not readily comparable to private companies or GSEs due to its 
small size, business complexity, and cooperative board structure. 
Farmer Mac has considered itself a start-up company--using 1996 as the 
initial year although it has been in business since 1987--and has 
compared itself to a technology company model because of its daily 
operational risks and demonstrated growth. Generally, start-up 
companies have aggressive compensation packages to attract highly 
qualified employees, paying a higher proportion of compensation in the 
form of equity incentives, such as stock options premised on future 
growth and earnings. Farmer Mac's total compensation has included an 
annual salary, an annual bonus, and stock options, which are included 
in its vesting program.

In 1995, the board retained a compensation consultant to establish a 
compensation package for its staff. Farmer Mac's total executive 
compensation was based on a number of factors--the compensation 
consultant's suggestions, the board's business plan targets, and the 
value of stock options granted. The consultant assesses the 
compensation package annually, and on a multiyear basis, takes into 
account pay levels and rate of increase at Farmer Mac and similar 
private companies and GSEs. In 2002, FCA retained an independent 
compensation consultant to determine if Farmer Mac's total executive 
compensation package was reasonable. We reviewed both Farmer Mac and 
FCA consultant reports. Both consultants provided a range of benchmarks 
to compare Farmer Mac's compensation, but used different assumptions 
that may not be entirely applicable to Farmer Mac. Specifically, we 
question whether the "start-up" assumption--used as an industry 
benchmark by Farmer Mac's consultant to develop its compensation 
package--was still valid, given the maturity of Farmer Mac. Further, 
Farmer Mac's consultant heavily weighted the housing GSEs as comparable 
peer organizations to ensure that Farmer Mac's compensation structure 
was competitive enough to attract and retain qualified executives. 
FCA's consultant also used the housing GSEs as benchmarks, in addition 
to mortgage banking organizations and financial service organizations 
because the various organizations more closely represented the 
positions from which executive management would be recruited. We 
question whether putting such heavy emphasis on housing GSEs as a 
benchmark is appropriate because they are so much larger and more 
complex than Farmer Mac, in terms of size and structure, earnings, 
portfolio, and operations. For example, Farmer Mac has 33 employees 
compared to Fannie Mae and Freddie Mac's 4,700 and 3,900 employees, 
respectively. As shown in table 3, Farmer Mac's compensation and 
options granted fell below the much larger housing GSEs.

Table 3: Annual Compensation and Options Granted for CEO's of Farmer 
Mac and Housing GSEs, 2002:

(Dollars in millions).

President & CEO, Farmer Mac; Loan & Guarantee Portfolio: $5.5 billion; 
Employees: 33; Annual salary: $447,480; Annual compensation bonus: 
$344,195; Options grant date present value: $1,150,783; Value of 
unexercised in-the-money options at year-end exercisable: $9,511,068.

CEO and Chairman, Fannie Mae; Loan & Guarantee Portfolio: $1.8 
trillion; Employees: 4,700; Annual salary: 992,250; Annual 
compensation bonus: 3,300,000; Options grant date present value: 
6,680,395; Value of unexercised in-the-money options at year-end 
exercisable: 1,441,600.

CEO and Chairman; Freddie Mac (A) Loan & Guarantee; Portfolio: $1.1 
trillion Employees: 3,900; Annual salary: 1,132,500; Annual 
compensation bonus: 2,123,438; Options grant date present value: 
3,899,741; Value of unexercised in-the-money options at year-end 
exercisable: 17,227,372.

Source: Compensation information from Farmer Mac and Fannie Mae proxy 
statements as of April 2003, and Freddie Mac's proxy statement as of 
April 2002. Portfolio size and employee data from Farmer Mac 2002 
Annual Report and Fannie Mae 2002 Annual Report, and Freddie Mac 2001 
Annual Report. Freddie Mac's 2002 and 2001 financial results subject to 
restatement.

Notes: Freddie Mac's loan and guarantee portfolio size is an estimate 
from its 2001 financial statements because Freddie Mac was in the 
process of restating its 2000 to 2002 financial statements. Freddie Mac 
was not expected to complete the restatement until November 30, 2003.

This table excludes long-term compensation, restricted stock awards, 
other annual compensation, securities underlying options, LTIP payouts, 
and all other compensation, which includes life insurance premiums, 
defined contribution pension plans, and Retirement Savings Plans for 
Employees.

[End of table]

However, when benchmark issues are set aside, and Farmer Mac is 
compared to public companies or GSEs, its total executive compensation 
was within the consultants' recommended parameters but its stock option 
vesting program appears generous compared to general industry 
practices. Under Farmer Mac's 1997 Stock Option Plan, Farmer Mac 
employees and directors have been granted options in stages, with one-
third of the options vested immediately on the date being granted, one-
third vested at the end of the following year, and the remainder vested 
in the second year. According to current practices of public and 
private companies with a public mission, these companies have average 
vesting periods of 4 to 5 years, with employees vesting 25 percent 
annually for 4 years after 1 year of employment. Generally, companies 
structure their vesting schedules to attract and to retain employees. 
Additionally, according to researchers, start-up companies use vesting 
programs to attract an important group of intellectual capital 
employees, and vest sooner to bolster income levels so that the 
employees can be compensated for their contributions. In these cases, 
more generous vesting programs serve the need to quickly develop the 
company to profitability, which may no longer be suitable for Farmer 
Mac's needs.

FCA Has Taken Steps to Enhance Oversight of Farmer Mac, but Faces 
Challenges That Could Limit the Effectiveness of Its Oversight:

FCA has recently taken several steps to strengthen its oversight of 
Farmer Mac, including instituting a more comprehensive safety and 
soundness examination and undertaking initiatives to expand its 
regulatory framework. However, as it continues to improve its oversight 
of Farmer Mac, FCA faces five major challenges. First, limitations 
exist in the model used to estimate Farmer Mac's credit risk. Second, 
FCA's regulation does not include a component to measure credit risk on 
liquidity investments held by Farmer Mac. Third, FCA's market risk and 
income models may understate estimated levels of required risk-based 
capital. Fourth, lack of criteria defining Farmer Mac's mission limits 
FCA's ability to effectively oversee Farmer Mac's mission achievement. 
Finally, FCA is challenged by regulating both a primary and secondary 
market.

FCA Has Expanded Its Farmer Mac Examination and Is Taking Actions to 
Improve Oversight:

FCA's June 2002 annual safety and soundness examination had a larger 
scope and employed more resources than its past examinations. The 
examination included a comprehensive review of Farmer Mac's financial 
condition, portfolio activity, risk management, and a review of board 
governance and executive compensation. FCA officials said that this 
CAMELS-based examination would serve as a guide for future Farmer Mac 
examinations.[Footnote 43] In addition, FCA was closely monitoring 
Farmer Mac's corrective actions to address identified weaknesses.

As part of its increased focus on Farmer Mac oversight, FCA formed a 
working group in 2002 that prepared a white paper on Farmer Mac's 
nonmission investments and liquidity requirements. FCA was developing 
regulations in this area to ensure that Farmer Mac's nonmission 
investments would be appropriate in both quality and quantity and that 
Farmer Mac's use of its GSE status to issue debt would be appropriate. 
According to FCA representatives, to date, FCA has not placed a 
regulatory limit on the level or quality of Farmer Mac's nonmission 
investments, nor has it regulated specific liquidity standards.

FCA also formed another working group in 2002 to study the implications 
of regulatory capital arbitrage between FCS institutions and Farmer 
Mac. The regulatory capital arbitrage working group provided a white 
paper to the FCA board that contained an in-depth analysis of the 
causes and sources of capital arbitrage. The white paper presented 
several options for how FCA could reduce potential safety and soundness 
issues that might arise when FCS institutions and Farmer Mac engaged in 
capital arbitrage activities to reduce capital required to be held. FCA 
said that the agency is still studying the issue and has made no 
decisions on any specific actions.

Notwithstanding these positive developments, FCA has not been updating 
and reformatting Farmer Mac's call reports, a tool used for off-site 
monitoring. Our review of yearend 2001 and 2002 call report schedules 
and corresponding instructions indicated that in some cases, they do 
not fully conform to FCA regulations nor have these documents been 
updated to reflect recent accounting changes. One of the discrepancies 
we noted was FCA's acceptance of Farmer Mac's inaccurate reporting of 
the amount of one of three categories in which Farmer Mac was required 
to maintain its minimum core capital. Although to date the amount of 
capital affected was very small, this discrepancy raises questions on 
FCA's oversight of this part of Farmer Mac's capital requirement. FCA 
officials responded that they do not believe this discrepancy weakens 
FCA's oversight of Farmer Mac's capital requirement. Further, FCA 
officials recognized that the call report instructions need to be 
revised and said that they have plans to update them but that resources 
were currently not available due to other priorities associated with 
their oversight of Farmer Mac. FCA officials said that outdated call 
reports were not a primary concern because they augment the call report 
information with various other sources, including SEC filings and risk-
based capital supporting data obtained from Farmer Mac.

FCA Faces Challenges as It Enhances Farmer Mac Oversight:

FCA has begun to strengthen its oversight of Farmer Mac, but the agency 
still faces a number of technical and supervisory challenges. These 
include deficiencies in the estimation and measurement of risk and 
regulatory management issues.

Limitations Exist in the Model Used to Estimate Farmer Mac's Credit 
Risk:

The model FCA used to estimate the amount of risk-based capital that is 
required to cover Farmer Mac's credit risk, utilizes the same data that 
are used in Farmer Mac's loan loss estimation model. As we discussed 
earlier, this model is limited by the poor data quality. We identified 
limitations related to using FCBT data and issues such as not modeling 
changes in interest rates, loan terms, or property values. For example, 
the model uses FCBT data to estimate loan losses even though Texas did 
not have the highest rates of default and severity of agricultural 
mortgage losses as required under Farmer Mac's statutory risk-based 
levels.[Footnote 44] Legislation on Farmer Mac's risk-based capital 
requirements requires FCA to stress test the model, based upon the 
worst experience for defaults and loss severities for a period of not 
less than 2 years for agricultural real estate loans in contiguous 
areas comprising at least 5 percent of the U.S. population.[Footnote 
45] Analysis by FCA's consultants indicates that Minnesota, Iowa, and 
Illinois experienced the greatest decrease in farmland prices in 1983 
and 1984. However, the loans in FCA's database are limited to Texas, 
which experienced the fourth greatest decrease in farmland prices. 
FCA's consultants found that FCBT had the only usable loan database for 
the purpose of building the credit risk model to estimate Farmer Mac's 
credit risk. Since the loan data were limited, it may not provide all 
data elements that would be desirable in a stress test. For example, 
the sample was small and it did not fully reveal the extent of 
restructuring of loans that could affect default estimates and losses. 
Additionally, the FCBT loan files did not show the extent to which loan 
terms had been changed to forestall foreclosures. Consequently, if some 
of these loans did have losses, which were not recorded in the 
database, the frequency of credit losses may be understated in the 
credit risk analysis. As we explained earlier, the loans in Farmer 
Mac's current portfolio tend to adjust for changes in interest rates 
more quickly than the loans issued by FCBT in the 1970s and 1980s. As 
such, loans in the current portfolio may be exposed to credit risk 
longer than were the FCBT loans used in estimating the credit risk 
model and therefore, the FCBT loans would not be representative of 
Farmer Mac's current risks.

We also identified limitations in the structure of FCA's model. One 
limitation is that FCA's credit risk model was constructed so that the 
expected losses in a stressed environment are the same no matter what 
appreciation or depreciation in farmland prices occurred over the life 
of the loan in any period other than the period of maximum stress. The 
model also does not consider the effect that interest rate changes may 
have on the probability of default, such as the increased default risk 
of fixed-rate loans with yield maintenance in times of falling interest 
rates, or the increased risk of adjustable rate loans at times of 
rising interest rates. Another limitation of this model is that it does 
not differentiate between loans with short-and long-amortization 
periods, although loans with shorter amortization periods are likely to 
have lower credit risk, holding other loan underwriting terms constant. 
Because these variables are not included in the credit risk model, by 
varying its mix of fixed-rate and adjustable-rate loans, or short-
versus long-amortization loans, Farmer Mac could change its credit risk 
profile with no resultant change in the regulatory capital for credit 
risk as measured by the FCA model. A more detailed discussion of the 
limitations of FCA's credit risk model is presented in appendix VI.

FCA's Regulation Does Not Capture Credit Risk on Farmer Mac's Liquidity 
Investments and AgVantage Bonds:

FCA's regulation for calculating Farmer Mac's risk-based capital does 
not assess the amount of capital that must be held against credit risk 
associated with assets in Farmer Mac's liquidity investment portfolio. 
As such, there is no credit risk capital charge against approximately 
37 percent of Farmer Mac's total balance sheet assets, which consist of 
liquidity investments such as commercial paper or corporate bonds. 
Although corporations with investment-grade ratings have relatively 
high credit quality, there is a possibility that they will default and 
fail to make all interest and principal payments in full and on 
schedule. In contrast, other financial regulators, including the 
Federal Housing Finance Board (FHFB), Federal Reserve Board, Office of 
the Comptroller of Currency, Office of Thrift Supervision, and Office 
of Federal Housing Enterprises Oversight (OFHEO), calculate the capital 
that must be held for the credit risk on investment securities, loans, 
and other assets, and also capital for the risk that a counterparty in 
a derivative transaction would fail to perform.

In addition, FHFB calculates required risk-based capital for advances 
that Federal Home Loan Banks make to their members. Farmer Mac's 
Agvantage bond program is structured similarly to the FHLB advances, in 
that both require overcollateralization using borrower mortgage assets 
as collateral.[Footnote 46] However, unlike FHFB, FCA does not include 
AgVantage bonds in its risk-based capital calculation.

Market Risk and Income Models May Understate Estimated Levels of 
Required Risk-based Capital:

FCA uses results from Farmer Mac's interest rate risk model to measure 
the level of market risk to which Farmer Mac is exposed and determine 
corresponding levels of risk-based capital. As discussed previously, 
the Farmer Mac interest rate risk model has limitations with regard to 
prepayment modeling and the effect of prepayment penalties. These 
limitations could lead to errors in measuring the prepayment risk to 
which Farmer Mac is exposed and weaken FCA's oversight of risk-based 
capital, in addition to affecting Farmer Mac's risk management.

Farmer Mac uses the estimated behavior of single-family residential 
mortgage benchmarks to estimate the prepayment risk of commercial 
agricultural mortgages. Using one type of mortgage as a benchmark for 
another may lead to an underestimate of the extension risk in Farmer 
Mac's commercial agricultural mortgage holdings. Extension risk is the 
tendency for expected lifetimes of a mortgage to lengthen when interest 
rates rise. Most single-family residential mortgages have due-on-sale 
clauses, which compel borrowers to pay off their loan balances when 
selling their property. However, commercial agricultural mortgages are 
more easily assumed when it is advantageous to do so, often in the form 
of a "wrap," in which the property is sold as part of a long-term 
contract, so that the title to the property does not formally change 
hands for several years. The result is, at times of rising interest 
rates, the average life of commercial agricultural mortgages will 
increase more than will the average life of residential mortgages.

Additionally, FCA has chosen to incorporate an estimate of Farmer Mac's 
earnings into its income model, that assumes the level of new business 
activity and profitability for Farmer Mac per year will be unchanged in 
a 10-year period (steady state approach). In effect, even in the stress 
test scenario, by holding new business activity level constant, losses 
can be compensated for with profits from new business. By not including 
specific instructions on this issue, the 1991 amendments to the 1971 
Act establishing Farmer Mac's risk-based capital standards gave FCA the 
choice of including or excluding an estimate of Farmer Mac's earnings 
over a 10-year stress period when calculating Farmer Mac's risk-based 
capital requirements.[Footnote 47] FCA officials said that they made a 
judgment to use the steady state approach because it allowed them to 
treat Farmer Mac as a going concern business, which they interpreted to 
be the intent of the statute. Further, FCA officials said that in 
developing the model, they found that using a steady state approach 
resulted in their use of fewer assumptions than would have been 
required by other approaches. In contrast, we have previously reported 
on the serious problems involved in estimating future income for GSEs 
since it is hard to determine what a reasonable level of activity, 
profits, or losses would be during a stressful period.[Footnote 48]

Consistent with our concerns, other regulators such as OFHEO and FHFB 
do not use an estimate of earnings on new business when calculating 
their regulatory capital requirements.[Footnote 49] In addition, OFHEO 
assumes that as Fannie Mae and Freddie Mac refinanced their short-term 
debt to support outstanding business, they could face higher interest 
rates caused by increasing risks of borrowing during a stressful 
period. Recently, OFHEO modified its stress test by increasing the 
short-term rates, which the model assumes will be paid by the housing 
enterprises by 10 basis points.[Footnote 50] If FCA were to make a 
similar adjustment to future borrowing costs for Farmer Mac in a stress 
environment, the effect would be to reduce the estimated amount of 
future income earned by Farmer Mac, hence increasing the level of 
capital required to be held.

Lack of Criteria and Procedures Limit FCA's Ability to Effectively 
Oversee Farmer Mac's Mission Achievement:

Although FCA has general regulatory authority over Farmer Mac for both 
safety and soundness oversight and mission regulation, FCA has focused 
primarily on safety and soundness.[Footnote 51] We recognize that 
balancing these two goals--safety and soundness oversight and mission 
regulation--is difficult and could create tensions. However, if FCA is 
to oversee Farmer Mac's mission achievement, a lack of criteria and 
processes to measure how Farmer Mac's activities and products have 
contributed to mission achievement will limit its effectiveness. As 
discussed earlier, Farmer Mac's enabling legislation does not establish 
specific mission obligations that include specific or measurable goals; 
rather, Farmer Mac's mission is broadly stated. FCA officials said that 
FCA's authority to establish specific and measurable goals is fact 
specific and would depend on the particular nature of the proposal. 
Further, unlike the Department of Housing and Urban Development (HUD), 
FCA has received no congressional direction to undertake an analysis to 
determine the net public policy benefit of Farmer Mac's actions.

FCA officials said that the continued combined effect of FCA's 
supervisory efforts and regulatory development plans would bring 
greater focus on Farmer Mac's accomplishment of its public policy 
purpose. The officials also said that FCA has taken various steps to 
indirectly monitor Farmer Mac's mission achievement, including looking 
at Farmer Mac's book of business to see how it has grown over time and 
to identify inappropriate activities and products.

Overseeing Both FCS Banks and Farmer Mac Is a Regulatory Challenge:

FCA's role as regulator of Farmer Mac and the FCS institutions raises a 
concern about regulatory conflict of interest. FCS is a primary market 
for agricultural real estate loans, while Farmer Mac is the secondary 
market for these loans. We have previously reported that to carry out 
oversight responsibilities effectively, a GSE regulatory structure must 
separate regulation of primary and secondary market 
participants.[Footnote 52] This criterion posits that a regulator 
overseeing both a GSE and its primary business partners could be 
subject to conflicts of interest. For example, if an FCS institution 
was in danger of failing, the regulator might be tempted to pressure a 
healthy GSE into increasing the price it pays the bank for loans. Or, 
if a GSE was in poor financial health, the regulator might be tempted 
to encourage the GSE counterparties to discontinue their relationships. 
On the other hand, we recognize that a single regulator could offer 
some benefits such as knowledge of the market and its participants, and 
the opportunity to observe the transactions and trends between the 
primary and secondary markets.

Congress recognized this potential regulation problem and it attempted 
to mitigate this by creating OSMO, a separate office within FCA to 
regulate Farmer Mac. As required by the 1987 Act, the director of OSMO 
is selected by and reports to the FCA Board.[Footnote 53] Yet, the 1987 
Act directs FCA examiners, who also examine FCS institutions, to 
examine Farmer Mac's financial transactions. The 1987 Act also charges 
FCA with ensuring that OSMO is adequately staffed to supervise Farmer 
Mac's secondary market activities; although, to the extent practicable, 
the personnel responsible for supervising the powers, functions, and 
duties of the corporation should not also be responsible for 
supervising the banks and associations of the Farm Credit System. While 
this regulatory structure provides for a degree of separation between 
FCA's responsibilities for FCS institutions and its responsibilities 
with respect to Farmer Mac, in practice, the FCS institutions and 
Farmer Mac are still subject to oversight by the same FCA board and 
reviewed by some of the same FCA examiners and analysts. Consequently, 
FCA could be subject to potential conflicts of interest. In our 
discussions with FCA officials, they said that they were aware of the 
need to maintain the proper balance in their oversight roles to avoid 
such potential conflicts.

Conclusions:

Government sponsorship of a financial institution, such as Farmer Mac, 
can generate a number of public benefits and costs, which are difficult 
to quantify. To the degree that lower funding costs and other benefits 
are passed on to borrowers in the affected financial sector, public 
benefits are generated. However, government sponsorship also generates 
potential public costs. One potential cost is the risk that taxpayers 
will be called upon if a GSE is unable to meet its financial 
obligations. In Farmer Mac's case, it would be the need to draw on its 
$1.5 billion line of credit with Treasury and the possibility that the 
federal government might appropriate further funds in the event that 
Farmer Mac faces financial difficulties. GSE status inherently weakens 
market discipline, which heightens the importance of internal and 
external oversight by the GSE's board of directors, auditors, and 
regulators, as well as transparency through financial reporting and 
credit ratings to the creditors and investors.

Farmer Mac's financial condition has improved since we last reported in 
1999; specifically, its income has increased and its capital continues 
to exceed required levels. Farmer Mac's risk profile has become more 
complex as a result of the growth in size and complexity of its loan 
and guarantee portfolio. Although the company has made progress over 
the past few years to enhance its credit controls, asset management, 
and reduction of asset liability mismatch, its efforts to measure and 
monitor its risks have not kept pace and could be improved. As its loan 
and guarantee portfolio ages and delinquencies increase, it is key for 
Farmer Mac to continue to manage its credit risk by improving its loan 
loss estimation model and documentation of policies, procedures, and 
management judgments related to loan purchases and guarantees. More 
importantly, the rapid growth of standby agreements has generated a 
need for Farmer Mac to consider a funding strategy that would allow it 
to meet unexpected demands to fund purchases of underlying impaired or 
defaulted loans, in the event of stressful economic conditions. A 
funding strategy would entail a comprehensive contingency funding 
liquidity plan and a detailed analysis of capital adequacy. As noted, a 
strategy that consists of selling AMBS to obtain funding would 
potentially be limited by the lack of knowledge of the depth and 
liquidity of the secondary market for AMBS.

Farmer Mac has increased its mission-related activities since we last 
reported on this in 1999, but it is still not apparent if sufficient 
public benefits are derived from these activities. The lack of specific 
or measurable mission goals in its statute beyond providing a secondary 
market and stable long-term financing does not allow for a meaningful 
assessment of whether Farmer Mac's activities are having the desired 
impact on the agricultural real estate market. Further, because Farmer 
Mac has elected to retain nearly all its AMBS in portfolio for 
profitability reasons, the depth and liquidity of the secondary market 
for AMBS is unknown.

Similar to other publicly traded companies, Farmer Mac is faced with 
the challenges of updating its corporate governance practices to comply 
with Sarbanes-Oxley, SEC rules, and proposed NYSE listing standards as 
they become effective. As a GSE, Farmer Mac, however, has a board 
structure set in statute, which hampers its efforts to comply with the 
stricter independence requirements in proposed NYSE listing standards, 
specifically those requirements calling for a fully independent and 
competent audit committee. Moreover, Farmer Mac's statutory governance 
structure has elements of a cooperative and investor-owned publicly 
traded company, which does not reflect the interest of all 
shareholders. While we did not draw conclusions on Farmer Mac's overall 
executive compensation, we would no longer consider Farmer Mac a start-
up company and the assumptions used to set its executive compensation 
may no longer be valid. Changes are needed to Farmer Mac's vesting 
program for stock options to bring them more in line with general 
industry practices and other GSEs. Farmer Mac could improve its 
training for directors, provide more transparency to its directors on 
the nomination process, and better document its succession plan for its 
executive management. These actions, along with obtaining a credit 
rating to provide transparency to the market, could also help Farmer 
Mac respond to criticisms and increased expectations in today's market 
environment.

In addition to Farmer Mac's internal management of risks, as a GSE, it 
is required to have regulatory oversight to ensure that it operates in 
a safe and sound manner. Beginning in 2002, FCA had improved its 
oversight of Farmer Mac, but continues to face significant challenges 
in sustaining and further enhancing its oversight. While FCA has 
improved its examination approach, more remains to be done to improve 
its assessment of risk-based capital and mission oversight. We 
discussed a number of issues related to the data and structure of FCA's 
risk-based capital model, but the overall impact these issues have on 
the estimate of risk-based capital for Farmer Mac's credit risk is 
uncertain. Some concerns, such as the potential undercounting of loans 
that experienced credit losses, or greater prepayment of FCBT loans 
relative to Farmer Mac loans, may result in the FCA credit risk model 
underestimating the credit risk capital requirement. Other issues, such 
as the lack of a variable to track land price changes for any but the 
most stressed year, may cause the model to overestimate the credit risk 
capital requirement. Augmented data and more analysis could better 
determine the relative magnitudes of these effects. While FCA's 
oversight of Farmer Mac typically has focused on safety and soundness, 
it lacks criteria and procedures to effectively oversee how well Farmer 
Mac achieves its mission. At the same time, Farmer Mac's enabling 
legislation is broadly stated and does not include any measurable goals 
or requirements to assess progress toward meeting its mission. More 
explicit mission goals or requirements would help FCA in improving its 
oversight of Farmer Mac.

Recommendations:

To help ensure that Farmer Mac's management can properly identify, 
manage, and control risks, we recommend that Farmer Mac management 
ensure that it has adequate staff resources and technical skills to 
oversee the following actions:

* Address weaknesses in its loan loss estimation model, which could 
affect the reasonableness and adequacy of the loan loss allowance, 
through the following actions:

* Include current data on farm loan payment, delinquency, and valuation 
for the loans included in the estimation model so that the estimation 
process reflects current loan and economic conditions;

* Explore other data sources that are relevant to Farmer Mac's current 
portfolio for estimating probability, amounts, and distribution of 
credit losses in its estimation model; and:

* Improve documentation of the results of the model compared to actual 
portfolio and economic conditions, and of the reconciliation to the 
amounts recorded in the financial statements.

* Continue to reduce its credit risk by improving its documentation of 
policies and procedures, and management's actions and judgments through 
the following actions:

* Continue to gather documentation supporting management's assessment 
of loans approved using underwriting standard 9, including 
quantification and evaluation of compensating risk factors, and develop 
a process for utilizing such information in the management decision 
process for future exceptions and for estimating credit losses, and:

* Improve documentation supporting and quantifying the effect of 
extracting specific loan loss estimates from the overall loss estimate 
to determine whether this approach differs materially from estimating 
specific loan losses separately.

* Reevaluate its current strategy of holding agricultural mortgage-
backed securities in portfolio and issuing debt to obtain funding.

* Develop a contingency funding liquidity plan to address potential 
vulnerabilities in less favorable capital markets conditions and 
liquidity needs arising from the rapid growth of standby agreements.

* Improve the quality of its prepayment model to ensure accurate 
interest rate risk measurements.

* Improve its analysis of capital adequacy to help ensure that capital 
would meet the needs of increasing and potential credit risks and 
growth.

Although the Farmer Mac board has taken steps to strengthen its 
corporate governance practices, we recommend that the Chairman, Farmer 
Mac, further enhance those practices by:

* reevaluating stock option levels and vesting period to ensure that 
they are not excessive in relation to comparable industry standards for 
vesting and waiting period for stock options;

* better communicating the criteria for identifying and selecting 
director nominees and the process to nominate directors among the 
directors;

* formalizing executive management succession plan and communicate plan 
with all board members to provide transparency; and:

* providing consistent training on governance and Farmer Mac related 
topics to all board members to increase directors' understanding of 
risks facing the corporation.

Finally, to improve the quality and effectiveness of FCA's oversight of 
Farmer Mac, we recommend that FCA implement the following steps:

* Continue to obtain more relevant and current data on farm loan 
behavior used in the risk-based capital model and consider more 
flexible modeling approaches to credit risk, such as those used by 
OFHEO for regulatory purposes or the Federal Housing Administration 
(FHA) for evaluating actuarial soundness;

* Continue to improve and formalize off-site monitoring of Farmer Mac, 
including reviews of Farmer Mac's regulatory reporting;

* Create a plan to implement actions currently under consideration to 
reduce potential safety and soundness issues that may arise from 
capital arbitrage activities of Farmer Mac and FCS institutions;

* Examine how other secondary market regulators developed regulations 
to require the GSEs to obtain a government risk credit ratings from 
nationally recognized statistical rating agencies; and:

* Assess and report on the impact Farmer Mac's activities has on the 
agricultural real estate lending market.

Matters for Congressional Consideration:

Congress may wish to consider the following legislative change:

* Establish clearer mission goals for Farmer Mac with respect to the 
agricultural real estate market to allow for a meaningful assessment of 
whether Farmer Mac had achieved its public policy goals;

* Allow FCA more flexibility in establishing capital standards that are 
commensurate with Farmer Mac's changing risk profile and in setting 
minimum capital requirements;

* If Congress intends for Farmer Mac to operate in a cooperative manner 
and maintain its current board structure of Class A and Class B stock, 
it may wish to consider making Farmer Mac a true cooperative entity 
like the Federal Home Loan Bank System, and rescind Farmer Mac's 
authority to issue Class C stock. However, if Congress intends for 
Farmer Mac to operate as a publicly traded company, it should consider 
amending (1) Farmer Mac's board structure to ensure an independent 
board and independent and competent audit committee and (2) the 
structure of Farmer Mac's Class C common stock to include a one share, 
one vote principle to provide the opportunity to better reflect all 
shareholder interests.

Agency Comments and Our Evaluation:

We requested comments on a draft of this report from the heads or their 
designees of the FCA, Farmer Mac, SEC, and Treasury. We received 
written comments from Farmer Mac and FCA that are summarized below and 
reprinted in appendixes VII and VIII, respectively. SEC did not provide 
comments. FCA, Farmer Mac, and Treasury also provided technical 
comments that we have incorporated as appropriate.

In commenting on this report, Farmer Mac stated that it agreed with the 
report's findings and conclusions on Farmer Mac's risk management 
practices and has taken a number of steps toward implementing the 
majority of the recommendations. While Farmer Mac seemed to agree with 
the report's recommendations to improve its analysis of capital 
adequacy, develop a contingency funding plan, and improve documentation 
of management exceptions to its eight major underwriting standards, 
Farmer Mac's comments did not address our recommendations to reevaluate 
its strategy of holding AMBS in its portfolio, improve the quality of 
its prepayment model, better communicate the criteria for selecting 
director nominees and provide consistent training to the board of 
directors.

Farmer Mac commented that in discussing the availability of the 
Treasury line of credit relative to AMBS that Farmer Mac or its 
affiliates hold, the report acknowledged that Farmer Mac has a legal 
opinion by its outside counsel stating that the Treasury line of credit 
would be available in those circumstances; therefore, the question is 
moot. In fact, the report discussed the line of credit because Treasury 
has expressed serious questions about whether Treasury is required to 
purchase Farmer Mac obligations to meet Farmer Mac-guaranteed 
liabilities on AMBS that are held by Farmer Mac or its affiliates, and 
therefore, this issue remains unresolved until that time when Farmer 
Mac approaches Treasury for assistance. Farmer Mac commented that if it 
were coming under pressure to fund its guarantee obligations, it could 
sell AMBS it held to third parties long before it needed to use the 
line of credit. As we stated in the report, however, the depth and 
liquidity of the demand for these securities in the current market is 
unknown. Therefore, Farmer Mac would be selling AMBS at the same time 
that it was coming under pressure to fund its guarantee obligations, 
which would most likely affect Farmer Mac's ability to sell these 
securities and the price at which it could sell them.

Farmer Mac seems to disagree with our concern on funding liquidity risk 
that might arise from standby agreements. Farmer Mac commented that the 
report posits a situation in which loan defaults go far beyond the 
default rate peak for agricultural loans within the Farm Credit System 
in 1986. We do not provide an estimate of the level of default rate at 
which Farmer Mac would need additional funding. The report stated that 
if rapid growth continues, standby agreements could generate 
substantial funding liquidity risk under stressful economic conditions. 
By using the default rate peak, Farmer Mac is alluding to the stressful 
conditions incorporated in the risk-based capital model. However, this 
model addresses credit risk, not liquidity risk. Under standby 
agreements, Farmer Mac would need to fund not only the net losses from 
foreclosures that were used in estimating the risk-based capital 
requirement but must fund the gross amount of loans that enter 
foreclosure and seriously delinquent loans presented for purchase to 
Farmer Mac. Other more technical comments provided by Farmer Mac and 
our detailed response is discussed in appendix VII.

Finally, FCA overall concurred with our report's findings and 
conclusions that are focused on FCA's work to oversee the safety and 
soundness of Farmer Mac. FCA also agreed to implement the 
recommendations for improving FCA's oversight of Farmer Mac contained 
in this report through current regulatory and examination work that is 
in process, and as necessary, new initiatives. In response to our 
recommendation regarding the risk-based capital model, FCA does not 
agree that additional data and modeling would add value, although FCA 
is studying the possibility of updating the data used in its model. As 
we stated in the report, the data used by FCA do not include all the 
components of credit losses, may not capture all the loans that 
experienced losses, and the loans used in the model have different 
interest rate characteristics than those currently purchased by Farmer 
Mac. Also as stated in the report, the key independent variable used in 
FCA's model--land price decline--is defined in such a way that the 
model will produce a biased estimate of the impact of land price 
declines on credit losses. FCA's technical comments and our detailed 
response are discussed in appendix VIII.

:

We are sending copies of this report to the Chairmen and Ranking 
Minority Members of the Senate Committee on Banking, Housing, and Urban 
Affairs; the House Committee on Financial Services; and the House 
Committee on Agriculture. We are also sending copies of this report to 
the President and Chief Executive Officer of Farmer Mac; the Chairman 
and Chief Executive Officer of the Farm Credit Administration, the 
Chairman of SEC, the Secretary of Treasury, and other interested 
parties. This report will also be available at no charge on GAO's 
Internet homepage at [Hyperlink, http://www.gao.gov] http://
www.gao.gov.

Please contact us at (202) 512-8678 if you or your staff have any 
questions concerning this work. Key contributors are ackknowledged in 
appendix IX.

[See PDF for image]

[End of figure]

Davi M. D'Agostino 

Director, Financial Markets and Community Investment:

Signed by Davi M. D'Agostino: 

Jeannette M. Franzel 

Director, Financial Management and Assurance:

Signed by Jeannette M. Franzel: 

[End of section]

Appendixes: 

Appendix I: Objectives, Scope, and Methodology:

As requested by the Senate Committee on Agriculture, Nutrition, and 
Forestry, we conducted a review of the Federal Agricultural Mortgage 
Corporation (Farmer Mac). Our objectives were to (1) assess Farmer 
Mac's current financial condition and risk management practices, (2) 
determine the extent to which Farmer Mac has achieved its statutory 
mission, (3) evaluate Farmer Mac's corporate governance as it pertains 
to board structure and oversight and executive compensation, and (4) 
evaluate the Farm Credit Administration's (FCA) oversight of Farmer 
Mac.

The focus of our review on Farmer Mac's secondary market activity in 
agricultural mortgages was on the Farmer Mac I Program because it is 
the primary program through which Farmer Mac conducts its secondary 
market activity. However, we included Farmer Mac II Program activity in 
our overall analysis of Farmer Mac's financial condition. To address 
our objectives overall, we reviewed the legislative history and 
statutory authorities governing Farmer Mac. We also reviewed relevant 
Farmer Mac public filings with the Securities and Exchange Commission 
(SEC) and regulatory reporting to the Farm Credit Administration (FCA), 
FCA regulatory reporting instructions, and examined copies of reports 
from Farmer Mac's regulator, external auditors, internal auditors, and 
held discussions with its external counsel and forensic accountants. 
Further, we held numerous discussions with Farmer Mac management and 
staff; FCA officials and examiners; and interviewed representatives of 
the American Bankers Association, the Farm Credit Council, and former 
FCA and Farmer Mac management.

To assess Farmer Mac's financial condition and risk management 
practices, we performed three major steps. First, we reviewed Farmer 
Mac's trends for earnings, capital, and asset (credit) quality, 
including return on average assets, return on common stockholders' 
equity, capital to assets ratio, nonperforming loans as a percentage of 
total loans, and the ratio of allowance for loan losses to 
nonperforming loans. In performing our trend analysis and cost of funds 
analysis, we did not verify the data provided by Farmer Mac. In 
addition, we did not audit Farmer Mac's financial statement or its loan 
loss allowance balances nor did we review any transactions or loan 
files.

Second, we determined how Farmer Mac compares to other entities. To do 
so, we identified appropriate measures of rates of return, capital, and 
asset quality for Farmer Mac and comparable entities. Because of its 
unique role, Farmer Mac does not have any direct peers. However, for 
purposes of our analysis, we determined that the following entities had 
similar characteristics that could be compared to Farmer Mac: Federal 
National Mortgage Association (Fannie Mae), Agricultural Credit 
Association (ACA), Federal Land Credit Association (FLCA), and 
commercial agriculture banks.[Footnote 54] While these organizations 
share some similar characteristics with Farmer Mac, distinct 
differences exist between each of these entities and Farmer Mac. For 
instance, while Fannie Mae is a government-sponsored entity (GSE) and 
publicly traded like Farmer Mac, Fannie Mae deals primarily with 
residential housing mortgages, which are less risky than the 
agriculture mortgages held by Farmer Mac. Farmer Mac's agricultural 
mortgages are commercial loans that fund a wide variety of agriculture 
activity (for example, poultry farms or orange groves), while Fannie 
Mae's single-family mortgages represent a fairly homogeneous asset. As 
a result, in the event of foreclosure, farm properties can be harder to 
appraise and more difficult to liquidate than single-family residences.

Like Farmer Mac, ACA and FLCA are both Farm Credit System (FCS) 
institutions and their business is farm related. However, unlike Farmer 
Mac, they originate loans instead of purchasing loans. Also, Farmer Mac 
is a publicly traded institution and therefore subject to SEC 
oversight, whereas ACA and FLCA are not publicly traded institutions. 
Also included in our comparisons are commercial agriculture banks, 
which are banks that have a higher proportion of farm loans to total 
loans than other commercial banks. Commercial agriculture banks 
originate a range of farm-related loans, unlike Farmer Mac, which buys 
or guarantees only agricultural mortgage loans, and does not originate 
loans.[Footnote 55] Due to the significant impact of the 1996 Act on 
Farmer Mac's operations, we analyzed Farmer Mac's financial performance 
for calendar years 1997 through 2002 and used that same period for our 
comparison of Farmer Mac's financial measures to other entities.

Third, we assessed Farmer Mac's risk management practices and exposure 
to credit, liquidity, interest rate, and legal risks. We (1) obtained 
Farmer Mac's written and oral responses to questions on measurement, 
analysis, and mitigation of those risks; (2) reviewed Farmer Mac Board-
approved policies and standards related to those risks; (3) reviewed 
methodologies for determining loan loss reserves, examined existing 
studies of loan performance and research on agricultural loan 
performance conducted by contractors working for FCA and Farmer Mac, 
and interviewed the contractors; (4) received a demonstration of the 
model used by Farmer Mac to measure market risk; (5) analyzed financial 
data relating to the liquidity portfolio, outstanding debt, derivatives 
and total loan portfolio (on-and off-balance sheet); (6) interviewed 
representatives from the investment community; and (7) examined copies 
of reports from FCA, external auditors, and internal auditors and held 
discussions with external counsel and forensic accountants.

To assess Farmer Mac mission accomplishment, we gained general 
background related to agricultural secondary markets and obtained a 
regulatory perspective on Farmer Mac activities from meetings with 
representatives from the U.S. Department of Agriculture (USDA) Economic 
Research Service, FCA's Office of Secondary Market Oversight (OSMO), 
and the Department of the Treasury's Office of Financial Institutions. 
To gain an understanding of the lenders' perspective on Farmer Mac's 
programs, we interviewed agricultural real estate lenders and banking 
associations. We also compared lending institutions' market share in 
the agricultural real estate market with their percentage of 
participation in Farmer Mac's programs. We measured the amount of 
Farmer Mac's secondary activity by analyzing Farmer Mac's portfolio 
growth by identifying growth by product type and the ratio of retained 
agricultural mortgage-backed securities (AMBS) to AMBS that are sold to 
investors. In addition, we compared average long-term fixed interest 
rates offered by Farmer Mac with average rates offered by agricultural 
real-estate lenders. To the extent possible, we relied on publicly 
available data; therefore, there could be some inconsistencies with 
some of the characteristics of the data sets used to compare interest 
rates.

To evaluate Farmer Mac's corporate governance practices, we reviewed 
Farmer Mac's enabling legislation to understand the legal authority, 
oversight, and structure of Farmer Mac and its Board of Directors. We 
analyzed the Sarbanes-Oxley Act of 2002, the recently proposed New York 
Stock Exchange (NYSE) listing standards, and spoke with NYSE 
representatives to identify the requirements that Farmer Mac would need 
to meet. We reviewed relevant GAO reports and other related literature, 
and attended relevant seminars to gain a better understanding of 
corporate governance best practices. We conducted structured interviews 
with all 15 members of Farmer Mac's current Board of Directors to 
obtain their perspectives on board governance and communication with 
management. Further, we reviewed selected information packages prepared 
for board members and board minutes. To evaluate Farmer Mac's executive 
compensation, we obtained, compared, and analyzed two consultant 
reports on Farmer Mac's compensation and stock option and vesting 
program policies. We compared Farmer Mac's executive packages to the 
housing GSEs. We also reviewed compensation policies for senior 
officers. In addition, we interviewed the corporate governance 
consultant retained by Farmer Mac to obtain her views on Farmer Mac's 
governance structure and practices.

To evaluate FCA's oversight of Farmer Mac, we reviewed examination 
scope and reports on Farmer Mac from 1999 through 2002. We reviewed 
Farmer Mac year-end 2001 and 2002 call reports and compared the 
instructions to the schedules and its legal requirements. We examined a 
copy of the spreadsheet model used by FCA to measure Farmer Mac's 
credit risk, examined the computer programs and data, which produced 
FCA's credit risk model, and interviewed the FCA contractors who built 
the model. Additionally, we examined regulations promulgated by other 
GSE regulators, such as Office of Federal Housing Enterprise Oversight 
(OFHEO) and the Federal Housing Finance Board, and we met with 
officials from OFHEO and the Department of Housing and Urban 
Development (HUD) to understand their examination programs.

We conducted our work in California, Indiana, New York, Virginia, and 
Washington, D.C., between August 2002 and May 2003 in accordance with 
generally accepted government auditing standards.

[End of section]

Appendix II: Farmer Mac's Programs and Products:

Program: Farmer Mac I [A]: Cash Window Program; Program Description: 
Sellers receive cash by selling 100 percent of qualifying first 
mortgage agricultural real estate loans directly to Farmer Mac; 
Product feature: Terms and rates are described below under the 
Full-Time Farm, Part-Time Farm, and AgVantage Programs.

Program: Farmer Mac I [A]: Full-Time Farm Program; Program 
Description: Designed for borrowers who live on agricultural 
properties and derive a significant portion of their income from farm 
employment; Product feature: Types of agricultural 
loans offered include: * 15-year fixed rate, 15-year maturity with 15-
or 25-year amortization and partial open prepayment available (annual, 
semiannual, or monthly payments); * 10-year fixed rate, 10-year 
maturity fully amortizing (semiannual or monthly payments); * 5-year 
reset loan with a 5-year term (renewable twice); 5-, 10-, 15-, or 25-
year amortization (annual, semiannual, or monthly payments); * 30-day, 
1-, 3-and 5-year ARMs (convertible to long-term, fixed rate), 15-year 
maturity, 15-or 25-year amortization, (semiannual or monthly payments); 
and; * facility loans, 10-or 15-year fixed rate maturity, and fully 
amortized.

Program: Farmer Mac I [A]: Part-Time Farm Program; Program Description: Designed for borrowers who live on agricultural 
properties with a valuable residence and derive a significant portion 
of their income from off-farm employment; Product feature: Farmer Mac 
offers a 15-and 30-year loan for single-family, 
detached residences; 3/1, 5/1, 7/1 and 10/1 ARMs and 15-and 30-year 
fixed rate mortgages (monthly payments).

Program: Farmer Mac I [A]: AgVantage Program; Program description: 
Farmer Mac I [A]: Farmer Mac purchases and guarantees timely payment 
of principle and interest on mortgage-backed bonds; Product feature: 
AgVantage bonds may range in maturity from short-term 
to 15 years and have low fixed or variable rates of interest.

Program: Farmer Mac I [A]: Swap Program; Program Description: Farmer 
Mac acquires eligible loans from sellers in exchange 
for Farmer Mac Guaranteed Securities backed by those loans; Product 
feature: Security terms, rates, etc., are negotiated 
with the seller on the basis of the characteristics of the loan.

Program: Farmer Mac I [A]: Long-Term Standby Purchase Commitments; 
Program Description: Farmer Mac commits to purchase 
loans from a segregated pool of loans on one or more undetermined 
future dates; Product feature: Terms are negotiated 
with institution based on the characteristics of the underlying loan.

Program: Farmer Mac II [B]: Cash Window Program; Program Description: 
Lenders receive cash by selling 100 percent of the 
guaranteed portion of USDA loans directly to Farmer Mac; Product 
feature: * 7-year fixed rate and 15-year fixed rate 
based on full amortization; * 5-or 10-year fixed rate based on full 
amortization with 5-or 10-year rate reset periods--which are tied to 
the Farmer Mac 5-or 10-year Reset Cost of Funds Index Net Yield; and; * 
floating rate is tied to Farmer Mac 3-month Cost of Funds Index's "Net 
Yield" with calendar quarter rate adjustments or The Wall Street 
Journal's Prime Rate.

Program: Farmer Mac II [B]: Swap Program; Program description: Farmer 
Mac I [A]: Lenders receive Farmer Mac-guaranteed securities in return 
for the guaranteed portion of USDA loans; Product feature: Farmer Mac 
I [A]: Security terms, rates, etc., are negotiated with the seller on 
the basis of the characteristics of the loan.

Sources: Farmer Mac and FCA.

[A] Farmer Mac I operates as a secondary mortgage market for high-
quality agricultural real estate and rural home mortgages. 
Participation is limited to financially healthy farmers as established 
in the Agricultural Credit Act of 1987.

[B] In the 1990 Act, Farmer Mac was authorized to serve as the pooler 
for secondary sales of agricultural and rural development loans that 
are guaranteed by USDA. This program benefits borrowers who are unable 
to get commercial credit at affordable rates because of financial 
problems.

[End of table]

[End of section]

Appendix III: Financial Trends and Comparisons with Other Entities:

Farmer Mac's increase in impaired loans and in write offs of bad loans 
is indicative of increasing credit risk. Farmer Mac's percentage of 
impaired loans[Footnote 56] to total outstanding post-1996 Act loans, 
AMBS, and standby agreements increased each year from 1997 through 
2001, and then decreased slightly, by 14 basis points[Footnote 57] from 
1.70 percent at December 31, 2001, to 1.56 percent at December 31, 
2002.[Footnote 58] (See fig. 6) On a comparative basis, the proportion 
of Farmer Mac's nonperforming loans to total loans is higher than other 
comparable entities. For instance, Agricultural Credit Associations' 
(ACA) and Federal Land Credit Associations' (FLCA) nonperforming loans 
to total loans at December 31, 2002, were .89 percent and .57 percent, 
respectively. See fig. 7.

Figure 6: Farmer Mac's Impaired Loans from 1997 to 2002:

[See PDF for image]

[End of figure]

Figure 7: Farmer Mac's Nonperforming to Total Loans Compared to Other 
Entities, as of December 31, 2002:

[See PDF for image]

[End of figure]

Farmer Mac's write offs of impaired loans have been limited to date, 
but delinquencies are increasing. During 2002, Farmer Mac wrote off 
$4.1 million of bad loans, or 8 basis points of post-1996 Act loans and 
guarantees,[Footnote 59] which was a significant increase over the $2.2 
million, or 6 basis points, written off in 2001.

Revenue Has Increased, but Some Financial Performance Indicators Lag 
Comparative Entities:

Farmer Mac's net interest income grew from $7.1 million in 1997 to 
$35.0 million in 2002. Net interest income is interest income generated 
from Farmer Mac Guaranteed Securities, loans, and investments, less 
interest expense, which Farmer Mac pays on its debt. Interest rates 
Farmer Mac earned on Farmer Mac Guaranteed Securities and loan products 
declined 177 basis points from 7.41 percent in 1997 to 5.64 percent in 
2002. During the same period, the weighted average interest rates that 
Farmer Mac paid on its debt decreased 216 basis point from 5.75 percent 
to 3.59 percent. The growth in Farmer Mac Guaranteed Securities and 
loans from $442 million and $47 million at year-end 1997 to $1.6 
billion and $966 million at year-end 2002, respectively, caused Farmer 
Mac's interest income to increase. See fig. 8.

Figure 8: Income by Program Assets:

[See PDF for image]

[End of figure]

Farmer Mac's return on assets (ROA) generally increased between 1997 
and 2002, but continued to lag behind other comparative entities. 
During this period, Farmer Mac's performance as measured by percentage 
return on average assets fluctuated from a low of .31 percent in 1999 
to a high of .60 percent in 2002. The increase in 2002 was driven by 
continued growth in the off-balance sheet standby agreement product, 
which experienced 42 percent growth in 2002 and 118 percent growth in 
2001. As previously mentioned, Farmer Mac earns and recognizes income 
from the standby agreements as commitment fees. The standby growth 
caused Farmer Mac's net income growth rate between 2001 and 2002 to 
exceed its average asset growth rate.

During the period 1997 to 2002, Farmer Mac's ROA was consistently lower 
than the ROA of the following comparative banking institutions: Fannie 
Mae (except in 2002), commercial agriculture banks, ACA, and FLCA. This 
indicates that Farmer Mac is using its assets differently than 
comparative banking entities. For instance, of its total assets, Farmer 
Mac had 17.1 percent in cash and 19.7 percent in investments at 
December 31, 2002, while Fannie Mae had 0.2 percent in cash and 6.7 
percent in investments. ACA and FLCA held even lower portions of their 
assets as cash and investments. See fig. 9.

Figure 9: Farmer Mac's ROA Compared to Other Entities:

[See PDF for image]

[End of figure]

Farmer Mac's return on average common stockholder equity (ROE) of 15.04 
percent for 2002 increased steadily from 7.57 percent in 1997. Between 
1997 and 2002, Farmer Mac's ROE remained well below Fannie Mae's ROE, 
which was 30.2 percent for 2002. However, for 2002, Farmer Mac's ROE 
exceeded the comparative banking institutions of commercial agriculture 
banks, ACA, and FLCA. One reason for the difference is that Farmer 
Mac's capital as a percentage of total assets is less than that of the 
comparative banking institutions, but greater than Fannie Mae's capital 
ratio. See fig. 10.

Figure 10: Farmer Mac's ROE Compared to Other Entities:

[See PDF for image]

[End of figure]

Farmer Mac's total capital (stockholder equity) to total assets of 4.35 
percent as of December 31, 2002, is significantly below ACA's and 
FLCA's ratios of 15.81 percent and 16.46 percent, respectively, but 
above Fannie Mae's ratio of 1.84 percent. See fig. 11. Capital's 
primary function is to support the institution's operations, act as a 
cushion to absorb unanticipated losses and declines in asset values 
that could otherwise cause an institution to fail, and provide 
protection to debt holders in the event of liquidation. A higher 
capital to assets ratio, such as ACA's and FLCA's compared to Farmer 
Mac's, indicates there is more coverage for potential financial losses. 
Because Fannie Mae's housing loans have different risks than 
agriculture loans, it is expected that its capital would be lower than 
Farmer Mac's, ACA's, and FLCA's. In general, since 1997, Farmer Mac has 
operated in economic times when agriculture land values have been 
rising and interest rates have been relatively low, experienced minimal 
credit losses, and has not experienced net income losses, so its 
capital has not been stressed and therefore has not demonstrated 
whether it can absorb unanticipated losses and declines in asset 
values.

Figure 11: Farmer Mac's Capital to Asset Ratios Compared to Other 
Entities:

[See PDF for image]

[End of figure]

As of December 31, 2002, Farmer Mac's capital was in excess of its 
statutory requirements. According to the 1991 Amendment to the 
Agricultural Credit Act of 1987 and the 1996 Act, Farmer Mac has the 
following capital requirements:

* Minimum required capital level is an amount of core capital equal to 
the sum of 2.75 percent of Farmer Mac's aggregate on-balance sheet 
assets, as calculated in accordance with generally accepted accounting 
principles (GAAP), plus .75 percent of the aggregate off-balance sheet 
obligations of Farmer Mac, specifically including the unpaid principal 
balance of outstanding Farmer Mac AMBS, instruments issued or 
guaranteed by Farmer Mac, and other off-balance sheet obligations.

* Critical capital level is an amount of core capital equal to 50 
percent of the total minimum capital requirement at that time.

* Core capital is the sum of par value of common and preferred stock 
plus paid-in capital and retained earnings, determined in accordance 
with GAAP.

[End of section]

Appendix IV: Farmer Mac's Underwriting Standards:

Underwriting standards are used by Farmer Mac to determine which 
mortgages it will buy, and then choose to either hold in its own 
portfolios of loans, place into mortgage pools to be sold to investors, 
or place under standby agreements. Generally, eligible loans must meet 
each of the underwriting standards. The standards are meant to limit 
the risk that the mortgages will create losses for Farmer Mac or the 
holders of mortgage pools by ensuring that the borrower has the ability 
to pay, is creditworthy, and is likely to meet scheduled payments and, 
in the event of the default, the value of the agriculture real estate 
limits any losses. Standards are tailored to loans depending upon 
whether the loan is newly originated or seasoned, based upon full-or 
part-time agricultural production, or for specialized facilities. 
Farmer Mac requires lenders originating and selling the loans to (1) 
ensure that loan documentation in each loan file conclusively supports 
determination of each standard and (2) provide representations and 
warranties to help ensure that the qualified loans conform to these 
standards and other requirements of Farmer Mac.

Farmer Mac has nine underwriting standards for newly originated loans 
that are based on credit ratios such as debt-to-assets, other 
quantitative measures such as loan-to-appraised value (LTV), and 
qualitative terms such as receipts supporting agricultural use of the 
property. These standards are a chapter in Farmer Mac's Seller/Servicer 
Guide, and provide guidelines to its staff and lenders, supported with 
detailed examples and explanatory comments for each standard. A summary 
of each of these nine standards is condensed below.

* Standard 1: Creditworthiness of Borrowers. A complete and current 
credit report must be obtained for each applicant and guarantor that 
includes historical experience, identification of all debts, and other 
pertinent information.

* Standard 2: Balance Sheet and Income Statements. This standard 
requires the loan applicant to provide fair market value balance sheets 
and income statements for at least the last 3 years.

* Standard 3: Debt-to-Asset (or Leverage) Ratio. The entity being 
financed should have a pro forma debt-to-asset ratio of 50 percent or 
less on a market value basis. The debt-to-asset ratio is calculated by 
dividing pro forma liabilities by pro forma assets. A pro forma ratio 
shows the impact of the amount borrowed on assets and liabilities.

* Standard 4: Liquidity and Earnings. The entity being financed should 
be able to generate sufficient liquidity and net earnings, after family 
living expenses and taxes, to meet all debt obligations as they come 
due over the term of the loan and provide a reasonable margin for 
capital replacement and contingencies. This standard is achieved by 
having a pro forma current ratio of not less than 1.0 and a pro forma 
total debt service ratio of not less than 1.25, after living expenses 
and taxes. The current ratio is calculated by dividing pro forma 
current assets by pro forma liabilities. Total debt service coverage 
ratio is calculated by dividing net operating income by annual debt 
service. Net income from farm and nonfarm sources may be included.

* Standard 5: Loan-to-Value (LTV) Ratio. The LTV should not exceed 70 
percent in the case of a typical Farmer Mac loan secured by 
agricultural real estate, 75 percent in the case of qualified facility 
loans, 60 percent for loans greater than $2.8 million, or 85 percent in 
the case of part-time farm loans with private mortgage insurance 
coverage required for amounts above 70 percent. The LTV ratio is 
important in determining the probability of default and the magnitude 
of loss.

* Standard 6: Minimum Acreage and Annual Receipts Requirement. 
Agricultural real estate must consist of at least 5 acres or be used to 
produce annual receipts of at least $5,000 to be eligible to secure a 
qualified loan.

* Standard 7: Loan Conditions. The loan (1) must be at a fixed payment 
level and either fully amortize the principal over a term not to exceed 
30 years or amortize the principal according to a schedule not to 
exceed 30 years and (2) mature no earlier than the time at which the 
remaining principal balance (i.e., balloon payment) of the loan equals 
50 percent of the original appraised value of the property securing the 
loan. The amortization is expected to match the useful life of the 
mortgaged asset and payments should match the earnings cycle of the 
farm operations. For facilities, the amortization schedule should not 
extend beyond the useful agricultural economic life of the facility.

* Standard 8: Rural Housing Loans Standards. Farmer Mac has adopted the 
credit underwriting standards applicable to Fannie Mae, adjusted to 
reflect the usual and customary characteristics of rural housing. These 
standards include, among other things, allowing loans secured by 
properties that are subject to unusual easements, having larger sites 
than those for normal residential properties in the area, and having 
property that is located in areas that are less than 25 percent 
developed.

* Standard 9: Nonconforming Loans. On a loan-by-loan determination, 
Farmer Mac may decide to accept loans that do not conform to one or 
more of the underwriting standards or conditions, with the exception of 
standard 5. Farmer Mac may accept those loans that have compensating 
strengths that outweigh their inability to meet all of the standards. 
Examples of compensating strengths include substantial borrower net 
worth or a larger borrower down payment. The granting of standard 9 
exceptions is not intended to provide a basis for waiving or lessening 
in any way Farmer Mac's focus on buying only high-quality loans.

[End of section]

Appendix V: Interest Rate Risk:

Asset-Liability Management:

As of December 31, 2002, over 70 percent of Farmer Mac's liabilities 
($2.9 billion) were short-term--maturing in 1 year or less--while most 
of the assets it held were agricultural real estate mortgages, which 
can have maturities of up to 30 years. As most of these longer-term 
assets are either fixed-interest rate loans or loans with adjustable 
rates that will adjust more than 1 year in the future, this would 
result in an asset liability mismatch, which would occur when assets 
and liabilities do not have the same maturity or interest rate 
characteristics. Farmer Mac's use of interest rate swaps substantially 
reduces this problem. In addition, Farmer Mac uses callable debt to 
mitigate the risk from prepayable mortgages.

Farmer Mac is subject to interest rate risk on its portfolio due to the 
potential timing differences in the cash-flow patterns of its assets 
and liabilities. Farmer Mac uses callable debt, derivatives and yield-
maintenance terms in its loan contracts to mitigate interest rate risk 
(IRR).[Footnote 60] Financial institutions often match the cash flow 
and duration of newly acquired assets with liabilities of equal cash 
flow and duration. In order to achieve an overall lower cost of funding 
for the assets it purchases, Farmer Mac relies on short-term discount 
notes as its primary source of funding. However, since funding longer-
term assets with short-term liabilities causes an asset-liability 
mismatch, Farmer Mac enters into derivative contracts to convert the 
short-term discount notes into longer-term liabilities, which more 
closely match the duration of the assets. The majority of Farmer Mac's 
interest rate contracts are floating to fixed-interest rate swaps, in 
which Farmer Mac pays fixed rates of interest to, and receives floating 
rates from, the derivative counterparty. If interest rates were to 
rise, Farmer Mac would have to pay higher rates when its discount notes 
matured and had to be reissued, but the interest it receives from the 
swaps would also rise, compensating Farmer Mac for the increased 
funding cost. Farmer Mac also enters into basis swaps in which it pays 
variable rates of interest based on its discount notes, and receives 
variable rates of interest based on another index, such as 
LIBOR.[Footnote 61] Farmer Mac also has prepayment penalties or yield-
maintenance terms on 57 percent of its outstanding balance of loans and 
guarantees (including 91 percent of loans with fixed-interest rates), 
which limits Farmer Mac's exposure to losses stemming from declines in 
interest rates. Prepayment penalties and yield-maintenance agreements 
reduce the borrower's incentive to refinance into a lower interest rate 
loan when interest rates drop, and produce additional revenue for the 
owner of the mortgage if it is refinanced at a time of falling interest 
rates.

Prepayment Model:

Prepayment models are an important component of interest-rate risk 
measurement. Approximately 57 percent of Farmer Mac's loan portfolio 
has some form of yield-maintenance protection, which mitigates the 
effects of loan prepayments. The fixed-rate loans that do not have 
yield maintenance expose Farmer Mac to prepayment risk. This is 
particularly true for the purchases of large portfolios of loans (bulk 
purchases) that include loans with characteristics different from the 
rest of the portfolio. For fixed-rate loans without yield-maintenance 
agreements, falling interest rates result in a loss for the financial 
institution if the mortgage is paid off early, as the owner of the 
mortgage can only reinvest the funds at a lower interest rate if the 
mortgage is paid off early. For fixed-rate loans with yield-maintenance 
agreements, falling rates may result in a gain for the financial 
institution, as any loans that do pay off early will pay a penalty that 
generally compensates the lender for the lower interest rate received 
on the reinvested funds.[Footnote 62] Prepayment models predict the 
number and timing of early payments, hence, the losses or gains that 
may result from changes in interest rates.

Farmer Mac's prepayment risk model was developed internally based on 
models that predict prepayment behavior for residential mortgage 
borrowers. Farmer Mac followed this approach due to the unavailability 
of external data on agricultural mortgage prepayments. But agricultural 
real-estate borrowers may behave differently than residential mortgage 
borrowers for many reasons. First, Farmer Mac's fixed-rate agricultural 
real-estate loans often have prepayment penalties or yield-maintenance 
agreements, which are rare for single-family residential borrowers. 
Therefore, at times of falling interest rates, single-family mortgages 
will experience waves of refinancing induced prepayment which will be 
absent for many types of agricultural mortgage. In addition, single-
family borrowers are influenced by price appreciation on single-family 
housing, and agricultural real estate may have significantly different 
patterns of price appreciation. Single-family prepayments are also 
determined in part by mobility and the sale of owner-occupied housing, 
and agricultural real estate may show different patterns of sale-
induced prepayment over time. Farmer Mac makes substantial downward 
revisions to prepayment speeds for loans with penalties or yield 
maintenance, but these adjustments are not based on a model of borrower 
behavior. Rather, they are based on long-run historical averages for 
prepayments on similar loan types. For loans that allow open 
prepayment, Farmer Mac uses a multiplicative adjustment factor applied 
to the prepayment speeds of single-family mortgages. These revisions to 
prepayment speeds more closely align the prepayment behavior of single-
family mortgages with the loans held or securitized by Farmer Mac. The 
adjustment factors are backtested over several previous quarters to 
ensure that they fit the recent past and are revised from time to time. 
However, because single-family prepayment rates fluctuate, sometimes 
substantially, for different reasons than do prepayment rates on 
agricultural mortgages, a simple proportional adjustment factor may be 
insufficient to capture the differences in prepayment behavior. For 
example, if agricultural real-estate prices were flat or falling while 
single family homes were appreciating rapidly, single-family 
prepayments may rise without a corresponding increase in agricultural 
prepayment rates, or vice versa. If the relative rate of agricultural 
mortgage prepayments to single-family mortgage prepayments were 
different for prepayments caused by property sales (which predominate 
at times of flat interest rates) than for prepayments caused by 
refinancing, a proportionate adjustment factor calculated at a time of 
flat interest rates would not provide a good forecast of agricultural 
mortgage behavior when rates are falling.

Loans with prepayment penalties are likely to experience higher default 
probabilities at times of falling interest rates. Yield-maintenance 
penalties have the effect of increasing the loan's payoff amount in a 
falling interest rate environment. This has an effect similar to an 
increase in the LTV ratio, a prime determinant of default in studies of 
borrower behavior.[Footnote 63] As a concrete example of this effect, 
consider a $700,000 loan on a $1,000,000 property. If the agricultural 
market is stressed, and the value of the farm falls to $800,000, a 
borrower may consider selling the property and using the proceeds to 
pay off the loan. If interest rates have fallen; however, and the loan 
payoff additionally includes a $150,000 prepayment penalty, the 
borrower would be unable to pay off the loan with the proceeds from the 
sale of the property and would therefore be more likely to default or 
to negotiate a costly restructuring. Farmer Mac's IRR model assumes 
that default behavior does not change when interest rates change, hence 
does not model an increased probability of failing to collect yield 
maintenance or prepayment penalties in times of falling rates.

Farmer Mac's IRR Measurement Process:

On a monthly basis, or more frequently if necessary, Farmer Mac 
measures its IRR using an industry standard package, Quantitative Risk 
Management (QRM).[Footnote 64] The primary IRR metric that is reported 
to the Farmer Mac board of directors is MVE-at-risk. Farmer Mac 
calculates MVE by first obtaining the market prices of Farmer Mac's 
assets, liabilities, and off-balance sheet obligations. Then Farmer Mac 
uses QRM to calculate the sensitivity of MVE to parallel changes of the 
Treasury yield curve of plus and minus 100, 200, and 300 basis 
points.[Footnote 65] In addition, on a quarterly basis, Farmer Mac 
management analyzes the effect that changes in interest rates have on 
the financial value of Farmer Mac. Farmer Mac management also managed 
NII in a similar fashion as MVE. Finally, Farmer Mac also measures the 
duration gap of its assets, liabilities, and off-balance sheet 
obligations. Other sensitivity analyses are done on a regular basis, 
such as examining the effects of changes in the prepayment speed 
assumptions for mortgages underlying the AMBS.[Footnote 66]

[End of section]

Appendix VI: Farm Credit Administration Credit Risk Model:

FCA measures the credit risk component of Farmer Mac's risk-based 
capital requirement with a statistical model that relates loan 
characteristics, such as the loan-to-value (LTV) ratio, and changes in 
agricultural real estate prices, to credit losses on loans secured by 
agricultural real estate. The estimated relationship between credit 
losses and the prediction variables is used to forecast the losses 
expected on agricultural real estate mortgages under a severe stress 
scenario, such as that experienced in Minnesota, Iowa, and Illinois in 
1983 and 1984.

The data used to estimate the credit loss model consist of loans from 
the Farm Credit Bank of Texas (FCBT) observed over the period 1979 to 
1992. This data source was identified by FCA's consultants who found 
that FCBT had the most reliable loan data for agricultural mortgage 
losses for the purpose of building the credit risk model to estimate 
Farmer Mac's credit risk. The data include several important 
underwriting variables: the LTV ratio, the ratio of the borrower's debt 
to the borrower's assets, and the ratio of the borrower's debt payments 
to farm income. The data also contain the dollar amount of the loan, 
the year in which the loan was written, the year in which the loan was 
foreclosed (for those loans that completed foreclosure), and the amount 
that was lost on the foreclosed loan. The data files used by the 
contractors did not contain information on other key variables, such as 
the amortization period of the loan, the interest rate on the loan, or 
an indicator of whether the loan was paid off early.

The model consists of three equations, estimated sequentially. In the 
first equation, the loss frequency equation, the probability that a 
loan will experience a credit loss at any point over its life is 
predicted by three underwriting variables--the LTV ratio, the debt-to-
asset ratio, the debt payment to farm income ratio--the dollar amount 
of the loan (in inflation adjusted dollars), and the maximum percentage 
decline in farmland value experienced over the life of the loan. 
Logistic regression is used to model the probability of a credit loss. 
Several of the explanatory variables are modified for use in the 
regression. The LTV ratio is raised to a power, the dollar amount of 
the loan is modified with an exponential function, and the decline in 
farmland value is adjusted downward with a multiplicative factor that 
varies with the age of the loan. The second equation multiplies the 
loss frequency by a loss severity, assumed to be a constant 20.9 
percent. The final equation uses a beta function to distribute the 
product of loss frequency and loss severity over time, so that the 
losses expected over the remaining lives of the loans may be isolated.

Data Limitations:

FCA's contractors and we have identified several shortcomings in the 
data used to estimate the credit risk model, including: (1) data have 
not been updated with post-1992 loan information; (2) data may not have 
captured all the credit losses experienced by the FCBT; (3) the data 
set consists entirely of loans in Texas; and (4) the data set does not 
contain information on prepayments.[Footnote 67] These shortcomings 
were noted by FCA's contractors in the Federal Register (Final Rule) 
document that presents the credit risk model. FCA's contractors told us 
that despite these flaws they believe this data set represents the best 
data available for estimating a credit risk model in a stressed time 
period.

We have identified other data shortcomings, which were not indicated in 
the Federal Register risk-based capital document. These include: (1) 
the FCBT data systems did not record all the components of loss on 
foreclosed loans; (2) the loans made by the FCBT from 1979 to 1992 had 
very different interest rate terms than the most common loans bought by 
Farmer Mac; and (3) the data set does not include other important 
predictors of credit loss, such as interest rate or amortization terms. 
These shortcomings limit the ability of the credit risk model to 
forecast the credit risk on loans held by Farmer Mac.

Restricting the data set to 1979 through 1992 creates the possibility 
that credit losses on the loans used in the data will be missed. For 
example, a 15-year loan originated in 1990 may experience a credit loss 
in any year from 1990 to 2005, but only credit losses that occur in 
1990 to 1992 will be predicted by the regression.[Footnote 68] Updating 
the data set with post-1992 borrower behavior would allow more credit 
losses to be observed in the data. Because a longer history is 
available for older loans, it is likely that fewer credit losses are 
missed on older loans than on newer loans. Because a key predictor, the 
greatest decline in land prices, varies with the age of loan, the 
result is likely to be a biased regression coefficient for this 
variable.

The data systems in use by FCBT did not identify all the loans that 
resulted in losses to the bank. Some loans that were merged or 
restructured may have resulted in losses to the bank, but these losses 
are not captured by the foreclosure variable used in the credit risk 
model. Thus, the frequency of credit losses may be understated in the 
model's forecast.

Additionally, the data system did not record the time value of money 
foregone during foreclosure, a process that could take 2 years. This 
has two implications for the credit risk model. First, some 
foreclosures, which appear to have had no credit loss may, in fact, 
have resulted in credit losses. Thus, the frequency of a credit loss 
would be understated in the model's forecast. Second, the model's 
estimate of severity given default may be different than the historical 
average. To the extent that costs are not captured on loans that 
resulted in credit losses, the calculated severity will understate the 
loss severity actually experienced by the bank. To the extent that some 
loans are excluded from the severity analysis because the database 
recorded that they had no credit losses, severity may be either 
understated or overstated, depending on the magnitude of the severity 
for these loans. The data used by FCA's contractors indicated that 62 
percent of the loans that went through foreclosure had no credit loss 
recorded. While the number of loans that may have been misidentified as 
having no losses is not known, it is potentially large.

The data set contains only Texas loans. Previous work by FCA's 
contractors indicates that a region consisting of Minnesota, Iowa, and 
Illinois was the area that experienced the highest level of stress as 
legally defined for FCA's credit risk test, and that Texas was the 
fourth most severely stressed geographic region in the mid-
1980s.[Footnote 69] Hence, the model must extrapolate credit losses to 
a stress situation beyond that contained in the data used to estimate 
the model. The form of extrapolation used in FCA's credit risk model 
assumes that there is a straight line relationship between land price 
declines and a function of the probability of credit loss. Without data 
on loans that experienced property price declines akin to those in the 
most stressed region of the country, it is impossible to know if the 
true relationship is linear or nonlinear.[Footnote 70]

The data used by FCA's contractors did not include information on 
whether or when the loan was prepaid. This has several consequences for 
the credit risk model. The model assigns the land value decline that 
occurred between 1985 and 1986 to any loan written between 1979 and 
1986 that had not entered foreclosure by 1986. It is possible that some 
of these loans had refinanced by 1985 as interest rates declined, so 
that these loans would not have been exposed to the 1985 and 1986 land 
price change. For these loans, the regression would be predicting the 
probability of a credit loss on a loan using a value for the predictor 
that occurred after the loan had been paid off. The lack of information 
on loan prepayment also precludes the measurement of the impact of loan 
duration on the probability of credit loss. It is likely that a loan 
that was active for 10 years is more likely to experience a credit loss 
than is an otherwise identical loan active for only 2 years, as it is 
exposed to the potential of adverse events for a longer time. But the 
data do not identify which loans were active for only 2 years versus 
those active for 10 years. To the extent that loans with lower credit 
risk as measured by underwriting variables, such as lower LTV ratios, 
are more likely to prepay, the underwriting variables in the regression 
are likely to capture both the direct effect of the underwriting 
variable on the probability of credit loss, and an indirect effect 
caused by the tendency of these higher credit quality loans to prepay 
more often; hence, be exposed to risk of a credit loss for a shorter 
period of time.[Footnote 71]

The loans now purchased by Farmer Mac have different interest rate 
terms than those used in FCA's credit risk model. Over the time period 
covered by the data, Farm Credit System (FCS) institutions, including 
FCBT, made loans with adjustable interest rates, in which the interest 
rate was tied to FCS' cost of funds. The average cost of funds changed 
more slowly than did the prevailing rate of interest, as FCS 
institutions used a mix of short-and long-term debt, and the average 
cost of funds was an average of rates on debt recently incurred and 
debt incurred over several previous years. Because of these interest 
rate terms, when interest rates fell after 1982, many farm credit 
borrowers found it advantageous to refinance their debt with other 
lenders. The mismatch between fixed rate liabilities and variable rate, 
prepayable assets was a cause of the FCS's financial problems in the 
mid-1980s.[Footnote 72] However, the bulk of the loans now purchased by 
Farmer Mac are either rapidly adjusting adjustable-rate mortgages, tied 
to short-term interest rates, or are fixed-rate loans with prepayment 
penalties or yield maintenance agreements. For these loans, there is 
little or no advantage in refinancing when interest rates drop. In the 
case of adjustable-rate loans, the interest rate on the mortgage will 
drop without the need to refinance, and in the case of fixed-rate loans 
the prepayment penalties or yield maintenance agreements increase the 
cost of refinancing, making it less advantageous. As interest rates 
generally declined over the period 1979 through1992 (the high point for 
interest rates was 1982, the low point was 1992), it is likely that a 
larger percentage of the loans in the data set paid off early than 
would be the case for the loans now purchased by Farmer Mac. Therefore, 
the loans purchased by Farmer Mac are likely to be exposed to adverse 
events for a longer time period than the loans used in estimating the 
credit risk model. This would have the effect of understating the 
credit risk capital requirement. The prevalence of yield maintenance 
agreements has another effect on the potential for credit losses in 
Farmer Mac's portfolio. As previously discussed, the fixed-rate loans 
now purchased by Farmer Mac that have yield maintenance agreements are 
likely to experience elevated credit risk in times of falling interest 
rates. A borrower in financial distress is more likely to go to 
foreclosure, and is more likely to impose a severe credit loss, if the 
value of the debt substantially exceeds the value of the collateral. 
After a fall in interest rates, fixed-rate loans with yield maintenance 
agreements will owe substantial amounts in excess of their unpaid 
principal balance. Therefore, these loans are more likely to have total 
obligations (unpaid principal balance plus yield maintenance) that 
exceed the value of the collateral, than would loans of otherwise 
similar characteristics that did not have yield maintenance agreements, 
such as those used in estimating FCA's credit risk model, resulting in 
an underestimate of credit risk by FCA's model.

Because the data set did not contain information on interest rates or 
amortization terms, these variables could not be included in the credit 
risk model regression analysis. Other studies of credit risk have found 
these to be important variables in predicting credit losses.[Footnote 
73] Loans which amortize faster are exposed to adverse events for a 
shorter period of time, and accumulate equity more rapidly, which 
reduces credit risk. Higher interest rates lead to higher payment 
burdens, which can put greater stress on borrower's financial 
resources. Adjustable rate mortgages are subject to "payment shock" in 
which defaults increase after a rise in interest rates, which leads to 
a rise in the mortgage payment.[Footnote 74] Since FCA's model does not 
assign higher credit risk to longer amortization loans, or to 
adjustable-rate loans in times of rising interest rates, Farmer Mac 
could increase its exposure to credit risk by buying more of these 
types of loans, without facing a higher risk-based capital requirement.

FCA's ability to estimate a detailed credit risk model was limited by 
the scarcity of relevant data for agricultural real estate loans. FCA's 
consultants identified the Farm Credit Bank of Texas' data from 1979 to 
1992 as the only available data set of agricultural loans observed 
during a stressed period.[Footnote 75] The data file used by the 
contractors had 19,418 loans, including 180 loans with credit losses. 
In contrast, the Office of Federal Housing Enterprise Oversight's 
(OFHEO) risk-based capital model for Fannie Mae and Freddie Mac thirty 
year fixed rate single-family loans is based on about 15 million loans, 
176,000 of which had credit losses. While none of the loans in FCA's 
model were observed during a stress event as severe as that called for 
in its risk-based capital statute, over 7,000 of the 30-year single 
family fixed rate loans used by OFHEO were observed during the 
benchmark stress event specified by OFHEO's risk-based capital 
legislation.

Model Limitations:

We also have identified several limitations in the form of the credit 
risk model used by FCA. These limitations include: (1) the methodology 
chosen by FCA's contractors; (2) use of an independent variable, 
greatest land price decline, whose value is a function of the event 
predicted by the regression; (3) transformations of the independent 
variables to enhance goodness of fit prior to and independent of the 
calculation of significance tests; and (4) the use of state averages to 
model credit risk on the long-term standby agreements.

FCA's credit risk model uses observations on loans, and predicts the 
probability that a loan will experience a credit loss at some point in 
its life. Many models of mortgage credit risk use a different 
structure, and predict the probability that a loan will experience a 
credit event over a defined time period, such as a quarter or a 
year.[Footnote 76] For example, our model of the Veteran's 
Administration credit subsidy costs and OFHEO's multifamily risk-based 
capital model predict annual probabilities of a credit event, while 
OFHEO's single-family model predicts quarterly probabilities of a 
credit event.[Footnote 77] These models have the advantage of 
accounting for the different level of risk inherent in loans that are 
active for longer or shorter periods, and can readily estimate the 
effects of predictor variables that change over time, such as interest 
rates or the value of collateral. The ability to incorporate such 
variables in the measurement of credit risk is important when the goal 
is to measure the risk of a pool of loans, some of which are new, and 
some of which have been active for a long time. For example, the credit 
risk on a loan originated 5 years ago in a state with a 50 percent rise 
in agricultural real estate prices over that 5-year period is likely to 
have less credit risk than an otherwise identical loan originated 5 
years ago in a state where agricultural real estate prices have 
remained constant. In order to capture the changing credit risk over 
time in a portfolio with seasoned loans, it is necessary to include 
measures of credit risk determinants that change over time.

The credit risk model does incorporate a variable, change in the value 
of agricultural real estate, which changes over time. However, its 
inclusion in FCA's model, which predicts lifetime credit event 
probabilities, instead of annual or quarterly probabilities, leads to 
biased estimates of the effects of land price changes on credit risk. 
The variable is defined as the greatest annual percentage decline in 
agricultural land price in Texas from the year that the loan is 
originated until either 1992 or the year of loan foreclosure, whichever 
comes first. The regression is designed to predict the probability of 
foreclosure with credit losses, but the variable's value is determined, 
in part, by whether the loan enters foreclosure. For example, a 1979 
loan that does not enter foreclosure will be assigned a value of--17 
percent--that being the maximum decline in land prices from 1979 to 
1992. However, a loan that enters foreclosure in 1980 would be assigned 
a value of 7 percent, as land prices rose over the short time that the 
loan survived.[Footnote 78] Thus, the definition of the land price 
variable is not independent of the event being estimated, and its 
estimated coefficient is likely to be biased.[Footnote 79] It can be 
shown that the structure of estimating a lifetime probability, combined 
with the definition of land price change used in FCA's credit risk 
model, can produce a statistically significant coefficient for the land 
price change variable, even if there were no effect of land price 
changes on credit risk.[Footnote 80] The land price change variable is 
the key variable used to extrapolate the stress scenario called for in 
the Farm Credit Act of 1971, as amended, and the regression may be 
estimating its impact in a biased fashion, which would lead to a 
misestimate of the credit risk in Farmer Mac's portfolio.

Significance tests reported with the regression results in the Federal 
Register document, which describes the credit-risk model, are likely to 
be biased in favor of a finding of significance. Several variables are 
transformed in various ways before their coefficients are estimated in 
the logistic regression. The greatest decline in land value variable is 
modified by a "dampening factor," which proportionately decreases the 
absolute value of the variable, the longer the loan is observed to have 
not defaulted. The LTV ratio is raised to a power. The loan size 
(dollar amount) variable is defined as 1 minus the exponential of the 
product of the loan size in thousands and the number -0.00538178, so 
that the variable is close to 0 when loan size is small and rises 
towards 1 as loan size increases. These transformations were not 
estimated as part of the logistic regression. Instead, several values 
were tried for each of these parameters, and the values giving the best 
goodness-of-fit measurements were used to transform the predictor 
variables prior to estimating the logistic regression. Such pre-testing 
leads to inflated tests of significance.[Footnote 81]

The estimated credit risk model is not used directly to produce an 
estimate of the credit risk inherent in Farmer Mac's standby 
agreements. Instead, the average credit risk for loans in each state is 
used as an estimate for the credit risk in the standby agreements. The 
regression-based model cannot be used because key underwriting 
variables for standby agreements are not reported to FCA. The use of 
state averages in place of credit risk calculations based on 
underwriting variables would allow Farmer Mac to purchase standby 
agreements with higher loan-to-value or debt-to-asset ratios; hence, 
higher credit risk, than is contained in their loan portfolio, without 
a commensurate increase in their risk-based capital requirements. 
Although Farmer Mac's current portfolio of standby agreements has, on 
average, lower LTV ratios than its on-balance sheet portfolio, the 
structure of the credit risk model provides Farmer Mac with the 
incentive to shift risk into standby agreements should the risk-based 
capital constraint become binding.

[End of section]

Appendix VII: Comments from the Federal Agricultural Mortgage 
Corporation:

FARMER MAC:

Federal Agricultural Mortgage Corporation 1133 Twenty-First Street 
N.W., Suite 600 Washington, D.C. 20036:

(202) 872-7700 FAX 872-7713:

September 19, 2003:

Thomas M. McCool Managing Director Financial Markets and Community 
Investment U.S. General Accounting Office:

Room 2A14:

441 G Street, N. W. Washington, DC 20548:

Re: GAO Draft Report:

Federal Agricultural Mortgage Corporation:

Dear Mr. McCool:

Thank you for this opportunity to comment on the draft of the above-
referenced General Accounting Office Report (the "Report"). We 
appreciate the cooperation afforded us by the GAO staff who worked on 
the Report. It was gratifying to note that, after the third examination 
and review of the Federal Agricultural Mortgage Corporation (Farmer 
Mac) by GAO over the course of the past six years, no serious 
deficiencies in the implementation of Farmer Mac's programs or mission 
were identified; recommendations were made for enhancing risk 
management procedures in light of the Corporation's continuing growth; 
and a number of positive findings were made about the safety and 
soundness of Farmer Mac's operations and about its progress in 
establishing the secondary market for agricultural mortgages.

In its last report on Farmer Mac, dated May 21, 1999, GAO concluded 
that "Farmer Mac's future viability depends on its growth potential in 
the secondary market for agricultural mortgages and the prospects for 
realizing that potential are unclear.... if FCS institutions or other 
lenders increase participation in Farmer Mac programs, Farmer Mac's 
financial condition could improve." Working with agricultural lenders 
throughout the Nation, Farmer Mac's Board and management have resolved 
favorably the uncertainties raised in that earlier report. Thus, we are 
pleased with the new Report's recognition of Farmer Mac's progress in 
the four intervening years, in terms of both its financial strength and 
the growth of its mission-related activities. At the same time, we 
share GAO's view, expressed in the Report, that Farmer Mac should 
continue to enhance its risk management, for the efficiency of our 
operations will be a key factor in the 
accomplishment of our mission. Farmer Mac has already implemented 
several enhancements to risk management procedures which are described 
later in this letter.

The Chairman and Ranking Member of the United States Senate Committee 
on Agriculture, Nutrition, and Forestry, in their letter request of 
June 26, 2002, sought GAO's assistance "to ensure that Farmer Mac's 
mission continues to be met in a financially sound manner." As to the 
six specific topics GAO was asked to address in that letter, the 
Report 

(1) found no current financial instability at Farmer Mac, noting that: 
independent accountants issued Farmer Mac a "clean," unqualified audit 
opinion; an outside forensic accounting firm determined Farmer Mac's 
reserve levels and methodology were "reasonable" under U.S. GAAP; 
Farmer Mac effectively managed its interest rate risk through asset/
liability matching and yield maintenance protection against prepayment 
risk; and Farmer Mac's controls over credit risks were generally sound, 
but could be improved in light of recent and expected future growth;

(2) found no significant shortcomings in corporate governance; 
acknowledged that Farmer Mac was taking actions to ensure that it 
complies with recent Sarbanes-Oxley Act and NYSE listing requirements, 
as they become effective; and noted areas in which Farmer Mac (like 
other private sector corporations) might have to update its corporate 
governance, including expanded director training;

(3) found executive compensation was in line with the recommendations 
of two independent consultants, but recommended that the timing of 
vesting of stock options be extended, which the Farmer Mac Board has 
done;

(4) found no irregularities in Farmer Mac's investment practices or 
strategy and that non-mission investments have been reduced as a 
percentage of mission-related assets;

(5) recommended that Congress reconsider the non-voting status of 
Farmer Mac Class C common stock; and:

(6) noted that Farmer Mac has increased its Congressional mission-
related activities (loan purchases, guarantees and commitments) since 
GAO's 1999 report, and recommended that Congress consider establishing 
more specific criteria for measurement of Farmer Mac's mission 
accomplishment.

In addition, the Report raised a number of problematic or hypothetical 
issues. For simplicity, we have paraphrased those issues in italics 
below with Farmer Mac's observations following them.

(A) It would be preferable if Farmer Mac used its own data instead of 
Farm Credit Bank of Texas (`FCBT') historical data for credit risk 
projections.

While the Report cites the conclusion of a prominent agricultural 
economic consultant to Farmer Mac's federal regulator that the FCBT 
loan data was the "best available" database for the purpose of building 
a model to estimate Farmer Mac's credit risk, it nevertheless concludes 
that it would be better if Farmer Mac were using its own historical 
database. While that conclusion may be theoretically correct, Farmer 
Mac notes that its own database is still relatively new and U.S. 
agriculture has not been through a significant downturn during the 
period it covers. The FCBT database contains loans screened for 
conformity to Farmer Mac credit underwriting standards and reflects 
loan performance during a series of economic events that resulted in 
the most severe loan losses in U.S. agricultural credit history. Until 
its portfolio of mortgages under guarantee and commitment seasons 
further, Farmer Mac expects to continue to use the FCBT database as a 
conservative benchmark for evaluation of its credit risk.

(B) Elements of Farmer Mac's prepayment model are based upon residential 
data, which may not reflectfully the prepayment characteristics of 
agricultural mortgages.

The prepayment model used by Farmer Mac is consistent with models used 
by other agricultural mortgage lenders; Farmer Mac has adjusted its 
model to reflect the differences in the behavior of agricultural and 
residential mortgage borrowers, with validation by a recognized outside 
expert on prepayment modeling; and the accuracy of Farmer Mac's model 
has been confirmed through "back-testing," i.e., verifying model 
forecasts against actual outcomes. No useable agricultural mortgage 
prepayment database exists and, when Farmer Mac's own historical 
database becomes statistically significant, it will revise its 
prepayment modeling accordingly.

(C) With respect to guaranteed AMBS held by Farmer Mac, the U.S. 
Treasury has questioned whether it would be required to allow Farmer 
Mac to draw upon its Treasury line of credit to support those 
guarantees.

While the Report acknowledges that Farmer Mac has a reasoned legal 
opinion of outside counsel stating that the Treasury line of credit 
would be available in those circumstances, the question is moot. No 
issue has been raised as to the availability of its Treasury line of 
credit relative to AMBS held by parties other than Farmer Mac and, if 
Farmer Mac were coming under pressure to fund its guarantee 
obligations, it is confident it could sell to third parties any AMBS it 
held long before it needed to use the Treasury line of credit.

(D) As Farmer Mac's Long-Term Standby Purchase Commitments ("LTSPCs") 
continue to grow, if risks were not closely managed and there were 
massive defaults on those loans far beyond the worst levels experienced 
in U.S. agricultural economic history, the Corporation could be 
required to acquire a high proportion of the outstanding loans covered 
by LTSPCs, resulting in a possible future funding risk.

GAO, in raising this issue, posits a situation in which loan defaults 
go far beyond the 13.8% default rate peak for all agricultural loans 
(including non-mortgage loans) within the Farm Credit System in 1986, 
the worst period in recorded U.S. agricultural economic history. 
Without discussing Fanner Mac's 90-day delinquency rate on LTSPCs 
referenced in the Report at 0.1%, we note that the additional funding 
required to acquire loans even in the implausible volume suggested in 
the Report would not be inconsistent with Farmer Mac's current levels 
of debt issuance in the capital markets.

(E) Farmer Mac creates "regulatory capital arbitrage "for lenders as a 
consequence of the application of its guarantee or commitment by 
reducing the regulatory risk weight assigned to mortgages by 80%.

The regulatory capital levels required (by OCC and FCA) for 
agricultural primary lenders take into account that the loans they hold 
include unsecured loans, loans secured by chattel (including crops in 
the ground), second mortgage loans, and first mortgage loans; only a 
subset of the last category would be eligible for a Farmer Mac 
guarantee or commitment. Rather than arbitrage, the reduction of 
capital requirements for first mortgage loans that bear Farmer Mac 
credit enhancements is appropriate to the reduced risk inherent in 
those loans, analogous to the regulatory capital treatment of 
residential mortgage loans credit enhanced by other GSEs. Comparison of 
non-stress test capital requirements for loans held by primary lenders 
to Fanner Mac's statutory minimum capital requirement is irrelevant. 
Farmer Mac's stress test based risk-based capital model is the 
appropriate means of determining the capital needed to support loans in 
Farmer Mac's programs. Farmer Mac is required to maintain the higher of 
statutory minimum and risk-based capital.

Fanner Mac recognizes that there are a number of areas in its business 
that will continue to need attention. Independent of, but consistent 
with the findings and recommendations in the Report, Fanner Mac has 
taken a number of steps to enhance its risk management practices that 
should be mentioned here. First, we have developed a loan 
classification system that is the basis for an internally developed 
capital adequacy model. Farmer Mac measures its capital adequacy 
against this model, in addition to the statutory minimum capital levels 
established by Congress and risk-based capital levels established by 
FCA. We expect to migrate Fanner Mac's loss allowance methodology from 
the current model based on the FCBT data to a methodology based on this 
loan classification system, which reflects Fanner Mac's own historical 
portfolio loss and 
loan performance experience. Second, we have formalized the 
Corporation's long-standing liquidity contingency funding plan, with 
Board action confirming the new plan. Third, we have enhanced 
documentation procedures regarding loan underwriting decisions, 
including those based on compensating strengths and subsequent 
performance of such loans, and regarding loan loss reserve methodology 
and procedures.

Farmer Mac's Board and management have cooperated with GAO throughout 
the process of GAO's research and drafting of the Report. We look 
forward to continuing the fulfillment of this Corporation's 
Congressional mission and are pleased to have been given this 
opportunity to share our thoughts on these important matters.

Very truly yours,

Henry D. Edelman 

President:

Signed by Henry D. Edelman: 

cc. Ms. Davi D'Agostino, Director, Financial Markets and Community 
Investment:

The following are GAO's comments on the Federal Agricultural Mortgage 
Corporation's letter dated September 19, 2003.

GAO Comments:

1. Farmer Mac commented that its own loan portfolio database is 
relatively new and that the U.S. agriculture has not been through a 
significant downturn during the period it covers. Further, Farmer Mac 
expects to continue to use the FCBT database as a conservative 
benchmark for evaluating credit risk. While it may be true that during 
the period of time Farmer Mac has been accumulating information to 
develop its own loan database, the U.S. agricultural industry has not 
faced a similar catastrophic decline as that experienced during the 
1980's as captured in the Texas data, we disagree with Farmer Mac's 
inference that its portfolio is too new to provide loan loss experience 
from which to estimate credit losses. Farmer Mac has been buying and 
retaining its portfolio of loans for over 7 years, and has been 
executing its guarantees under standby commitments for over 3 years. 
Accounting industry guidance suggests, "Two to three years of lending 
experience normally would provide data that is more relevant than peer 
group experience." Further, because Farmer Mac's loan portfolio has 
characteristics, which differ from the FCBT data used in the model, and 
quantification of the effect of these differences-whether it would 
increase, decrease, or have no material impact to the allowance-has not 
been made by Farmer Mac, we believe that Farmer Mac should use the more 
relevant data. Farmer Mac asserts, however, that the FCBT is a more 
conservative tool to benchmark the allowance because it includes an 
economically depressed time period. In fact, the loan loss allowance 
should reflect current environmental factors and conditions that could 
cause probable future losses rather than the most severe loss situation 
in history. We believe that the most appropriate approach would be for 
Farmer Mac to use its own data, which provides relevant and comparable 
loan characteristics, in its loan loss methodology while also applying 
appropriate "stress testing" approaches to reflect any potential or 
likely future downturns or economically depressed conditions.

2. Farmer Mac commented that an outside expert on prepayment modeling 
has validated the adjustments that Farmer Mac made to its prepayment 
model, and that since no useable agricultural mortgage database exists, 
Farmer Mac will revise its prepayment modeling accordingly when its 
historical database becomes statistically significant. In making that 
comment, Farmer Mac seems to disagree with our recommendation that it 
should improve the quality of its prepayment model to ensure accurate 
measurements of interest rate risk. However, as stated in our report, 
Farmer Mac management noted that they are currently working with an 
outside expert to develop an agricultural mortgage prepayment model to 
better model prepayment risk.

3. Farmer Mac commented that the reduction of capital requirements for 
mortgage loans that bear Farmer Mac credit enhancements is not 
arbitrage but is analogous to the regulatory capital treatment of loans 
enhanced by Fannie Mae and Freddie Mac guarantee or commitment. 
Referring to table 2 in the report, Farmer Mac commented that comparing 
Farmer Mac's statutory capital minimum requirement to the capital 
requirement for primary lenders is irrelevant and stated that Farmer 
Mac is required to maintain the higher of statutory minimum and risk-
based capital. First, because all of Farmer Mac's current participants 
in standby agreements are FCS institutions (another GSE), the report 
discusses the potential reduction of the sum of capital required to be 
held by the Farm Credit System and Farmer Mac without a corresponding 
reduction in risk. In this regard, a reduction in capital requirements 
for loans bearing Farmer Mac credit enhancements is not analogous to 
the housing GSEs because these GSEs are enhancing loan credit from 
commercial lenders, not another GSE. The intent of table 2 is not to 
compare the capital levels of Farmer Mac with primary lenders, but 
rather, to demonstrate the reduction of capital for loans enhanced by 
Farmer Mac guarantee or commitment. We agree and the draft report 
clearly states that Farmer Mac must meet the higher of statutory 
minimum or risk-based capital requirement. As such, we have analyzed 
the risk-based capital model and have identified some limitations that 
are discussed in the report.

[End of section]

Appendix VIII: Comments from the Farm Credit Administration:

Farm Credit Administration:

1501 Farm Credit Drive McLean, Virginia 22102-5090 (703) 883-4000:

August 21, 2003:

[See PDF for image]

[End of figure]

Ms. Davi M. D'Agostino Director, Financial Markets and Community 
Investment United States General Accounting Office 441 G Street, NW:

Washington, DC 20548:

Dear Ms. D'Agostino:

The Farm Credit Administration (FCA) appreciates the opportunity to 
comment on the General Accounting Office's (GAO) draft report entitled 
Farmer Mac: Greater Attention to Risk Management, Mission, and 
Corporate Governance is Needed (the Report). While we have some 
suggestions and clarifying comments regarding conclusions reached about 
FCA, overall we believe the Report is a fair representation of our work 
to oversee the safety and soundness of the Federal Agricultural 
Mortgage Corporation (Farmer Mac or Corporation).' Moreover, we expect 
the Report to add value to our work on several initiatives already 
underway at FCA.

The Report includes the following five recommendations for FCA's 
oversight of Farmer Mac:

* consider potential improvements to the Risk-based Capital Stress Test 
(RBC model);

* improve and formalize offsite monitoring including regulatory 
reporting;

* reduce potential safety and soundness concerns that may arise from 
"capital arbitrage" activities of Farmer Mac and Farm Credit System 
(FCS or System) banks and associations;

* examine how other secondary market regulators developed regulations 
to require government-sponsored enterprises to obtain a risk rating 
from nationally recognized statistical rating agencies; and:

* assess and report on the impact of Farmer Mac's activities on 
agricultural real estate lending markets.

FCA will fully consider and incorporate the Report's recommendations 
into its oversight of Farmer Mac both through current regulatory and 
examination work in process and, as necessary, new initiatives.

GAO recognizes that work relating to several items discussed in the 
Report is currently underway within FCA. This work includes two 
projects on FCA's current regulatory agenda addressing Farmer Mac's 
liquidity and nonprogram investments, and revisions to the RBC model.	
Further initiatives include plans for updating Call Report formats and 
instructions to 
improve offsite monitoring, and possible regulatory action on two items 
- the capital issues arising when System institutions enter into Long-
term Standby Purchase Commitments (Standbys) with Farmer Mac, and FCS 
credit exposure to guarantors that do not have a credit rating, such as 
Farmer Mac. As part of that process, we will examine how other 
regulators have imposed credit rating requirements.

In addition, the Report includes suggestions for Congressional 
consideration. Regarding Farmer Mac's mission, GAO suggests Congress 
establish clearer mission goals for Farmer Mac. It also suggests 
Congress allow FCA more flexibility to establish capital standards 
commensurate with Farmer Mac's changing risk profile and in setting 
minimum capital standards. We support these suggestions and will assist 
Congress as they are considered. In addition, we will reexamine 
authorities provided in existing legislation for FCA to pursue these 
suggestions.

Without making a recommendation, the Report points out a potential 
conflict of interest resulting from a single regulator overseeing both 
primary and secondary market institutions. The Report also notes FCA's 
awareness of this potential conflict and alludes to our belief that we 
are successfully managing any associated risk. In response, we 
emphasize that FCA is committed to delivering the highest standard of 
financial institution supervision. Congress established FCA's Office of 
Secondary Market Oversight (OSMO) as a separate office to oversee 
Farmer Mac with foresight of the issue raised by the Report. FCA 
continues to preserve the independence of OSMO within its 
organizational structure. FCA will periodically reassess the changing 
dynamics involved in regulating both FCS lending institutions and 
Farmer Mac and will remain vigilant in addressing any concerns arising 
from this dual responsibility.

FCA's comments on the recommendations regarding the RBC model are 
enclosed. These comments focus on the unique aspects of risk analysis 
of agricultural credit and the scientific properties of the model we 
developed under guidance provided in Title VIII of the Farm Credit Act 
of 1971, as amended (Act). Generally, FCA would stress that the RBC 
model was developed based on reasoned and conservative judgment, the 
best available data, accepted econometric methodologies, transparent 
procedures, and, most significantly, in conformance with the governing 
statutory provisions of the Act.

As envisioned by FCA in the RBC model's final rule, ongoing changes to 
the model are anticipated based on: developments in regulatory 
guidelines (e.g., New Basel Accord); new types of business activity; 
institutional change; financial innovations; new data sources and 
availabilities; refinements in methodologies; financial market 
conditions; and statutory amendments. As part of our planned regulatory 
project on revisions to the RBC model, we also will consider GAO's 
findings and recommendations.

Again, we thank you for the opportunity to provide these comments and 
the technical comments that were provided separately. We hope you find 
them useful as the final report is published.

Sincerely,

Signed by: 

Michael M. Reyna: 
Chairman and Chief Executive Officer:

Enclosure:

Enclosure:

Farm Credit Administration (FCA) Comments Specific To The Risk-based 
Capital Stress Test (RBC model) As Discussed In The General Accounting 
Office (GAO) Draft Report Entitled Farmer Mac: Greater Attention To 
Risk Management, Mission, And Corporate Governance Is Needed (the 
Report):

FCA receives the recommendations in the Report related to the RBC model 
as constructive commentary. Nevertheless, below are several responses 
that offer an expanded perspective on specific issues raised in the 
Report. We trust our comments will add value for users of the Report.

Steady-state Approach:

GAO questions FCA's use of a steady-state approach to the RBC model, 
which necessitates future earnings assumptions, in contrast to the 
Office of Federal Housing Enterprise Oversight's (OFHEO) use of a run-
off approach. It is important to note that OFHEO's statute has specific 
requirements regarding new business and earnings assumptions. The 
authorizing statute for FCA, the Farm Credit Act of 1971, as amended, 
(Act) is silent on the issue. Thus, FCA has the authority to interpret 
the Act and use whatever approach is reasonable and produces a 
stressful model that is most suitable for Farmer Mac and agricultural 
loans.

FCA thoroughly considered whether to use a steady-state or run-off 
approach. Although GAO questions the use of earnings estimates, FCA 
found in developing the model that using a steady-state approach 
resulted in our having to make fewer assumptions than would have been 
required by a run-off model.

FCA also believed that the Act is best read to treat Farmer Mac as a 
going concern. Section 8.32 of the Act states that the test must 
determine the amount of regulatory capital necessary for Farmer Mac to 
maintain positive capital during a 10-year stress period. The 
requirement that Farmer Mac maintain positive capital during the 10-
year period implies that Farmer Mac would remain a going concern during 
that period.

Congress considered both FCA's and OFHEO's mandate to develop a risk-
based capital test at the same time and chose to include specific 
instructions for OFHEO on this issue, while leaving FCA room to 
determine the best approach for Farmer Mac. Although GAO economists 
favor a run-off approach, FCA believes we adopted the best approach at 
the time given our statutory mandate.

Data Limitations:

GAO recommends FCA obtain "more relevant and current data" to support 
the loan loss estimation in the RBC model. Regarding relevancy, we 
would note that GAO staff made inquiries to identify a more relevant 
data set but were unable to provide FCA with any suggestions for a more 
suitable set of data.

In addition, during the development of the risk-based capital 
regulations, the FCA documented its exhaustive search for data on 
agricultural mortgage losses. FCA requested public input on whether 
more relevant data existed and received no information on other more 
usable, relevant data.

Regarding currency, current data is likely to be of limited use due to 
the constraints of the Act. The most stressful 2-year period in 
agricultural lending is required, but few would argue that such events 
have occurred over the past decade.

The data utilized in the study were the most comprehensive source 
available for farm real estate loans, with a relatively large set of 
loss experiences that reflected the severe farm financial adversities 
of the 1980s. The properties of these data were consistent with 
statutory specifications for estimating worst-case historic conditions 
and allowed estimates of the frequency and severity of loan loss.

Extensive evidence of model validation is found in the proposed and 
final rules for the RBC model and indicates strong explanatory capacity 
and forecast capability in determining risk-based capital requirements. 
Included in the validation process are: tests of logic, judgment, and 
experience of the FCA work team and outside consultants; external 
reviews and feedback; verification of model construction; consistency 
of the magnitude and location of worst-case conditions with findings in 
previous studies and data compilations by FCA, the U.S. Department of 
Agriculture, Federal Reserve economists, and academic economists; 
quality of econometric results; sensitivity of capital requirements to 
changing economic conditions; and comparisons of simulated to actual 
losses.

While we can agree that an improved data set would be beneficial - 
locating, testing, and evaluating new data will be a challenge. Still, 
we are pursing that end with the analysis of post-1992 data from the 
Farm Credit Bank of Texas (FCBT) as described below.

The Report also states that the RBC model did not use post-1992 data, 
which implies that FCA had usable data after 1992 and opted to exclude 
it from the RBC model. At the time of developing the test, the post-
1992 data available were not in a usable form. Our consultants are 
currently working with the FCBT to analyze post-1992 information and to 
render the data usable. Once the post-1992 information is usable, FCA 
intends to incorporate it into the RBC model as appropriate.

Servicing Records:

The Report states that FCA had access to servicing records of the FCBT, 
but that we did not use it in our analysis of the data, except to 
produce estimates of loan loss severity. In fact, the data in the 
servicing records were not detailed payment records. FCA reviewed the 
servicing records and concluded they did not contain information that 
would enhance the quality of the loss estimates. We would note that the 
Report does not suggest how that information would have been useful in 
analyzing the FCBT data or in developing a loss frequency estimation 
regression in the RBC model, nor did GAO review the data in these 
records to our knowledge.

Yield Maintenance Provisions:

The Report asserts that yield maintenance provisions and prepayment 
penalties in loan contracts reduce the borrowers' incentive to prepay 
or refinance loans in periods of falling interest rates, lengthen their 
exposure to credit risk, may lead to loan balances exceeding collateral 
values, and increase capital requirements. However, falling interest 
rates, with other factors held constant, would tend to increase rather 
than decrease present market values of farmland, a nondepreciable asset 
that is the underlying source of collateral for most farm mortgage 
loans. In addition, default rate studies, including the pioneering work 
by Edward Altman, generally indicate that default frequencies are 
considerably higher earlier in the lives of 
loans. Thus, we believe much of the credit risk will have dissipated on 
seasoned loans. These time patterns characterize the FCBT data and the 
RBC model developed by FCA.

The Use of Land Value Decline as an Independent Variable:

The Report suggests that the accuracy of the credit loss regression is 
reduced by using the variable of minimum land price decline whose value 
is a function of the event predicted by the regression. However, this 
position is inconsistent with the theory and empirical evidence 
indicating that stress-induced reductions in anticipated farm income 
cause declines in present values of farmland. That is, stress 
conditions lead to less optimistic expectations of farm income and 
these pessimistic expectations put downward pressure on the present 
values of farmland. Thus, FCA believes the direction of the functional 
relationship in the RBC model is valid.

Credit Risk Not Captured in the Model:

The Report referenced three types of instruments that are not subject 
to credit risk in the model, AgVantage bonds, nonprogram investments, 
and counterparty swaps. When the RBC model regulation was finalized, 
FCA did not believe that AgVantage bonds should be stressed against the 
loan loss portion of the model because the bonds are general 
obligations of the issuing institutions. Although pools of qualified 
loans serve as collateral for the bonds, the ultimate payment of the 
bonds is not solely dependent upon payment on the underlying loans.

Swap agreements have relatively little associated risk compared to 
loans because a counterparty's default results in, at most, the loss of 
a single periodic payment. Although we do not believe current exposures 
are material, we do agree that, as Farmer Mac and its swap portfolio 
grow, the materiality of this item could increase.

Historically, the investment portfolio has had a low proportion of 
assets rated less than "A." Therefore, FCA suggests that credit risk is 
low and currently risk exposure is immaterial on these items, and that 
the Report overstates the impact of their current exclusion. 
Nevertheless, FCA is considering a revision to the RBC model regulation 
to address risk on nonprogram investments. Further, FCA is working 
toward regulating Farmer Mac's nonprogram investments (including 
setting minimum quality standards) and liquidity requirements.

As FCA works to revise the RBC model, questions regarding certain 
statutory constraints have been raised. Having noted GAO's other 
suggestions to Congress, we believe further enhancement to the RBC 
model and overall capital adequacy measurements are possible with an 
adjustment to the Act's prescribed method of shocking interest rates. 
Rather than a one-time shock to interest rates, FCA suggests that 
applying interest rate volatility over the RBC model's 10-year 
estimation period would be both more stressful and realistic. In 
addition, should Congress decide to address GAO's suggestion to expand 
FCA's statutory flexibility to set minimum capital levels, we will be 
pleased to offer other suggestions for establishing minimum regulatory 
capital levels.

[NOTES]: 

[1] Farmer Mac is established separately by Title VIII of the Farm 
Credit Act of 1971, as amended (12 U.S.C. 2279aa-2279cc). Subtitles A 
and B of Title VIII authorize the FCA to examine the Corporation and 
provide for the regulation and general supervision of its safe and 
sound performance, including promulgation of regulatory capital 
standards.

The following are GAO's comments on the Farm Credit Administration's 
letter dated August 21, 2003.

GAO Comments:

Steady-state Approach:

1. FCA commented that it had the authority to use whatever approach is 
reasonable to produce a stressful model that is most suitable for 
Farmer Mac and agricultural loans. It also stated that the 1987 Act is 
best read to treat Farmer Mac as a going concern. We agree and believe 
that our report clearly indicates that FCA had the authority to build a 
risk-based capital test using either a steady-state or a run-off 
approach. However, we do not agree with FCA's view that the statute's 
requirement for positive capital throughout a stress scenario implies a 
preference for a steady-state approach. The Federal Housing Enterprises 
Financial Safety and Soundness Act of 1992 (FHEFSA), which sets the 
requirements for OFHEO's risk-based capital test, also requires 
positive capital throughout a stress scenario, but requires an initial 
run-off approach, followed by mandated studies of the steady-state 
approach.

2. FCA commented that using a steady-state approach resulted in their 
having to make fewer assumptions. We believe that the assumptions 
required for a steady-state approach are difficult to support. The key 
assumption of a steady-state approach is that the volume of new 
business will exactly match the run-off of old business, even during a 
stressed period. Additionally, assumptions concerning the level of 
profitability, or unprofitability, of new business during a stressed 
period must be made in order to implement a steady-state approach. Both 
of these assumptions are difficult to base on data for financial 
institutions in stressed time periods.

Data Limitations:

3. FCA commented that GAO staff were unable to provide suggestions for a 
more suitable data set. In our report, we recognized that FCB Texas 
data was the most comprehensive data source available and did not 
suggest that the FCB Texas data be replaced with a more suitable data 
set. Rather, we recommend that the FCB Texas data be brought current, 
if possible, and that data from other sources be used to model risks 
such as payment shocks on adjustable-rate mortgages or amortization 
terms that cannot be easily modeled with the Texas data. Updating the 
Texas Bank data would improve the credit risk estimation model by 
addressing an issue raised in the report, that loans originated in 1992 
or earlier, some of which would have experienced the land price 
stresses of the mid 1980's, may still result in credit losses after 
1992. Only by updating the data set with post 1992 foreclosures can the 
model capture the lifetime credit experience of these loans.

Additionally, the credit risk model uses a regression framework to 
extrapolate losses based on the Texas stress event to a more severe 
stress event such as that which occurred in the Upper Midwest. The 
extrapolation relies on the slope of the land price decline variable 
estimated by the regression. Since 71 percent of the loans in the Texas 
data file used by the contractors, comprising 176 of the 180 credit 
losses, are associated with a land price decline of 17 percent, and 
another 25 percent of the Texas data are associated with land price 
declines of 2 or 4 percent (these loans have no credit losses), there 
is very little variation in the data used to estimate the slope with 
respect to minimum land price changes. The loans associated with a 17 
percent price decline are all observed for 7 to 13 years after 
origination, while the loans associated with 2 or 4 percent declines 
are all observed for only 0 to 5 years after origination. Augmenting 
the Texas data to include credit losses over less stressful time 
periods should reduce the bias and increase the precision of the 
estimate of the land price decline - credit loss relationship, upon 
which the extrapolation used by FCA is based.

4. FCA commented that the magnitude and location of worst-case 
conditions of its model validation process is consistent with findings 
in studies and data compilations by a number of economists. Therefore, 
it is evident that the FCA model has strong forecasting capability in 
determining risk-based capital requirements. We do not dispute FCA's 
finding that the Upper Midwest in the mid-1980's was a high stress 
event for agricultural real estate. We disagree that FCA has presented 
evidence of the model's forecasting ability. Without post 1992 data on 
loans, there are no data with which out of sample forecasts can be made 
to test the model's forecasting ability. Additionally, we have noted in 
the report that in-sample goodness-of-fit statistics presented with the 
model are likely to be biased, based on the fact that nonlinear 
transformations of certain variables, such as loan-to-value, were 
fitted prior to the estimation of the regression model.

5. We have modified the text of the report to indicate that FCA did not 
have post 1992 data available in a ready to use format, and to 
recognize that FCA is engaged in an effort to incorporate post 1992 FCB 
Texas data.

Servicing Records:

6. In commenting on a section of the draft report that discusses data 
limitations, FCA stated that it had reviewed the servicing records of 
FCB of Texas, which were not detailed payment records, and concluded 
that they did not contain information that would enhance the quality of 
the loss estimates. We modified the text of the report to delete 
references to servicing records.

Yield Maintenance Provisions:

7. In commenting on a section of the draft report that discusses the 
effect of yield maintenance provisions and prepayment penalties in the 
credit risk model, FCA stated that falling interest rates, with other 
factors held constant, would tend to increase rather than decrease 
present market values of farmland. We agree that, with other factors 
held constant, declining interest rates will tend to increase the value 
of agricultural real estate. However, other factors are often not 
constant. For example, a decline in inflation will lower both interest 
rates and anticipated cash flows, so that real estate values will not 
necessarily increase when interest rates decline. FCA's stress test is 
based upon falling Texas land prices in 1985 and 1986. From their peak 
in 1985, Texas agricultural real estate values fell by 25 percent over 
the next 2 years, although the interest rate on 10 year Treasury bonds 
had fallen from 10.6 percent to 8.4 percent over the same time period. 
Additionally, yield maintenance provisions increase the borrower's 
obligation even when interest rates are unchanged, because the present 
value of the spread between the loan rate and the rate on comparable 
Treasury securities must be paid when a loan is terminated. Farmer 
Mac's seller-servicer manual gives an example in which there is an 8 
percent yield maintenance penalty despite unchanging interest rates. 
Because yield maintenance penalties and land prices do not always move 
in equal and opposite proportions, we believe that each should be 
considered as independent variables in a credit risk regression.

8. FCA stated that default rate studies generally indicate that default 
frequencies are considerably higher earlier in the lives of loans and 
that these time patterns characterize the FCBT data and the risk-based 
capital model. We do not agree that default frequencies are higher in 
the early years of a loan's life based on performance of the loans in 
the FCBT data. In this data, no credit losses occurred in the year of 
loan origination, and less than 4 percent of the credit losses occurred 
within the subsequent 2 years. Further, about 25 percent of the credit 
losses occurred 9 years or more after origination and the median year 
of foreclosure in the FCBT data is the 7th. Nevertheless, large yield 
maintenance penalties and substantial refinancing incentives can occur 
early in a loan's life. Therefore, we continue to believe that it is 
important to consider the effects of prepayment and yield maintenance 
when estimating a credit risk model.

The Use of Land Value Decline:

9. FCA referred to a section of the draft of this report that discusses 
how the variable of minimum land price decline affects the accuracy of 
the credit loss regression. FCA commented that GAO's position, which 
suggests the use of this variable would reduce the accuracy of the 
model, is inconsistent with theory and empirical evidence and that the 
direction of the functional relationship in the risk-based capital 
model is valid. We agree that the direction of the effect of land 
prices on credit losses in FCA's credit risk model is consistent with 
theory and empirical evidence. However, FCA's implementation of the 
risk-based capital test relies on the magnitude, as well as the 
direction, of this relationship. It is still the case that using a land 
price decline variable that is defined, in part, by the event that the 
regression seeks to predict, will produce a biased estimate of the 
magnitude of the effect of land price changes on credit risk.

Credit Risk Not Captured:

10. In commenting on the draft report discussion of the three types of 
instruments that are not subject to credit risk in the risk-based 
capital model, FCA stated that the current risk exposures on these 
instruments were immaterial. We recognize that Agvantage bonds are 
backed by both mortgage collateral and the general obligation of the 
issuing institutions. Issuing institutions are likely to be stressed at 
a time of falling farmland values as contemplated by the RBC stress 
test. While we agree that multiple layers of backing for these bonds is 
likely to result in a small credit risk, they are still at some risk of 
loss in a stressed time period. The Federal Home Loan Bank System uses 
a similar product (Advances) with even more layers of backing (mortgage 
pools, general obligations of the originating institutions, and the so-
called superlien, giving Home Loan Banks first priority on the assets 
of originating depository institutions), yet the Federal Housing 
Finance Board assigns a small, but nonzero credit risk charge to these 
assets.

We agree that the credit risk stemming from counterparty risk on swap 
transactions, and the credit risk on many of the assets in Farmer Mac's 
liquidity portfolio, is likely to be small. However, we believe that 
credit risk can be easily accounted for, and the text of our report 
notes that it is accounted for in the risk-based capital models of 
other regulators, such as OFHEO and the FHFB. Doing so would increase 
the accuracy of FCA's risk-based capital calculation for Farmer Mac, 
and would provide an incentive for Farmer Mac to do business with 
higher rated counterparties and to hold lower risk assets, if the risk 
based capital constraint becomes binding.

[End of section]

Appendix IX: GAO Contacts and Staff Acknowledgments:

GAO Contacts:

Davi D'Agostino, (202) 512-8678 Jeanette Franzel, (202) 512-9471:

Acknowledgments:

In addition to those individuals named above, Rachel DeMarcus, Debra 
Johnson, Austin Kelly, Paul Kinney, Bettye Massenburg, Kimberley 
McGatlin, Nicholas Satriano, John Treanor, and Karen Tremba made key 
contributions to this report.

[End of section]

Glossary of Terms:

Amortization:

The process of making regular, periodic decreases in the book or 
carrying value of an asset.

Basis points:

A basis point is equal to one hundredth of a percent. It is used to 
measure changes in or differences between yields or interest.

Capital:

For financial purposes, capital is generally defined as the long-term 
funding for a firm that cushions the firm against unexpected losses.

Credit risk:

The possibility of financial loss resulting from default by borrowers 
on farming assets that have lost value or other parties' failing to 
meet their obligations.

Discount notes:

Discount notes are unsecured general corporate obligations that are 
issued at a discount but mature at face value. Their maturities range 
from overnight to one year.

Duration:

A measure of the average timing of cash flows from an asset or a 
liability. It is computed by summing the present values of all future 
cash flows after multiplying each by the time until receipt, and then 
dividing that product by the sum of the present value of the future 
cash flows without weighting them for the time of receipt.

Interest rate risk:

Interest rate risk is the potential that changes in prevailing interest 
rates will adversely affect assets, liabilities, capital, income or 
expenses at different times in different amounts.

Interest rate swap:

A financial instrument representing a transaction in which two parties 
agree to swap or exchange net cash flows, on agreed-upon dates, for an 
agreed-upon period of time, for interest on an agree-upon principal 
amount. The agreed upon principal amount, called the notional amount, 
is never exchanged. Only the net interest cash flows are remitted. In 
the simplest form of interest rate swap, one party agrees to swap 
fixed-rate loan payments with the floating-rate payments of another 
party.

Liquidity:

Both the capacity and the perceived capacity to meet all obligations 
whenever due, without a material increase in cost, and to take 
advantage of business opportunities important to the future of the 
enterprise. The capacity and the perceived ability to meet known near-
term and long-term funding commitments whole supporting selective 
business expansion.

Liquidity contingency risk:

The risk that future events may require a materially larger amount of 
liquidity than the financial institution currently requires. It is of 
the three primary components of liquidity risk along with mismatch 
liquidity risk and market liquidity risk.

Medium term notes (MTN):

Medium term notes are debt securities that may be issued with floating 
or fixed interest rates with maturities ranging from nine months to 
thirty years or longer. An advantage of MTNs over corporate bonds is 
that they tend to be more flexible in terms of maturities and interest 
rates.

Operations risk:

The risk that an entity may be exposed to financial loss from 
inadequate systems, management failure, faulty controls, or human 
error.

Prepayment risk:

The risk that prepayments will speed or slow and therefore change the 
yield and/or life of the security.

Return on average assets:

Return on average assets is net income for the year divided by the 
average total assets of the year.

Return on equity:

Return on average common stockholder equity is the net income for the 
year less preferred stockholder dividends divided by the average common 
stockholder equity for the year and demonstrates how well the company 
is performing for its common stock shareholders.

Yield maintenance:

A prepayment premium that allows investors to attain the same yield as 
if the borrower made all scheduled mortgage payments until maturity.

(250095):

FOOTNOTES

[1] As used in this report, a GSE is a federally chartered, privately 
owned corporation established by Congress to provide a continuing 
source of credit nationwide to a specific economic sector.

[2] U.S. General Accounting Office, Government Sponsored Enterprises: 
Federal Oversight Needed for Nonmortgage Investments, GAO/GGD-98-48 
(Washington, D.C.: Mar. 11, 1998). U.S. General Accounting Office, 
Farmer Mac: Revised Charter Enhances Secondary Market Activity, but 
Growth Depends on Various Factors, GAO/GGD-99-85 (Washington, D.C.: May 
21, 1999). In this report, we reviewed the progress that Farmer Mac had 
made in achieving its statutory mission and examined its future 
viability. 

[3] See Background section for definitions.

[4] Pub. L. No. 100-233.

[5] Id.

[6] Pub. L. No. 104-105.

[7] SPI is the right to receive a portion of the principal and interest 
payments on a loan or pool of loans, but only after other investors in 
the Farmer Mac-guaranteed securities backed by these pools have 
received all payments due to them. Originators could have retained SPIs 
in the loans they sold to Farmer Mac or they could have sold SPIs to a 
pooler.

[8] Zions Bank, a national bank chartered by the Office of the 
Comptroller of Currency, is referred to as a related party of Farmer 
Mac's because it is the largest holder of Farmer Mac's Class A voting 
Common Stock and a major holder of Class C nonvoting Common Stock. In 
addition, Zions Bank's Executive Vice President is on Farmer Mac's 
Board of Directors, and Zions Bank sells loans to Farmer Mac and serves 
as a Central Servicer of loans for Farmer Mac. Zions Bank also acted as 
an underwriter, agent, and dealer regarding Farmer Mac's discount and 
medium-term notes.

[9] These off-balance sheet obligations were disclosed in Farmer Mac's 
SEC filings.

[10] Per Farmer Mac's 2002 Annual Report, impaired assets are loans 
that are 90 days or more past due, in foreclosure, loans performing in 
bankruptcy, either under their original loans terms or a court-approved 
bankruptcy plan, and real estate owned, which is real estate acquired 
through foreclosure.

[11] A basis point is equal to one hundredth of a percent. 

[12] Loans written off are losses on the outstanding balance of the 
loan, any interest payments previously accrued or advanced, and 
expected collateral liquidation costs. The post-1996 Act loans and 
guarantees are post-1996 Act loans held and loans underlying the 
guaranteed securities and standby agreements, which represent the 
credit risk on loans and guarantees assumed by Farmer Mac.

[13] Statement of Financial Accounting Standard 5: Accounting for 
Contingencies, issued March 1975. Statement of Financial Accounting 
Standard 114: Accounting by Creditors for Impairment of a Loan, an 
amendment of FASB Statements No. 5 and 15, issued May 1993.

[14] Liquid assets are primarily cash and cash equivalents on the 
balance sheet. Farmer Mac refers to these as the Liquidity Investment 
Portfolio.

[15] Discount notes are unsecured general corporate obligations that 
are issued at a discount but mature at face value. Their maturities 
range from overnight to 1 year. 

[16] Medium-term notes (MTN) are debt securities that may be issued 
with floating or fixed interest rates with maturities ranging from 9 
months to 30 years or longer. An advantage of MTNs over corporate bonds 
is that they tend to be more flexible in terms of maturities and 
interest rates. 

[17] There is no statutory or regulatory requirement for Farmer Mac to 
obtain a credit rating.

[18] A rating agency, such as Moody's or Standard and Poors, provides 
its opinion on the creditworthiness of an entity and the financial 
obligations issued by an entity, using a credit rating system. The 
ratings may range from AAA (high quality) to D (in default). Bonds 
rated "BBB" or higher are widely considered "investment grade." This 
means the quality of the securities is high enough for a prudent 
investor to purchase them. 

[19] Agency papers are short-term debt securities that are 
predominantly issued by GSE and federal agencies.

[20] For purposes of this report, we define liquidity as both the 
capacity and the perceived capacity to meet obligations as they come 
due without a material increase in the cost to the institution.

[21] We did not include Freddie Mac's liquidity reserve since at the 
time of this report, Freddie Mac was in the process of restating its 
financial position.

[22] A repurchase agreement is a form of secured, short-term borrowing 
in which a security is sold with a simultaneous agreement to buy it 
back from the purchaser at a future date.

[23] Funding liquidity risk is the potential that an institution would 
be unable to meet its obligations as they come due because of an 
inability to liquidate a sufficient quantity of assets or to obtain a 
sufficient quantity of new liabilities.

[24] See Interest rate risk section and appendix III for further 
discussion of interest rate swaps.

[25] Farmer Mac noted that between 1996 and 2000, $553 million of AMBS 
were sold. In addition, Farmer Mac noted that traders advised 
management that they believed Farmer Mac could re-enter the AMBS market 
and achieve pricing relative to comparable Fannie Mae securities at 
least as favorable as that achieved in 1996-1998.

[26] Yield maintenance is a penalty paid by borrowers to lenders when a 
loan is paid off before its scheduled maturity. It is calculated so 
that the lender is at least made whole in a time of falling interest 
rates.

[27] As discussed later in this report, Farmer Mac has chosen to retain 
the majority of the loans it has purchased and securitized as AMBS.

[28] The MVE is the difference between the present values of cash flows 
associated with assets, minus the present value of cash flows 
associated with liabilities and obligations. MVE represents the current 
economic or financial value of the firm as opposed to the accounting-
based value represented on the balance sheet.

[29] Duration gap is the difference between the average timing of the 
cash flows of the assets and the average timing of the cash flows of 
the liabilities. For a further description of duration, see the 
Glossary.

[30] Farmer Mac is required to comply with the higher of the minimum 
capital requirement or the risk-based capital requirement.

[31] 12 C.F.R.§650.22(a).

[32] The minimum capital requirement is an amount of core capital equal 
to the sum of 2.75 percent of Farmer Mac's aggregate on-balance-sheet 
assets, as calculated for regulatory purposes, plus 0.75 percent of the 
aggregate off-balance-sheet obligations of Farmer Mac.

[33] Both Treasury and Farmer Mac are in agreement that the authority 
of Treasury to purchase obligations to enable Farmer Mac to fulfill its 
guarantee obligations does not extend to the standby agreements because 
they do not involve Farmer Mac's guarantee liabilities.

[34] 12 U.S.C.2279aa-13.

[35] Letter dated April 13, 1997, from then-Under Secretary for 
Domestic Finance, John D. Hawke, Jr., to Marsha P. Martin, then-
Chairman of the Farm Credit Administration.

[36] A spread is the difference between two prices or two rates.

[37] The single AMBS transaction made by Farmer Mac that year was a 
$47.7 million sale to a related party, which represented only 2 percent 
of Farmer Mac's loan purchase, guarantee, and commitment activity for 
the year. 

[38] General farm debt includes more than agricultural real estate 
mortgages; however, it is a proxy for the relative proportion of farm 
borrowing in a region (USDA, Economic Research Service).

[39] U.S. General Accounting Office, Government Sponsored Enterprises: 
Federal Oversight Needed for Nonmortgage Investments, GAO/GGD-98-48 
(Washington, D.C.: Mar. 11, 1998), p.17.

[40] The authority to issue Class C common, nonvoting stock was added 
as an amendment to the proposed legislation that became the 1987 Act by 
Senator Leahy, who explained the purpose of the amendment as follows: 
"….amendments establish that while the initially issued stock is voting 
and fairly distributed between the Farm Credit System and non-Farm 
Credit System participants, the corporation has the authority to issue 
additional nonvoting common and preferred stock if it is determined by 
the mortgage corporation that the corporation should raise additional 
capital."

[41] Statement of Congressman Bereuter, H11869-01, Congressional 
Record. In addition, the conference report cites testimony given in 
hearings held prior to enactment of the bill that indicate that FCS 
spokepersons argued that the secondary market mechanism should operate 
as an arm of the FCS and private lenders believed that the FCS should 
have a much more limited involvement in the secondary market and that 
additional control over a large secondary market operation would give 
the FCS an unfair competitive advantage.

[42] Listed companies are required to disclose these determinations. 
The proposed standards also contain descriptions of relationships in 
which a director is presumed not to be independent until 5 years after 
the relationship ceases. These relationships are as follows: (1) The 
director or an immediate family member receives more than $100,000 per 
year in direct compensation from the listed company, other than 
director and committee fees and pensions. (2) The director or an 
immediate family member is affiliated with or employed in a 
professional capacity by a present or former internal or external 
auditor of the company. (3) The director or an immediate family member 
is employed as an executive officer of another company where any of the 
listed company's present executives serves on that company's 
compensation committee. (4) The director or immediate family member is 
an executive officer of another company that accounts for (a) at least 
2% or $1 million, whichever is greater, or the listed company's 
consolidated gross revenues, or (b) for which the listed company 
accounts for at least 2% or $1 million, whichever is greater of the 
other company's gross revenues. 

[43] CAMELS refer to six components of a financial institution's 
performance - capital adequacy, asset quality, management, earnings, 
liquidity, and sensitivity to market risk. 

[44] 12 U.S.C. §2279bb-1.

[45] Stress tests are computer simulations that demonstrate how a 
firm's financial holdings and obligations would perform under adverse 
economic conditions. Generally, stress tests simulate an economic 
environment considered to be a worst-case scenario for the type of 
business a firm runs.

[46] With the AgVantage program Farmer Mac, in effect, purchases bonds 
from agricultural lenders with the lenders' using agricultural 
mortgages as collateral.

[47] Pub. L. No. 102-237 states that the risk-based capital test must 
determine the amount of regulatory capital for Farmer Mac that is 
sufficient for Farmer Mac to maintain positive capital during a 10-year 
period.

[48] U.S. General Accounting Office, OFHEO's Risk-based Capital Stress 
Test: Incorporating New Business Is Not Advisable, GAO-02-521 
(Washington, D.C.: June 28, 2002).

[49] 12 U.S.C. § 4611 required the Congressional Budget Office and us 
to study whether OFEHO should incorporate new business assumptions into 
the stress test used to establish risk-based capital requirements for 
the housing GSEs. The Director of OFHEO may, after consideration of 
these studies, assume that the GSE conducts additional new business 
during the stress period.

[50] 12 C.F.R. 1750.

[51] Under the Farm Credit Act of 1971, FCA has general authority to 
examine and supervise the safety and sound performance of the powers, 
functions, and duties of Farmer Mac and its affiliates (12 
U.S.C.§2279aa-11(a)).

[52] U.S. General Accounting Office, Options for Federal Oversight of 
GSEs, GAO/GGD-91-90 (Washington D.C.: May 22, 1991). Other criteria for 
effective oversight are independence and objectivity, prominence, 
economy and efficiency, and consistency.

[53] 12 U.S.C. §2279aa-(11)(a)(3)(C).

[54] Federal Home Loan Mortgage Corporation (Freddie Mac) was not 
included in the analysis because it was in the process of restating its 
2000 to 2002 financial statements. Freddie Mac was not expected to 
complete the restatement until November 30, 2003.

[55] For purposes of this study, Commercial Agriculture Banks reflect 
the combined performance of banks "that have a proportion of farm loans 
to total loans that is greater than the unweighted average at all 
banks" and were obtained from the fourth quarter 2002 and first quarter 
2003 Board of Governors of the Federal Reserve System Agricultural 
Finance Databook.

[56] Post-1996 Act loans and guarantees are Farmer Mac I loans and 
guarantees that Farmer Mac acquired or guaranteed after the 1996 Farm 
Act was passed.

[57] A basis point is equal to one hundredth of a percent. It is used 
to measure changes in or differences between yields or interest rates.

[58] Impaired loans are analyzed on a loan-by-loan basis to measure 
impairment on the current value of the collateral for each loan 
relative to the total amount due from the borrower. Farmer Mac 
specifically determines an allowance for the loan for the difference 
between the recorded amount due and its current collateral value, less 
estimated costs to liquidate the collateral.

[59] Loans written off are losses on the outstanding balance of the 
loan, any interest payments previously accrued or advanced, and 
expected collateral liquidation costs. The post-1996 Act loans and 
guarantees are post-1996 Act loans held and loans underlying the 
guaranteed securities and standby agreement, which represent the credit 
risk on loans and guarantees assumed by Farmer Mac.

[60] Interest rate risk is the potential that changes in prevailing 
interest rates will adversely affect assets, liabilities, capital, 
income, or expenses at different times in different amounts.

[61] London Interbank Offered Rate (LIBOR) is the rate that the most 
creditworthy international banks dealing in Eurodollars charge each 
other for large loans. The LIBOR rate is usually the base for other 
large Eurodollar loans to less creditworthy corporate and government 
borrowers.

[62] Yield maintenance is designed to compensate lenders for loss in 
market value when loans are paid off early in falling rate 
environments. The yield-maintenance penalty formula tends to slightly 
overcompensate lenders for early repayment because the formula does not 
consider the effect of amortization, and the formula uses the gross 
spread between the interest rate on the mortgage (net of servicing 
fees) and a Treasury security of comparable maturity, although some of 
that spread represents the higher cost of agency debt, and not the net 
interest margin on the loan. Because yield maintenance is not collected 
for the last six months of a loan's life, it may less than fully 
compensate the lender when a loan is paid off near its maturity date.

[63] Numerous studies of the performance of commercial mortgage 
behavior incorporate this effect. It is generally done by using the 
market, as opposed to the book value of the mortgage when calculating 
loan-to-value ratios. The market value is calculated by taking the 
stream of payments of the mortgage, discounted at the currently 
prevailing interest rate. The market value of a mortgage rises when 
interest rates fall, in line with the yield-maintenance payment. Market 
value and yield maintenance are two different approaches to calculating 
the same concept, namely, the value of the mortgage to investors. Some 
examples of papers that use market value of the mortgage as a predictor 
of loan default include Vandell, Barnes, Hartzell, Kraft, and Wendt 
(1993) "Commercial Mortgage Defaults: Proportional Hazards Estimation 
Using Individual Loan Histories," American Real Estate and Urban 
Economics Association Journal, V 21 Number 4, pp. 451-480, or Huang, 
and Ondrich (2002) "Pay, Stay, or Walk Away: A Hazard Rate Analysis of 
FHA Multifamily Mortgages," Journal of Housing Research, V 13 Number 1, 
pp. 85-117. 

[64] QRM is a commercial software used to manage IRR.

[65] Yield curve is a graph showing the relationship between the yield 
on bonds of the same credit quality but different maturities.

[66] Prepayment speed is the rate at which mortgages pay off before 
their scheduled maturity.

[67] Farmer Mac uses the same data for its credit risk models and 
therefore faces the same limitations.

[68] This problem is known as "right censoring" in statistical 
analysis. SeeYamaguchi, Kazuo, 1991, Event History Analysis (Sage 
Publications, Inc., Newbury Park, CA) pp. 3-9.

[69] 12 CFR Part 650.

[70] Problems relating to extrapolation and the form of the 
relationship are discussed in Snedecor, George, and Cochran, William, 
1967, Statistical Methods, Sixth Edition (Iowa State University Press, 
Ames, IA) p. 144 and p. 456.

[71] Yamaguchi calls this nonindependent censoring. See Yamaguchi, op. 
cit., p. 6 and pp.169-172.

[72] General Accounting Office, Preliminary Analysis of the Financial 
Condition of the Farm Credit System, GAO/GGD-86-13-BR (Washington, 
D.C.: Oct. 4, 1985).

[73] General Accounting Office, Mortgage Financing: FHA Has Achieved 
Its Home Mortgage Capital Reserve Target, GAO/RCED-96-50 (Washington, 
D.C.: April 12, 1996) and Office of Federal Housing Enterprise 
Oversight, 1999, Risk-Based Capital Regulation: Second Notice of 
Proposed Rulemaking, Federal Register: (Volume 64, Number 70) (April 
13, 1999).

[74] Price-Waterhouse, 1997, An Actuarial Review for Fiscal Year 1996 
of the Federal Housing Administration's Mutual Mortage Insurance Fund, 
(Washington, D.C., Feb. 14, 1997). 

[75] Data after 1992 were not readily useable, as the Texas Bank 
changed computer systems and post-1992 data could not be readily linked 
to earlier loans. FCA noted they are now studying the data to determine 
if it is possible to link post-1992 data to earlier loans.

[76] Different models use different measures of credit risk, such as a 
loan terminating in a claim (such as our study of VA Subsidy Rates, 
Homeownership: Appropriations Made to Finance VA's Housing Program May 
be Overestimated (GAO-RCED-93-173)) or delinquency, Calem, P. and 
Wachter, S. 1991, Community Reinvestment and Credit Risk: Evidence from 
an Affordable-Home-Loan Program, Real Estate Economics, V. 27 #1, 
pp.105-134. The term credit event is used as a general description of 
these various definitions.

[77] Such models are known as hazard models. Yamaguchi, op. cit., p. 9.

[78] The land price change variable is then modified by a "dampening 
factor" before entering the regression, but the value of the 
transformed variable is still determined, in part, by whether and when 
the loan enters foreclosure.

[79] Working, E "What do Statistical Demand Curves Show?", 1927, 
Quarterly Journal of Economics V 41 #1.

[80] A similar example is cited in Yamaguchi, op. cit., pp. 26-27. 
Yamaguchi concludes that independent variables, which are determined by 
life course characteristics, can only be used if their value is 
determined prior to the observation entering the period in which they 
are subject to the risk of experiencing the event to be modeled. 

[81] See Kennedy, P, 1987, A Guide to Econometrics, 2nd Edition, (MIT 
Press, Cambridge, MA), p. 164.

GAO's Mission:

The General Accounting Office, the investigative arm of Congress, 
exists to support Congress in meeting its constitutional 
responsibilities and to help improve the performance and accountability 
of the federal government for the American people. GAO examines the use 
of public funds; evaluates federal programs and policies; and provides 
analyses, recommendations, and other assistance to help Congress make 
informed oversight, policy, and funding decisions. GAO's commitment to 
good government is reflected in its core values of accountability, 
integrity, and reliability.

Obtaining Copies of GAO Reports and Testimony:

The fastest and easiest way to obtain copies of GAO documents at no 
cost is through the Internet. GAO's Web site ( www.gao.gov ) contains 
abstracts and full-text files of current reports and testimony and an 
expanding archive of older products. The Web site features a search 
engine to help you locate documents using key words and phrases. You 
can print these documents in their entirety, including charts and other 
graphics.

Each day, GAO issues a list of newly released reports, testimony, and 
correspondence. GAO posts this list, known as "Today's Reports," on its 
Web site daily. The list contains links to the full-text document 
files. To have GAO e-mail this list to you every afternoon, go to 
www.gao.gov and select "Subscribe to e-mail alerts" under the "Order 
GAO Products" heading.

Order by Mail or Phone:

The first copy of each printed report is free. Additional copies are $2 
each. A check or money order should be made out to the Superintendent 
of Documents. GAO also accepts VISA and Mastercard. Orders for 100 or 
more copies mailed to a single address are discounted 25 percent. 
Orders should be sent to:

U.S. General Accounting Office

441 G Street NW,

Room LM Washington,

D.C. 20548:

To order by Phone: 	

	Voice: (202) 512-6000:

	TDD: (202) 512-2537:

	Fax: (202) 512-6061:

To Report Fraud, Waste, and Abuse in Federal Programs:

Contact:

Web site: www.gao.gov/fraudnet/fraudnet.htm E-mail: fraudnet@gao.gov

Automated answering system: (800) 424-5454 or (202) 512-7470:

Public Affairs:

Jeff Nelligan, managing director, NelliganJ@gao.gov (202) 512-4800 U.S.

General Accounting Office, 441 G Street NW, Room 7149 Washington, D.C.

20548: