Farmer Mac: Some Progress Made, but Greater Attention to Risk	 
Management, Mission, and Corporate Governance Is Needed 	 
(16-OCT-03, GAO-04-116).					 
                                                                 
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).					 
-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-04-116 					        
    ACCNO:   A08727						        
  TITLE:     Farmer Mac: Some Progress Made, but Greater Attention to 
Risk Management, Mission, and Corporate Governance Is Needed	 
     DATE:   10/16/2003 
  SUBJECT:   Agricultural assistance				 
	     Agricultural programs				 
	     Farm credit					 
	     Financial analysis 				 
	     Program evaluation 				 
	     Risk management					 
	     Strategic planning 				 
	     Government sponsored enterprises			 
	     Agency missions					 
	     Investments					 
	     Loans						 
	     Securities 					 

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Product.                                                 **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO-04-116

United States General Accounting Office

GAO

                       Report to Congressional Requesters

October 2003

FARMER MAC

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

                                       a

GAO-04-116

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

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).

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.

October 2003

FARMER MAC

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

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.

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

1 2 7

12

30

38

47 54 57 59 59

Appendixes Appendix I: Appendix II: Appendix III:

Appendix IV: Appendix V:

Appendix VI:

Appendix VII:

Appendix VIII:

Objectives, Scope, and Methodology 63

Farmer Mac's Programs and Products 67

Financial Trends and Comparisons with Other Entities 69 Revenue Has
Increased, but Some Financial Performance Indicators Lag Comparative
Entities 72

Farmer Mac's Underwriting Standards 78

Interest Rate Risk 81 Asset-Liability Management 81 Prepayment Model 82
Farmer Mac's IRR Measurement Process 84

Farm Credit Administration Credit Risk Model 86
Data Limitations 87
Model Limitations 91

Comments from the Federal Agricultural Mortgage
Corporation 95
GAO Comments 100

Comments from the Farm Credit Administration 102

                                    Contents

GAO Comments 107

Appendix IX:	GAO Contacts and Staff Acknowledgments 112 GAO Contacts 112
Acknowledgments 112

  Glossary of Terms

Tables           Table 1: Loans Covered by Standby Agreements           13 
           Table 2: Comparison of Total Minimum Capital Levels Per $100 of 
                                        Loans                              36 
             Table 3: Annual Compensation and Options Granted for CEO's of 
                             Farmer Mac and Housing GSEs                   46 

Figures Figure 1:

Figure 2:

Figure 3:

Figure 4:

Figure 5:

Figure 6: Figure 7:

Figure 8: Figure 9:

Percentage of Outstanding Balance of Loans, AMBS and
Standby Agreements, as of December 31, 2002
Long-term Interest Rates on Loans Secured by
Agricultural Real Estate
Securitization Status of Farmer Mac I Portfolio, as of
December 31, 2002
Farmer Mac Portfolio Exposure by Loan Origination
Type
Farmer Mac I Geographic Concentration of Exposure by
Region, as of December 31, 2002
Farmer Mac's Impaired Loans from 1997 to 2002
Farmer Mac's Nonperforming to Total Loans Compared
to Other Entities, as of December 31, 2002
Income by Program Assets
Farmer Mac's ROA Compared to Other Entities

                                       6

                                       32

                                       33

                                       35

                                     37 70

71 72 74 75

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

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

Contents

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 U.S. Department of Agriculture

Contents

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. However, 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.

A

United States General Accounting Office Washington, D.C. 20548

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)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.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

1As 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.

2U.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.

measured the risks it faces-credit, liquidity, and interest rate risk.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).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.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

3See Background section for definitions.
4Pub. L. No. 100-233.
5Id.

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.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. 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,

6Pub. L. No. 104-105.

7SPI 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.

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. 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

8Zions 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.

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

                             1% Pre 1996 Act loans

                                 Farmer Mac II

                                 Loans and AMBS

                               Standby agreements

Source: GAO analysis of data from Farmer Mac 2002 SEC 10-K filing.

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:

o 	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.

o 	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").

o 	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.

o 	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

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

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.

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 1998 1999 2000 2001 2002

Loans covered by
standby agreements $0 $0.6 billion $0.9 billion $1.9 billion $2.7 billion

Percentage increase
from previous year N/A N/A 50% 111% 42%

Source: Farmer Mac 2002 SEC 10K filing.

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.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 Farmer Mac has established an allowance for
loan losses and reserve for

Charge-offs May Indicate losses (a loan loss allowance) to cover
estimated, probable loan losses for

Increasing Credit Risk	its current portfolio of loans, commitments, and
guarantees. The loan loss allowance has increased substantially from $1.6
million at December 31,

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

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, 1997.10 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
points11 of post-1996 Act loans and guarantees, which was a significant
increase over the $2.2 million, or 6 basis points, written off in 2001.12

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

10Per 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.

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

12Loans 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.

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 Farmer Mac uses underwriting standards and
processes for monitoring the

and Servicing Procedures Were Clear, but Exceptions Were Not Consistently
Well Documented

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 As part of the financial monitoring and
reporting process, Farmer Mac's

Mac's Monitoring and management is responsible for assessing the current
level of risk

Assessment of Changes in Credit associated with individual loans and loan
portfolios that have been

Risk	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."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 Reviews of Farmer Mac conducted in 2002
concluded that Farmer Mac had

Estimation Model Was Weak	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

13Statement 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.

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.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.

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

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. 15 However, these events temporarily affected the interest rates
on medium-term notes.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.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.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.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,

15Discount 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.

16Medium-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.

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

18A 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.

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

who purchase debt securities, due to internal policies that prohibit
purchasing unrated financial instruments.

Farmer Mac Has Maintained Farmer Mac's liquidity investment portfolio was
a secondary source of

Close to 30 Days of Liquidity	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.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.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.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

20For 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.

21We 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.

22A 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.

funds could potentially be impaired by external or internal events. 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.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 In addition to meeting liquidity demands from
expected obligations,

Agreements Can Potentially Farmer Mac may face unexpected demands on
funding liquidity should

Create Unexpected Demands for lenders that participate in the standby
agreements exercise their contracts.

Additional Funding Liquidity	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.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

23Funding 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.

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

25Farmer 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.

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.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.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

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

26Yield 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.

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

28The 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.

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

o 	duration gap calculation, which measures the interest rate mismatch
between Farmer Mac' assets and liabilities. 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

29Duration 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.

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.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.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

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

3112 C.F.R.S:650.22(a).

capital requirement,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.

32The 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.

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.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.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.

33Both 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.

3412 U.S.C.2279aa-13.

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),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.

35Letter 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.

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

                                       10

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.
36 According to USDA, holding AMBS has typically been more profitable but

                                       8

                                       6

                                       4

                                       2

                                       0

Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4
2000 2001 2002
Year and quarter

Commercial banks Farm Credit System

Farmer Mac I

                             10-year Treasury bond

Sources: GAO analysis of Farmer Mac data and USDA's Agricultural Income &
                 Finance Outlook, AIS80, March 11, 2003, p.63.

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

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

also more risky than selling AMBS to investors. During 2002, Farmer Mac
did not make any sales of AMBS to unrelated parties.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

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

37The 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.

Securitized and sold

                            Securitized and retained

Unsecuritized loans

Standby agreements

                    Source: GAO analysis of Farmer Mac data.

Farmer Mac's Business Activities Are Largely Concentrated

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 Percent
100

80

60

40

20

0 1996 1997 1998 1999 2000 2001 2002 Year

Commercial banks

Farm credit system

Mortgage brokers

Insurance companies

Source: GAO analysis of Farmer Mac data.

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 institution    Farmer Mac     Total capital 
                    FCS loan              $7.00           N/A           $7.00 
           Standby agreement              $1.40         $0.75           $2.15 
                   Loan sale              $0.00        $2.75a           $2.75 
               Swap for AMBS              $1.40         $0.75           $2.15 

Source: Farm Credit System.

aThis assumes that Farmer Mac retains the loan or securitizes the loan and
holds the AMBS on its balance sheet.

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.38 Greater geographic
diversification of Farmer Mac loans would lower risks of concentration and
mitigate risks associated with the lower capital requirements.

38General 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).

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

                              Mid-south Northeast

                                   Southeast

                                   Mid-north

                                   Northwest

                                   Southwest

                    Source: GAO analysis of Farmer Mac data.

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.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

39U.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.

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.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.

40The 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."

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.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

41Statement 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.

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.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.

42Listed 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.

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

                              Dollars in millions

                                                                     Value of
                                                                  unexercised
                                                                 in-the-money
                                              Annual Options grant options at
                                           compensation date present year-end
                                        Annual salary bonus value exercisable

President & CEO, Farmer Mac

Loan & Guarantee Portfolio: $5.5 billion

Employees: 33 $447,480 $344,195 $1,150,783 $9,511,068

CEO and Chairman, Fannie Mae

Loan & Guarantee Portfolio: $1.8 trillion

Employees: 4,700 992,250 3,300,000 6,680,395 1,441,600

CEO and Chairman Freddie Mac (A)

Loan & Guarantee Portfolio: $1.1 trillion

Employees: 3,900 1,132,500 2,123,438 3,899,741 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.

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.43
In addition, FCA was closely monitoring Farmer Mac's corrective actions to
address identified weaknesses.

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

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

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

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.

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.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.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

4412 U.S.C. S:2279bb-1.

45Stress 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.

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

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 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. 46 However, unlike FHFB, FCA does not include AgVantage bonds
in its risk-based capital calculation.

Market Risk and Income Models FCA uses results from Farmer Mac's interest
rate risk model to measure the May Understate Estimated level of market
risk to which Farmer Mac is exposed and determine Levels of Required
Risk-based corresponding levels of risk-based capital. As discussed
previously, the

Capital

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

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

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.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.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.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.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.

47Pub. 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.

48U.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).

4912 U.S.C. S: 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.

5012 C.F.R. 1750.

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

Overseeing Both FCS Banks and Farmer Mac Is a Regulatory Challenge

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.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.

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.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

51Under 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.S:2279aa-11(a)).

52U.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.

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.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

5312 U.S.C. S:2279aa-(11)(a)(3)(C).

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:

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

o 	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;

o 	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

o 	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.

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

                                       o

                                       o

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.

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

o 	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.

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

o 	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

o 	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;

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

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

o 	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:

o 	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;

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

o 	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;

o 	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

o 	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:

o 	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;

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

o 	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 	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

Our Evaluation	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
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.

Davi M. D'Agostino

Director, Financial Markets and Community Investment

Jeannette M. Franzel

Director, Financial Management and Assurance

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

Appendix I
Objectives, Scope, and Methodology

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.1
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. 2 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,

1Federal 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.

2For 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.

Appendix I
Objectives, Scope, and Methodology

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

Appendix I
Objectives, Scope, and Methodology

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.

Appendix II

                       Farmer Mac's Programs and Products

Program Program description Product feature

Farmer Mac I a

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

Full-Time Farm Designed for borrowers who live on

Program	agricultural properties and derive a significant portion of their
income from farm employment.

Types of agricultural loans offered include:

o  15-year fixed rate, 15-year maturity with 15- or 25-year amortization
and partial open prepayment available (annual, semiannual, or monthly
payments);

o  10-year fixed rate, 10-year maturity fully amortizing (semiannual or
monthly payments);

o  5-year reset loan with a 5-year term (renewable twice); 5-, 10-, 15-,
or 25-year amortization (annual, semiannual, or monthly payments);

o  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

o  facility loans,10- or 15-year fixed rate maturity, and fully amortized.

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

AgVantage Program	Farmer Mac purchases and guarantees timely AgVantage
bonds may range in maturity from short-term to payment of principle and
interest on 15 years and have low fixed or variable rates of interest.
mortgage-backed bonds.

Swap Program	Farmer Mac acquires eligible loans from Security terms,
rates, etc., are negotiated with the seller on sellers in exchange for
Farmer Mac the basis of the characteristics of the loan. Guaranteed
Securities backed by those loans.

Long-Term Standby Farmer Mac commits to purchase loans from Terms are
negotiated with institution based on the

Purchase Commitments	a segregated pool of loans on one or more
characteristics of the underlying loan. undetermined future dates.

                                  Appendix II
                       Farmer Mac's Programs and Products

                         (Continued From Previous Page)

                  Program Program description Product feature

Farmer Mac II b

Cash Window Program	Lenders receive cash by selling 100 percent of  o 
7-year fixed rate and 15-year fixed rate based on full the guaranteed
portion of USDA loans directly amortization; to Farmer Mac.

o  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

o  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.

Swap Program	Lenders receive Farmer Mac-guaranteed Security terms, rates,
etc., are negotiated with the seller on securities in return for the
guaranteed portion the basis of the characteristics of the loan. of USDA
loans.

Sources: Farmer Mac and FCA.

aFarmer 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.

bIn 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.

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
loans1 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 points2 from 1.70 percent at December 31, 2001, to
1.56 percent at December 31, 2002.3 (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.

1Post-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.

2A basis point is equal to one hundredth of a percent. It is used to
measure changes in or differences between yields or interest rates.

3Impaired 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.

Appendix III
Financial Trends and Comparisons with
Other Entities

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

Dollars in millions Percent

80 2.0

70

60 1.5

50

40 1.0

30

20 0.5

10

0 0.0
1997 1998 1999 2000 2001 2002
Year

Impaired loans to total loans Total impaired loans Source: GAO's analysis
of Farmer Mac's data.

Appendix III
Financial Trends and Comparisons with
Other Entities

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

2.0

1.5

1.0

0.5

0.0 Farmer ACA FLCA Mac

Source: GAO's analysis of Farmer Mac and FCA data.

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,4 which was a significant increase over the $2.2 million, or 6
basis points, written off in 2001.

4Loans 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.

                                  Appendix III
                     Financial Trends and Comparisons with
                                 Other Entities

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
Dollars in millions
250

200

150

100

50

0 1997 1998 1999 2000 2001 2002 Year

Commitment and guarantee fees
Loan interest income
Farmer Mac guarantee securities
Investment interest income

                  Source: GAO's analysis of Farmer Mac's data.

Appendix III
Financial Trends and Comparisons with
Other Entities

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.

Appendix III
Financial Trends and Comparisons with
Other Entities

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

2.5

2.0

1.5

1.0

0.5

0.0 1997 1998 1999 2000 2001 2002

Year

ACA

Commercial agriculture banks

Fannie Mae

Farmer Mac

FLCA

Source: GAO's analysis of Farmer Mac, FCA, Federal Reserve, and Fannie Mae
data.

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.

Appendix III
Financial Trends and Comparisons with
Other Entities

Figure 10: Farmer Mac's ROE Compared to Other Entities Percent 40

35

30

25

20

15

10

5 1997 1998 1999 2000 2001 2002

Year

ACA

Commercial agriculture banks

Fannie Mae

Farmer Mac

FLCA

Source: GAO's analysis of Farmer Mac, FCA, Federal Reserve, and Fannie Mae
data.

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

Appendix III
Financial Trends and Comparisons with
Other Entities

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
Percent

20

15

10

5

0 1997 1998 1999 2000 2001 2002

Year

ACA

Farmer Mac

FLCA

Fannie Mae

Source: GAO analysis of Farmer Mac, FCA, and Fannie Mae data.

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:

o 	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

Appendix III
Financial Trends and Comparisons with
Other Entities

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.

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

o 	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.

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.

o 	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.

o 	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.

o 	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.

Appendix IV
Farmer Mac's Underwriting Standards

o 	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.

o 	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.

o 	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.

o 	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.

o 	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

Appendix IV
Farmer Mac's Underwriting Standards

having property that is located in areas that are less than 25 percent
developed.

o 	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.

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).1 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.2 Farmer Mac also has prepayment penalties or
yield-maintenance terms on 57

1Interest 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.

2London 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.

                         Appendix V Interest Rate Risk

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.3 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

3Yield 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.

Appendix V Interest Rate Risk

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,

                         Appendix V Interest Rate Risk

a prime determinant of default in studies of borrower behavior.4 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).5 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.6 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

4Numerous 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.

5QRM is a commercial software used to manage IRR.

6Yield curve is a graph showing the relationship between the yield on
bonds of the same credit quality but different maturities.

Appendix V Interest Rate Risk

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.7

7 Prepayment speed is the rate at which mortgages pay off before their scheduled
                                   maturity.

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.

            Appendix VI Farm Credit Administration Credit Risk Model

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.1 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.2 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

1Farmer Mac uses the same data for its credit risk models and therefore
faces the same limitations.

2This problem is known as "right censoring" in statistical analysis.
SeeYamaguchi, Kazuo, 1991, Event History Analysis (Sage Publications,
Inc., Newbury Park, CA) pp. 3-9.

Appendix VI Farm Credit Administration Credit Risk Model

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.3 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.4

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

312 CFR Part 650.

4Problems 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.

Appendix VI Farm Credit Administration Credit Risk Model

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.5

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.6 However, the bulk of the loans now purchased
by Farmer Mac are either rapidly adjusting adjustable-rate mortgages, tied

5Yamaguchi calls this nonindependent censoring. See Yamaguchi, op. cit.,
p. 6 and pp.169-172.

6General Accounting Office, Preliminary Analysis of the Financial
Condition of the Farm Credit System, GAO/GGD-86-13-BR (Washington, D.C.:
Oct. 4, 1985).

Appendix VI Farm Credit Administration Credit Risk Model

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.7 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

7General 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).

            Appendix VI Farm Credit Administration Credit Risk Model

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.8 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.9 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.

8Price-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).

9Data 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.

Appendix VI Farm Credit Administration Credit Risk Model

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. 10 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.11 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

10Different 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.

11Such models are known as hazard models. Yamaguchi, op. cit., p. 9.

Appendix VI Farm Credit Administration Credit Risk Model

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.12 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.13 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.14 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

12The 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.

13Working, E "What do Statistical Demand Curves Show?", 1927, Quarterly
Journal of Economics V 41 #1.

14A 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.

Appendix VI Farm Credit Administration Credit Risk Model

transform the predictor variables prior to estimating the logistic
regression. Such pre-testing leads to inflated tests of significance.15

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.

15See Kennedy, P, 1987, A Guide to Econometrics, 2nd Edition, (MIT Press,
Cambridge, MA),

p. 164.

Appendix VII

Comments from the Federal Agricultural Mortgage Corporation

Note: GAO comments supplementing those in the report text appear at the
end of this appendix.

Appendix VII
Comments from the Federal Agricultural
Mortgage Corporation

Appendix VII
Comments from the Federal Agricultural
Mortgage Corporation

                                 See comment 1.

                                 See comment 2.

Appendix VII
Comments from the Federal Agricultural
Mortgage Corporation

                                 See comment 3.

Appendix VII
Comments from the Federal Agricultural
Mortgage Corporation

                                  Appendix VII
                     Comments from the Federal Agricultural
                              Mortgage Corporation

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

GAO Comments 1.

2.

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.

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

Appendix VII
Comments from the Federal Agricultural
Mortgage Corporation

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.

Appendix VIII

Comments from the Farm Credit Administration

Note: GAO comments supplementing those in the report text appear at the
end of this appendix.

Appendix VIII
Comments from the Farm Credit
Administration

Appendix VIII
Comments from the Farm Credit
Administration

                             See comments 1 and 2.

See comment 3.

Appendix VIII
Comments from the Farm Credit
Administration

See comment 4.

See comment 5.

See comment 6.

See comments 7 and 8.

Appendix VIII
Comments from the Farm Credit
Administration

                                 See comment 9.

                                See comment 10.

                                 Appendix VIII
                         Comments from the Farm Credit
                                 Administration

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

Appendix VIII
Comments from the Farm Credit
Administration

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

Appendix VIII
Comments from the Farm Credit
Administration

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

Appendix VIII
Comments from the Farm Credit
Administration

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-

Appendix VIII
Comments from the Farm Credit
Administration

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.

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.

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.

                               Glossary of Terms

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-uponperiod of time,for interest on anagree-upon
principalamount. 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.

                               Glossary of Terms

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.

GAO's Mission	The General Accounting Office, the audit, evaluation and
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, 	Contact: Web site: www.gao.gov/fraudnet/fraudnet.htm

Waste, and Abuse in E-mail: [email protected]

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

Public Affairs	Jeff Nelligan, Managing Director, [email protected] (202)
512-4800 U.S. General Accounting Office, 441 G Street NW, Room 7149
Washington, D.C. 20548

                               Presorted Standard
                              Postage & Fees Paid
                                      GAO
                                Permit No. GI00

United States
General Accounting Office
Washington, D.C. 20548-0001

Official Business
Penalty for Private Use $300

Address Service Requested
*** End of document. ***