-------------------------Indexing Terms------------------------- 
REPORTNUM:   GAO-01-343SP						        

TITLE:     Final Report Regarding the Findings of the Study Group On the Feasibility of
Using Alternative Financial Instruments For Determining Lender Yield Under the Federal Family Education Loan Program

DATE:   01/19/2001 
				                                                                         
-----------------------------------------------------------------

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** 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-01-343S
     
Final Report Regarding the Findings of the Study Group On the Feasibility of
Using Alternative Financial Instruments

For Determining Lender Yield Under the Federal Family Education Loan Program

January 19, 2001 Submitted by the General Accounting Office and the
Department of Education

GAO- 01- 343SP

Table of Contents CHAPTER 1: AN OVERVIEW OF MAJOR FEDERAL STUDENT LOAN
PROGRAMS AND PARTICIPANTS 1

Introduction
...........................................................................................................1
Purpose of the Study
............................................................................................1
Overview of Major Federal Student Loan Programs and Their
.............................3 Budget Scoring for Student Loan Programs
.........................................................9 Participants in
the Student Loan Program
..........................................................10 For-Profit and
Not-For Profit Lenders Face Similar Risks that Constrain their Operations
..........................................................................................................19

Basic Operating Strategies For Lenders
.............................................................22 Relative
Importance of Revenues and Expenses
...............................................24

CHAPTER 2: PROPERTIES OF RELEVANT INTEREST RATES AND THEIR DETERMINATION
..............................................................................................26

Treasury Bill
Rates..............................................................................................26
Commercial Paper and London Interbank Offer Rates
.......................................32 Spreads Between Rates and
Volatility of Spreads..............................................38 The
Outlook for Interest Rates and Spreads Among Them
................................43

CHAPTER 3: LIQUIDITY OF ALTERNATIVE INSTRUMENTS FOR SETTING LENDER
YIELDS??????????????????????????..46

Liquidity Is The Ability To Buy Or Sell An Asset Approximately At The
Current Market
Price........................................................................................................46

Liquidity of Various Market
Instruments..............................................................47
Liquidity Looking Forward: Is the Past
Prologue?...............................................49

Table of Contents CHAPTER 4: HOW LENDERS HAVE MANAGED INTEREST RATE RISK??72
Lenders? Yields Usually Move in Tandem with T-Bill Rates
................................75 Lenders? Interest Expenses Are Related to
But Do Not Move in Tandem With Their Market Reference
Rates............................................................................78

Credit Risk and Term to Maturity Affect Borrowing Margins Paid by
Lenders?.79 Basis-Risk, Hedging, and Expected Returns???????????????.83
Asset Backed Securities Are Especially Sensitive to Changes in Interest
Rates on Instruments with Longer Terms to Maturity??????????? .90

Comparing Volatility of Spreads Under Different Funding Options?????..94
CHAPTER 5: SUMMARY AND ANALYSES OF STUDY ISSUES 102 Issue 1: The Historical
Liquidity of the Market for Each Instrument, and a Historical Comparison of
the
Spreads...............................................................104

Issue 2: The Historical Volatility of the Rates and Projections of Future
Volatility
..........................................................................................................................105

Issue 3: Recent Changes in the Liquidity of the Market for Each Such
Instrument, and Projections of Future Liquidity in a Balanced Budget
Environment With Low Interest Rates
...................................................................................................106

Issue 4: The Cost or Savings to Different-Sized Lenders of Basing Lender
Yield on an Alternative Instrument, and the Effect of Such Change on the
Diversity of
Lenders.............................................................................................................108

Issue 5: The Cost or Savings to the Federal Government of Basing Lender
Yield on an Alternative Instrument
.............................................................................113

Issue 6: Any Possible Risks or Benefits to the Student Loan Programs Under
the Higher Education Act of 1965 and to Student
Borrowers..................................115

Table of Contents APPENDIX 1: LEGISLATIVE MANDATE IN THE HIGHER EDUCATION
ACT OF 1998 (P.L. 105-244)???????????????????? 117

APPENDIX 2: GOVERNMENT AND NON-GOVERNMENT STUDY GROUP
MEMBERS????????????????????????? 118

APPENDIX 3: A BASIC CHRONOLOGY OF BORROWER RATES; LENDER YIELDS; LOAN
PROGRAM ACTIVITY; AND CERTAIN QUARTERLY INTEREST
RATES???????????????????????? 120

APPENDIX 4: CBO METHODOLOGY FOR BUDGET SCORING?????..124 APPENDIX 5: DATA
AND CALCULATIONS FOR THE VOLATILITY ANALYSIS IN CHAPTER
2????????????????????????. 127

APPENDIX 6: DERIVATION OF COMMERCIAL PAPER INTEREST RATES (FROM FEDERAL
RESERVE BOARD OF GOVERNORS)???????? 128

APPENDIX 7: THE BRITISH BANKERS ASSOCIATION (BBA) AND THE LIBOR
FIXING????????????????????????????.. 131

APPENDIX 8: INTERVIEWS OF FFELP INDUSTRY
REPRESENTATIVES???????????????????..? 136

APPENDIX 9: SURVEYS OF LENDERS AND SECONDARY MARKET
INSTITUTIONS????????????????????????. 143

APPENDIX 10: TREASURY DEPARTMENT SYNTHETIC SWAP MODEL TO ESTIMATE SWAP
RATES??????????????????? 145

APPENDIX 11: RECOMMENDATION OF THE FFELP INDUSTRY MEMBERS OF THE STUDY GROUP
ON THE FEASIBILITY OF ALTERNATIVE FINANCIAL INSTRUMENTS FOR DETERMINING
LENDER YIELD?????????..151

APPENDIX 12: ADMINISTRATION?S EVALUATION OF BASIS RISK FOR THE GOVERNMENT
UNDER ALTERNATIVE FINANCIAL INSTRUMENTS 154

APPENDIX 13: PRINCIPLES OFFERED BY NON-LENDER, NON- GOVERNMENT MEMBERS, AND
RESPONSE TO PRINCIPLES, OFFERED BY FFLEP COMMUNITY MEMBERS????????????????..
158

Principles Offered by Non-Lender, Non-Government Members
.......................158

Table of Contents Response to Principles, Offered by FFELP Community Members
...................159 APPENDIX 14: FFELP COMMUNITY MEMBER COMMENTS ON THIS
REPORT AND THE IMPLICATIONS OF BUDGET SURPLUSES????. 163

APPENDIX 15: TREASURY STATEMENT REGARDING BUDGET SURPLUSES AND IMPLICATIONS
FOR LIQUIDITY??????????????. 172

GLOSSARY?????????????????????????.. 176

CHAPTER 1: AN OVERVIEW OF MAJOR FEDERAL STUDENT LOAN PROGRAMS AND
PARTICIPANTS

INTRODUCTION

As mandated by section 802 of the 1998 reauthorization of the Higher
Education Act, as amended (HEA), this report presents the results of the
Study Group on the feasibility of using alternative financial instruments
for determining lender yields under the Federal Family Education Loan
Program (FFELP). As required, members of the Study Group were convened by
the Comptroller General and the Secretary of Education. In accordance with
section 802, the other members were the Secretary of the Treasury, the
Director of the Office of Management and Budget, the Director of the
Congressional Budget Office, representatives of entities making FFELP loans,
other entities in the financial services community, other participants in
the student loan programs, and other individuals designated by the
Comptroller General and the Secretary of Education. The mandate is included
in appendix I, and appendix II is a complete list of the non-government
members and government agency members of the Study Group.

Section 802 of the HEA requires the Comptroller General and the Secretary of
Education to submit a final report regarding the findings of the Study
Group. The report begins with general descriptive material about FFELP and,
in order to put FFELP in context, the William D. Ford Federal Direct Loan
Program (DL). The report addresses the six specific issues raised in the
legislative mandate (enumerated in the next section). Where members of the
Study Group disagree with any of these portions of the report, that
disagreement is noted, and/or their separate views are included as
appendices.

Although section 802 did not require the Study Group to make a
recommendation about the use of an alternative financial instrument, the
Study Group did attempt to do so. Unfortunately, it was not successful.
Instead, the report includes different positions presented by Study Group
members. They include: the full text of the Study Group?s FFELP lenders?
proposal regarding an alternative financial instrument for determining
lender yields (appendix XI); the Administration?s position (appendix XII),
FFELP lenders? comments on the report (appendix XIV), Treasury
Undersecretary Gensler?s remarks on debt management at the February 2000
mid-quarter refunding announcement (appendix XV), and a proposed set of
principles from the non-FFELP lender and non-government members of the Study
Group to be used when considering changes in the financial instrument to be
used to determine lender yields (appendix XIII).

FFELP provides a credit guarantee and a set of subsidies to lenders in order
to encourage a diversity of lender participation in the guaranteed loan
program, so that all postsecondary school students have access to moderate
interest rate loans. Federal student loan policy has dictated uniform terms
and conditions on student loans. This results in the federal government
paying subsidies (in-school

Chapter 1 2 interest, special allowances, and/or guarantee payments) on all
loans when some loans may be profitable without payments and others may not
be.

Section 802 mandates that the Study Group analyze 91-day Treasury bills,
commercial paper (CP), and the London Interbank Offer Rate (LIBOR) as
potential bases for setting the index of lender yields. 1 Treasury bill
rates, a credit risk free rate in the short-term sector of financial
markets, have been a benchmark in money markets for a long time. Since the
1980s, with the growing breadth, sophistication, and internationalization of
money markets, some private rates, LIBOR and CP, have become increasingly
important. Both have been proposed as alternatives to the Treasury bill rate
for the determination of lender yields on FFELP loans because they are
considered better able to reflect money market conditions and lenders? cost
of funds.

When section 802 was enacted, both the student rate and lender yield were
indexed to the 91-day Treasury bill for most Stafford loans, the primary
component of FFELP loans. However, at the end of 1999, legislation was
enacted for Stafford loans to change the index on which the lender yields
are based to commercial paper rates for most of the loans. 2 This
legislation did not change the index for student borrower rates. The change
to commercial paper will affect new Stafford FFELP loans made between
January 1, 2000, and June 30, 2003. For loans made after July 1, 2003, the
index for both students and lenders will change to a Treasury rate of
comparable maturity to student loans plus 1%. The formulae for determining
borrower rates and lenders? yields will be discussed in more detail in
subsequent chapters because the components of the formulae are significant
contributors to lenders? risks in the current program. Lender yields in this
report are gross of any discounts or fees absorbed by lenders or guarantee
agencies.

PURPOSE OF THE STUDY

In response to the Section 802 mandate, this report evaluates the 91-day
Treasury bill, 30-day and 90-day commercial paper, and the 90-day LIBOR in
terms of:

The historical liquidity of the market for each, and a historical comparison
of the spread between: (1) the 30-day and 90-day commercial paper rate, and
the 91-day Treasury bill rate; and (2) the spread between the LIBOR and the
91-day Treasury bill rate;

The historical volatility of the rates and projections of future volatility;

Recent changes in the liquidity of the market for each such instrument in a
balanced Federal budget environment and a low-interest rate environment, and
projections of future liquidity assuming the federal budget remains in
balance;

1 Commercial paper is short term unsecured lending by firms with strong
credit ratings. LIBOR measures the rate for loans among internationally
active banks. 2 In 1999, approximately 87 percent of all federal student
loans were Stafford loans.

Chapter 1 3

The cost or savings to lenders with small, medium, and large student loan
portfolios of basing lender yield on either the 30-day or 90-day commercial
paper rate or the LIBOR while continuing to base the borrower rate on the
91-day Treasury bill, and the effect of such change on the diversity of
lenders participating in the program.

The cost or savings to the federal government of basing lender yield on
either the 30-day or 90-day commercial paper rate or LIBOR while continuing
to base the borrower rate on the 91-day Treasury bill; and finally

Any possible risks or benefits to the student loan programs under the Higher
Education Act of 1965 and to student borrowers.

This report also discusses topics that provide important context for the
issues raised in the legislative mandate. These include

The major Federal student loan programs;

The current economic and financial characteristics of the lender industry;

How alternative instruments can affect lenders? risks and returns, as well
as the risks faced by different size lenders when rates and spreads are
volatile and markets are illiquid; and

How using alternative financial instruments could affect the costs or
savings and risk for the federal budget and other federal concerns; and
could affect various non- lender participants in the programs.

The analyses in this report reflect discussions at the meetings of the Study
Group, interviews with Study Group members and other industry participants,
discussions with Study Group participants after the last meeting of the
Study Group, and reviews of drafts by Study Group participants. Analyses by
Education, Treasury, and GAO, subsequent to the last meeting, were included
in the drafts reviewed by Study Group members.

To address this mandate, this report analyzes the implications for lenders
and other parties of prospective loans under alternative formulae and
focuses on Stafford loans, the bulk of student loans.

OVERVIEW OF MAJOR FEDERAL STUDENT LOAN PROGRAMS

This chapter provides a broad overview of FFELP and DL, the two major
federal student loan programs--including a description of borrowers served,
the types of loans available, and the participants in the programs. It also
generally discusses the goals, constraints, and possible operating
strategies of FFELP lenders.

Chapter 1 4 TWO MAJOR FEDERAL STUDENT LOAN PROGRAMS FFELP and DL loans for
students pursuing postsecondary education include subsidized Stafford loans,
unsubsidized Stafford loans, PLUS loans, and consolidation loans. All FFELP
and DL loans receive federal support at least to the extent of (1) a
subsidized guarantee or direct federal credit exposure 3 , (2) a wedge
between the FFELP lender yields and borrower rates, and (3) subsidized caps
on borrower rates. At the end of federal fiscal year 1999, the total amount
of outstanding federally guaranteed loans, including direct loans, was about
$170 billion. (See table 1.1.) Most loans are Stafford loans to student.
PLUS loans represent loans to guardians or parents for the benefit of the
student and consolidated loans represent new loans used to replace existing
loans and do not increase outstanding student loan balances. (See table 1.1
and 1.2.)

3 Lenders pay the U.S. government an origination fee of 50 basis points (a
basis point is one- hundredth of a percent) for each loan. The lender
assesses a 3% fee on the borrower, which is paid to the federal government.
However, the assessment can be discounted to the student, and the student
may not pay the full 3%, even though the fee still is paid to the
government. The guarantee agency assesses a 1% fee on the borrower, which is
deposited in the federal reserve funds to pay for defaults. Although the
reserve funds are federal property, the Higher Education Act allows
guarantee agencies to discount the fee, which most do. Recently, the
Department of Education reduced fees for Direct Lending in response to the
reductions in FFEL fees charged by lenders and guarantee agencies.

Chapter 1 5

TABLE 1.1: OUTSTANDING LOAN BALANCES END OF FISCAL YEAR 1999 ($ MILLIONS)
FFEL PROGRAM DL PROGRAM TOTAL LOAN TYPE DOLLARS % of

Loan Type

DOLLARS % of Loan Type

DOLLARS % of Total Loans Stafford 1

Subsidized $65,443 77% $19,699 23% 85,142 49% Unsubsidized 28,565 72 11,216
28 39,781 23

PLUS 2 12,706 82 2,848 18 15,554 9

Consolidated 3 20,008 62 12,067 38 32,075 19

TOTAL 126,727 73 45,830 27 172,552 100 Notes: 1. The Federal government pays
the interest on subsidized Stafford loans while the

student is in school or otherwise not in repayment, while interest accrues
to the loan balance on unsubsidized Stafford loans. 2. PLUS loans are loans
made to parents and guardians for the benefit of their dependent

students. This figure also includes SLS loans, which provide supplemental
loans to students until the unsubsidized Stafford loan program was created
in 1994. 3. Consolidated loans replace one or more existing loans and do not
contribute to the

increase in outstanding balances. Percentages may not sum due to rounding.
Source: Department of Education

TABLE 1.2: LOAN ORIGINATIONS IN FISCAL YEAR 1999 ($ MILLIONS) FFEL PROGRAM
DL PROGRAM TOTAL LOAN TYPE DOLLARS % of

Loan Type

DOLLARS % of Loan Type

DOLLARS % of Total Loans Stafford

Subsidized $10,427 66% $5,318 34% $15,745 37% Unsubsidized 7,721 69 3,437 31
11,208 26

PLUS 1,908 61 1,198 39 3,106 7

Consolidated 4,720 37 8,006 63 12,726 30

TOTAL 24,826 58 17,959 42 42,785 100 Note: Percentages may not sum due to
rounding. Source: Department of Education

Chapter 1 6 STUDENT AND PARENT ELIGIBILITY FFELP and DL loans are virtually
a federal entitlement. Every student pursuing a postsecondary education on
at least a half-time basis at a school meeting certain gatekeeping
requirements may obtain funds directly and/or through his/her parents. For
the students and families served by these loans, the programs finance a
large proportion of the cost of both undergraduate and graduate
postsecondary education. The loans can be used at state four-year colleges
and universities, 2-year community colleges, private colleges and
universities, and for-profit trade and technical schools (commonly referred
to as proprietary schools).

THE STRUCTURE OF THE LENDER YIELD FORMULA IN STAFFORD LOANS Although
Congress has changed the maximum borrower rate and lender yield under FFELP
numerous times since 1965, this report will focus on comparing T- bill,
LIBOR and commercial-paper based formulae as required by section 802. 4 By
concentrating on the interest rate index for Stafford loans, the dominant
class of loans outstanding, we can illustrate the implications of
alternative indices for setting rates for lenders. In addition, by
concentrating on prospective loans, this analysis illustrates the
implications of alternative yield formulae based on 91-day Treasury bill
rates, commercial paper, or LIBOR interest rates. Historic rate changes for
Stafford and other loans are reported in appendix III.

For the rest of this report, the following terminology is used in describing
interest rates and lender yields.

Reference rate -- an interest rate on a cash market instrument used, or
referred to, in a formula for calculating another rate.

Markup -- an amount determined by law to be added to a reference rate (such
as a Treasury bill rate) to determine a new derived rate.

Student borrower rate -- a reference rate plus a markup.

Lender?s yield or interest income -- a reference rate plus a markup.

Margin -- a market-based add-on to a market rate; a margin can be negative,
and it changes based on market conditions and the creditworthiness of the
borrower.

Lender?s interest expense -- a market reference rate plus a margin.

Market rate -- the interest rate for commonly traded cash market financial
instruments such as Treasury bills, CP, or LIBOR.

Lender?s spread -- the difference between a lender yield and its funding
cost (interest income minus interest expense).

4 Before the change to a commercial paper based formula, the lenders yield
on student loans had been linked to Treasury issues.

Chapter 1 7

Special Allowance Payment (SAP) -- a payment that the federal government
makes to FFELP lenders that equals the difference between the rate a student
pays (which is capped) and what the current formula provides for lender
yields (which is not capped).

In this report, we assume the T-bill formula for lender yield put in place
in 1998 in the legislation that mandated this study. (See in table 1.3.) For
Stafford loans made on or after July 1, 1998 through December 31, 1999, the
student rate is the sum of the 91-day Treasury bill reference rate and a
markup. For loans originated in this period, the student markup is 1.7%
while the student is in school, grace or deferment 5 and 2.3% while the
student borrower is in repayment. The student borrower rate is adjusted
annually and fixed for the whole year. The maximum student borrower rate is
8.25%. The lender yield is the 91-day Treasury bill reference rate plus a
lender markup. The 91-day Treasury reference rate for lenders is set
quarterly based on the weekly auctions of Treasury bills during the quarter.
The lender markup is 2.2% while the student is in school, grace or deferment
and 2.8% when the student borrower is in repayment.

The formula for the lender yield includes a SAP which is intended to
maintain a yield or spread for the lender. The SAP permits students to pay
lower rates while encouraging lender participation in FFELP. The SAP
incorporated an explicit general subsidy of 0.5 percent for all Stafford
loans. The SAP is the lender rate minus the rate paid by students as long as
the difference is positive. If the difference is negative, the lender
receives the student rate and no SAP. If the quarterly lender yield
declines, so does the SAP generally. The minimum or ?floor? yield to lenders
is the student rate for the year. As described previously, the government
assumes the responsibility for borrowers? interest payments on "subsidized"
Stafford loans before the loans enter repayment and while repayment is
deferred.

5 Under a Stafford loan student have a 6-month grace period after ceasing
enrollment at least half time to start repaying the loan. In addition a
student while in school or deferment need not make payments on loans.

Chapter 1 8 Table 1.3: Maximum Student Borrower Rates and Lender Yields on
Stafford Loans

Originated From July 1,1998, To December 31, 1999

Reference Rate

Markup Reset Period

Maximum Rate or Cap for Student borrower

Minimum Yield or Floor For Lender

Rates and Yields While The Student Borrower Is In School, Grace, or
Deferment

Student Borrower Rate

91-day T-Bill

1.70 % Yearly 8.25% Student borrowers are not subject to the lender minimum
yield

Lender Yield

91-day T-Bill

2.20% Quarterly Lenders are not subject to the student rate cap

Student borrower rate, which is reset annually

Rates and Yields While The Student Borrower Is In Repayment Student Borrower
Rate

91-day T-Bill

2.30% Yearly 8.25% Student borrowers are not subject to the lender minimum
yield

Lender Yield

91-day T-Bill

2.80 % Quarterly Lenders are not subject to the student rate cap

Student borrower rate which is reset annually

NOTE: If the lender yield exceeds the students? rate, the government pays
the difference, i.e., the SAP. Source: Analysis

Chapter 1 9 Since the special allowance for the last three quarters of the
year is based upon the difference between the annual borrower rate and
quarterly rates, any drop in rates during those quarters reduces the general
subsidy, and to that extent it is symmetrical. The SAP is asymmetrical,
however, to the extent that lender yield, set quarterly, fall more than 0.5
percent below the borrowers? rate set in July, i.e., the reference rate
falls more than 0.5 percentage points. In such situations, the lender would
receive the student rate. The SAP is also asymmetric when the lenders?
formula rate exceeds the cap on the student rate.

FFELP lenders may offer lower interest rates to borrowers either when the
loan is originated or after periods of on-time payments. Prior to 1998,
lenders generally charged FFELP borrowers the maximum interest rate.
Occasionally, lenders offer some reduction in the interest rate after a
period of on-time payments. More recently, however, lenders have expanded
the availability of interest reductions for on-time payment and a few now
offer unconditional interest rate reductions.

BUDGET SCORING FOR STUDENT LOAN PROGRAMS

The cost of FFELP to the government, for budgetary purposes, is determined
following the terms of the Federal Credit Reform Act of 1990. The act
requires that the federal budget reflect the net present value of the total
cost to the government of loan programs in the budget year in which the loan
commitments are made. This means that the amount that is estimated to be the
government ?subsidy? over the life of the loan is recorded as part of the
federal budget in the initial year of the loan. The act therefore requires
agencies to estimate the cost of extending or guaranteeing credit, called
the subsidy cost. This cost is the present value 6 of cash disbursements by
the government, minus estimated payments to the government both calculated
over the life of the loan. For loan guarantees, the subsidy cost is the
present value of cash flows from estimated payments by the government (for
defaults and delinquencies, interest rate, the SAPs, and other payments)
minus estimated payments to the government (for loan origination and other
fees, penalties, and recoveries). The Credit Reform Act specifically
excluded administrative costs from the subsidy calculation. Changes in the
terms of FFELP loans require new estimations of subsidy costs. For example,
when the SAP terms are adjusted, the flow of payments over the life of the
loan changes as well, so the subsidy cost of the program is different. (See
appendix IV for an extended discussion of the Congressional Budget Office
(CBO) methodology for estimating budget effects.)

6 Present value is the value today of a stream of payments in the future,
discounted at a certain interest rate. Generally, when calculating the
present value of loan subsidy costs, agencies must use a discount rate that
is the average annual interest rate for marketable U.S. Treasury securities
with similar maturities to the guarantee, as specified in the Credit Reform
Act.

Chapter 1 10 The presence of the borrowers? interest rate cap and the floor
on lenders? yield poses particular problems for the budget scoring of
student loans. For example, under the standard scoring procedures, there is
no cost reflected in the budget for the interest rate cap when the interest
rate forecast is low. However, it is possible that at some point over the
life of the loan, the interest rate cap will be exceeded. The CBO uses a
methodology called ?probabilistic scoring? to capture the likely budgetary
costs associated with the caps and floors. CBO looks at historic volatility
of interest rates and estimates the likelihood associated with rates
exceeding caps, or falling fast enough to activate the floor, over the life
of the loans. CBO translates this likelihood into an estimated outlay for
its estimates of the student loan baseline and proposals affecting the
student loan program. 7

PARTICIPANTS IN THE STUDENT LOAN PROGRAM

This section describes the various types of entities involved in the student
loan programs in addition to borrowers: schools, the Department of
Education, lenders, loan servicers, guaranty agencies, investors, investment
banks, and credit-rating agencies. This discussion of program participants
provides the context needed to understand the discussion of possible changes
in the financial instrument used to determine the reference rate for lender
yield. 8

SCHOOLS Schools initially determine whether to participate in FFELP, DL, or
both. 9 Once the school decides which program it will participlate in, it
makes the certification necessary for a borrower to obtain a loan. Schools
in the FFELP program often provide borrowers a preferred or recommenced list
of lenders based on the services, loan terms (such as rate discounts for
on-time performance), and key service factors offered by the lenders. 10
However, borrowers in FFELP program schools are legally free to choose among
all eligible FFELP lenders.

THE U.S. DEPARTMENT OF EDUCATION 7 OMB did not object when CBO, in the fall
of 1999, used probabilistic scoring to determine the budget costs of
switching the reference rate to commercial paper. 8 Emerging Web products
may affect the relationships among program participants.

9 After incurring debt in an available program, individual borrowers may,
however, change programs by taking out a consolidation loan in the other
program. 10 Service competition by lenders involving ?inducements,? such as
unsolicited mailing of loan

applications or paying schools for referrals of loan applications, is
illegal.

Chapter 1 11 The Higher Education Act provides the broad structure of
program requirements, then authorizes the Secretary of Education (?the
Secretary?) to administer the program. Among the responsibilities of the
Secretary is the promulgation of regulations to provide detail on how the
requirements will be implemented. Sometimes, the statutory requirement is
very specific, in which case the regulation will simply restate the
statutory language (e.g., loan limits). In other cases, the statute
expresses the requirement in only the broadest terms, and gives extensive
authority to the Secretary to define standards of compliance (e.g.,
collection due diligence).

The Department is responsible for

determining a student?s eligibility to receive federal student financial
assistance;

gatekeeping, monitoring, and enforcement activities for postsecondary
schools;

recognizing accrediting agencies and administering the Quality Assurance and
Experimental Sites program, the Default Reduction Initiative, and Closed
School activities;

monitoring the participation of guarantors, lenders, secondary markets, and
third-party servicers in the Federal Family Education Loan Program;

collecting and resolving defaulted Federal Family Education Loans, Perkins
Loans, and Federal Direct Student Loans;

maintaining a centralized database on individuals that apply for and receive
federal student financial assistance,

! managing the financial aspects of the Student Financial Aid Programs, such
as receipt, disbursement, accounting, and financial reporting for program
funds;

! developing and disseminating information about the federal student loan
programs;

! developing cost estimates for the student loan programs; and

! providing technical support and information for financial aid
administrators. LENDERS General eligibility to originate and hold loans
under FFELP is limited by the HEA primarily to: (1) banks and certain other
savings institutions; (2) pension funds; (3) insurance companies; (4) one
state or private, nonprofit agency for each state; and (5) with certain
limitations, schools. Although other financial institutions and investors
are not eligible for direct participation as lenders in FFELP, the HEA
authorizes the use of trustee banks as eligible lenders to hold loans for
the benefit of others without regard to the latter?s own eligibility. Sallie
Mae 11 is eligible to hold loans originated by eligible originators.
Secondary

11 In 1972, Congress chartered Sallie Mae as a shareholder-owned
government-sponsored enterprise (GSE) with a purpose of providing liquidity
for the student loan market. The GSE provides this liquidity through the
direct purchase of insured student loans from eligible lenders

Chapter 1 12 markets 12 use trustee banks to originate loans, if they are
not directly eligible to do so themselves.

An eligible lender can approve and originate loans. 13 Once the loan is
originated, the lender can:

(1) keep the loan on its books and earn either a positive or negative return
and interest spread based on the lender yield and its own interest and other
expenses; (2) sell the loan to a purchasing lender and record the gain or
loss on the sale; or (3) securitize the loan by selling the loan to a trust
that has beneficial ownership

of the loans and funds its holdings by selling debt to investors, and book a
gain or loss depending on the terms of the transaction.

Some loan holders, who are not eligible lenders, purchase loans from
originators and some eligible lenders both originate and purchase loans.
These include:

the Student Loan Marketing Association (Sallie Mae), a state-chartered
holding company that owns both a federally chartered government-sponsored
enterprise, which is being liquidated, and a state-chartered subsidiary;

state agencies and private, not-for-profit entities; and

for-profit lenders, such as banks. In 1998 Sallie Mae was the largest holder
of FFELP student loans and banks dominated originations of FFELP loans. (See
tables 1.4 and 1.5.) 14 Both holdings and originations were concentrated in
larger institutions as shown in tables 1.6 and 1.7. The top 10 institutions
held 59 percent of outstanding loan balances and the top 10 institutions
originated 48 percent of the loan volume in 1998. However, ongoing market
developments are leading to changes in the concentration of both
originations and outstanding balances.

and through warehousing advances, which are loans to lenders secured by
insured student loans. The Student Loan Organization Act of 1997 authorized
the restructuring of Sallie Mae as a fully private state-chartered
corporation. 12 Secondary markets are financial institutions that are
specifically chartered to purchase student

loans from lenders and provide liquidity to the student loan market. 13 The
Higher Education Act prohibits discrimination by all eligible lenders based
on the standard

non-economic factors (such as race, religion etc.). There is no general
affirmative mandatory service requirement by any lender, and lenders are
expressively authorized to consider the financial situation of student
borrowers in determining whether to make loans. Certain forms of economic
discrimination are generally prohibited in consolidation loans. In making
PLUS loans, lenders are required to determine that a borrower or an endorser
does not have an adverse credit history. 14 Due to the data collection
cycles in student lending, the most recent data available to

determine concentrations are from 1998.

Chapter 1 13

Table 1.4: Total FFELP Program Loan Balances

By Type of Lender in 1998

Type of Lender $ billions % of total Sallie Mae $ 38.4 31.6% Secondary
Markets 29.4 24.2 Banks 51.7 42.5 Other 2.2 2.2 Total 121.7 100.0 Source:
U.S. Department of Education

Table 1.5: Total FFELP Program Originations

By Type of Lender in 1998

Type of Lender $ billions % of total Banks $18.0 80.4% Secondary Markets 2.7
12.1 Other 1.7 7.5 Total 22.4 100.0 Source: U.S. Department of Education

Chapter 1 14

Table 1.6: Concentration of FFELP Program Outstanding Loan

Balances Held By Type of Lender in 1998

$ Billion % of Total Top 10 $72.1 59.2% Top 25 93.6 76.9 Top 50 107.3 88.2
Top 75 113.1 92.9 Top 100 115.6 95.0 Total Lenders 121.7 100 Source: U.S.
Department of Education

Table 1.7: Concentration of Originations in FFELP Loans in 1998

$ billions % of total Top 10 $10.8 48.2% Top 25 15.0 67.0 Top 50 17.6 78.6
Top 75 18.9 84.4 Top 100 19.7 87.9 Total Lenders 22.4 100.0 Source: U.S.
Department of Education

Chapter 1 15 Under the HEA, each state can designate a not-for-profit
secondary market to help ensure that every student at every eligible
institution can receive a loan, and provide liquidity to originators.

LOAN SERVICERS Loan servicers undertake the processing necessary to ensure
that cash flows of the loans are recorded and transferred to and from
lenders, guaranty agencies, and the Department of Education. Holders of
student loans can service their own loans or contract out for loan
servicing. Sallie Mae, Citibank, Secondary Market Services, and some of the
other largest student loan lenders 15 tend to do their own servicing. Other
lenders often arrange for some or all of the FFELP loan origination and/or
maintenance functions to be performed by a third-party servicer. 16 If
servicing does not conform to detailed procedures promulgated by the U.S.
Department of Education, the lender may not be reimbursed should the
borrower default.

GUARANTY AGENCIES Guaranty agencies confirm eligibility, guarantee loans,
monitor status of loans, provide delinquency counseling/default aversion and
provide claims adjustments. FFELP loans are guaranteed as to 98 percent of
principal and accrued interest (100 percent in the case of death,
disability, bankruptcy discharge, closed schools and loans-of-last-resort)
by 1 of 36 state and private, nonprofit agencies designated on a state or
national basis by the Secretary of Education. 17 The federal government
reinsures up to 95 percent of the guaranty agencies' risk 18 and pays them
for loan-processing/issuance, account-maintenance and default- aversion
fees, and collection retentions. In addition, the federal government

15 Some large holders of student loans provide loan servicing to other
holders that may or may not be selling loans to such holders. Under such
arrangements, the secondary market may perform some or all of the marketing,
origination, funding, and/or servicing functions for the originating lender
of record while the student is in school. The loan may also be sold to the
secondary market when or before the loan enters repayment pursuant to a
forward-purchase agreement, often at a predetermined price. 16 Third-party
servicers are not subject to eligibility limitations by organizational type;
however

they must meet federal requirements of administrative capability and
financial responsibility, and they are subject to audit by the Department of
Education. Eligible lenders remain responsible for the performance of their
legal duties despite any delegation of functions to third-party servicers
and must monitor the latters' activities. 17 Loans under both programs must
be unsecured. FFELP lenders are neither prohibited from

obtaining endorsements of their loans by creditworthy individuals nor
required to obtain such private guarantees. There is no approved form for
lenders to secure endorsements on Stafford loans, and they do not normally
do so. Endorsements are used to overcome adverse credit histories for the
eligible borrowers in PLUS loans.

18 Reinsurance is reduced as low as 75 percent for high levels of default on
loans guaranteed by the particular guaranty agency.

Chapter 1 16 permits them to hold reserve fund accounts, which belong to the
federal government. The federal reserve accounts include any borrower-paid
guaranty fees, to cover their guaranty obligations and operational expenses.
The government also directly guarantees or reinsures FFELP lenders against
the inability of guaranty agencies to fulfill their guarantees because of
insolvency.

Federal reinsurance is the ultimate support for the value of a loan if the
borrower defaults, but it is available only if the guaranty agency correctly
enforces federal regulations and attempts to collect from delinquent
borrowers. If the loan servicing and collections are not done in accordance
with federal regulations, the federal guarantee can be voided and create
losses for the guaranty agency and the lender. As long as all parties in the
process are in compliance with the regulations, the guarantee substantially
eliminates losses due to default.

Guaranty agencies are authorized to collect a single insurance premium from
FFELP borrowers of not more than 1 percent of the principal amount of their
loans to cover the portion of their guaranty risk that is not federally
reinsured. Prior to reauthorization of the HEA in 1998, some guaranty
agencies had generally or selectively reduced or eliminated this insurance
premium. Since reauthorization, the elimination of guarantee fees has become
widespread due to market pressures. In addition, federal law states that the
reserves of the guaranty agencies are federal monies. Thus, any reserves
might be claimed by the federal government at some future time. 19

Figure 1.1 depicts the cash flows in a FFELP program loan, from the student
borrower and to and from the loan servicer, lender, guarantee agency, and
the Department of Education.

INVESTORS, INVESTMENT BANKS, AND CREDIT RATING AGENCIES Investors are a
crucial source of funding for student lending, because most of the funds
supporting student lending by FFELP lenders are raised in the capital
markets. Lenders issue debt or equity capital or create ABS to fund student
loans, which are sold to investors in capital markets. Credit rating
agencies and investment banks are crucial links between investors and
student lending. Investment banks advise lenders on the best funding
mechanisms and actually market or sell the equity, debt, or ABS to
investors. Credit rating agencies evaluate this debt or ABS in terms of
credit risk. The resulting credit ratings affect the interest that must be
paid to investors. As part of the evaluation, credit

19 Because the reserves belong to the federal government, interviews with
industry participants indicate that guaranty agencies have responded by
lowering or not charging the fee to gain competitive advantage. This
behavior depletes the reserves, while offering lower premiums to attract or
maintain current business. In addition, industry participants told us that
at least some guaranty agencies will be forced to re-institute the fees or
be unable to fulfill their guarantees once the reserves have been depleted.

Chapter 1 17 agencies evaluate the risks posed by servicers, guarantee
agencies, and regulatory changes.

Chapter 1 18 Investors and credit rating agencies consider the following
factors:

! the quality, term and status of the financed loan portfolio and underlying
loans;

! the current market funding costs;

! the extent of equity funding;

! the guarantees and other credit support for the issuance;

! the management capabilities of the issuer; and

Fi gure 1.1: Cash Flows From a FFEL Program Loan Bor r ower

Ser vi cer Lender

Guaranty agency Department of Educat i on Loan payments

Origination fee

Guarantee fee payments Loan payments net

of servicing fees Origination fees

Loan processi ng, account maint enance, default aversi on and reinsure costs
Default claims

payments Loan disbursement

Recoveries

Chapter 1 19

! the quality and cost of the servicing support for the loans. Such criteria
affect the financial instruments issued by both not-for-profit and for-
profit student lenders.

FOR-PROFIT AND NOT-FOR PROFIT LENDERS FACE SIMILAR RISKS THAT CONSTRAIN
THEIR OPERATIONS

For-profit lenders and not-for profit lenders have different goals and
constraints. Not-for-profit lenders involved in student lending are created
to serve the social good but many for-profit lenders, see student lending as
one of many lines of business. For-profit-lenders, therefore, must evaluate
whether and how to provide student loans. Both for-profit and not-for-profit
lenders must satisfy their customers, employees, other members of the
community, and investors in their debt. In addition, for-profit lenders must
also provide returns for their stockholders.

Almost all funding for student lenders comes from the capital markets. If
the funding is debt based, the lender must ensure that the interest and
principal owed to investors are paid on a timely basis. If investors who
provide funds to student lenders perceive a risk that might endanger
interest and principal payments, they may require a higher interest rate or
even refuse to lend funds to student lenders. For-profit student loan
lenders also obtain funding from equity stockholders. Stockholders, like
debt holders, consider the risk and return from investing but generally
expect higher returns because they face higher risks. Stockholders? risks
and returns are higher because stockholders? returns are based on expected
profits, which are a residual after all other obligations including interest
and principal are paid. If a student lender does exceptionally well, a large
profit is expected. If the student lender does poorly, there may be little
or no profit for the stockholder. If the student lender fails, the
stockholders may not even receive their initial equity investment because
other claimants against the failed student lender must be satisfied before
the stockholders.

Our review of annual reports and offerings to sell debt and our interviews
with for-profit institutions confirm a trade-off between risk and return. In
practice, however, for-profit firms measure returns and risks in different
ways or use several measures simultaneously. Returns are often discussed in
terms of operating profits or gains on sale. Risks are often discussed in
terms of volatility of the returns or the probability of losses of different
magnitudes.

Investors and investment banks view risks similarly. In discussing risks,
they often mention

interest rate risk or basis risk,

prepayment risk,

Chapter 1 20

credit risk and servicing risk,

political or regulatory risk, and

management risk. INTEREST RATE RISK Interest rate risk is the risk that
changes in interest rate levels, and the spreads among different interest
rates will increase the volatility of returns, reduce returns, or even
create losses. Interest rate risk is seen in the volatility of FFELP
lenders? returns and is an important issue for lenders because interest
expense is their largest expense. Their interest spread, a primary
determinant of the returns and risks, is the difference between interest
income, based on the formula yield and interest expenses, which usually vary
with movements in commercial rates such as LIBOR or CP. Even if the lender?s
return is based on the sale of loans, interest risk is important because
higher levels of interest rate risk decrease the value of the loan and the
gain on sale that can be earned.

Basis risk is a form of interest rate risk in which the interest income and
interest expenses are based on different instruments. When a FFELP lender?s
yield is based on the T-bill rate while its interest expense is tied to
LIBOR or commercial paper rates, basis risk is an important concern. If the
yield and interest expenses are both tied to the same interest rate, basis
risk is essentially eliminated. 20

PREPAYMENT RISK Prepayment risks exist when the rate of prepayment is
quicker or slower than the expected prepayment rate. A student loan is a
long term loan and can last up to 30 years after the student has entered
repayment. Lenders must make decisions about the amount and timing of debt
used to support the student loans. In making these decisions, lenders
determine the advantages and risks of alternative maturity patterns in the
debt. The decisions take into account the expected amount and maturity of
the student loan assets. However, lenders do not know when students will
enter repayment since they can defer repayment while pursuing further
education. The lender also does not know if the borrower will default. In
addition, a student may have several loans with one or more lenders and can
decide to combine such loans into one consolidation loan, which can have a
longer term to maturity than the original loans. Consequently, unexpected
movements in the amounts or maturity of outstanding student loans held by
lenders can create financial risks for the lenders.

20 A residual amount of basis risk may remain if the noninterest terms on a
lender?s funding differ from the noninterest terms on student loans. An
example would be timing difference in the contracts.

Chapter 1 21 Prepayment risk exists because (1) the returns on reinvestment
of prepaid loans fluctuate as interest rates fluctuate and (2) the
opportunity to recover the cost of originations (including any up-front fees
paid to the Department) or premiums paid to acquire the loans must be
amortized over a shorter period than originally expected. Financial analyses
for lenders assume a rate of prepayment by students when determining funding
needs. Unexpected early or late prepayments affect lenders adversely since
their funding amount and the timing of their own funding are based on the
expected prepayment rates on student loan assets.

CREDIT RISK AND SERVICING RISK Credit risk is the risk of loss from borrower
delinquencies (late payments) or defaults. Delinquencies lead to late
interest income and lead to collection costs. This lost income and
collection costs can create a need for added funding. Credit risk for
lenders is almost completely ameliorated by guaranty agencies and the
federal reinsurance guarantee. However, some credit risk remains even if the
guarantee remains in place. For example, when a loan is sold or when loans
are consolidated, credit risk exists until the payments are correctly
redirected to the new loan holder.

Servicing risk, that is, mistakes and errors, which may occur when servicing
the loan, can create credit risks for the lender, because improper servicing
can void the federal guarantee on student loans. The servicing requirements
are highly regulated with specific due diligence requirements detailed in
regulation.

POLITICAL OR REGULATORY RISK Political or regulatory risk exists for the
lenders, as they do in any government program. The lenders? yields and
eligibility requirements are set by federal statute and enforced by the
Department of Education, while interest costs are determined in the market.
Numerous statutory changes have occurred in the interest rate subsidies and
regulations dealing with guarantee fees and payments in the event of a
student borrower default. As shown in appendix III, lender yields have been
changed four times since 1992. The creation of the DL program in 1993 added
another element of political risk. With DL in place, the FFELP industry is
regulated by its direct competitor, which lenders told us poses additional
risk to the FFELP lenders. Going forward, the FFELP industry will continue
to face political risk. According to lenders, a change to a reference rate
based on a long-term Treasury rate with a 1% markup in July 2003 as required
under the current law, would worsen their trade-off between risks and return
and discourage future participation. Switching to a long-term reference rate
with a 1% markup as specified in the law ?need not imply an immediate crisis
in the market

Chapter 1 22 for guaranteed student loan but it could be problematic for
lenders in the longer term.? according to a Treasury study in 1998. 21

Management Risk Management risk is the risk that management will be unable
to manage the lender appropriately or make management errors. Mismanagement
by the lender can lead to bankruptcy and missed payments to investors.
Furthermore, mismanagement of a not-for-profit guaranty agency, independent
servicers, or secondary market can lead to bankruptcy or a state takeover of
the failing entity and create risks for lenders who depend on the guarantee
agency, secondary market, or independent servicer.

BASIC OPERATING STRATEGIES FOR LENDERS

The strategy that a lender adopts can be affected by the size of its student
loan portfolio, management concerns, and financial abilities. The four basic
funding options are

originating and permanently holding loans;

originating, holding temporarily, and selling loans;

buying and holding loans; and

securitizing loans. ORIGINATING AND PERMANENTLY HOLDING LOANS Originators or
others that hold loans on book fund these loans by selling securities or
equity or accepting deposits. Banks often use some or all such forms of
funding, while not-for-profit state secondary markets only issue debt
because they cannot accept deposits and cannot sell equity. Lenders? returns
depend on the interest rate spread (interest income less interest expense),
and other expenses, including servicing, default, and administrative
expenses and taxes.

ORIGINATING, HOLDING TEMPORARILY, AND SELLING LOANS Most lenders fund new
loans by selling loans that they have originated to other lenders in
secondary markets. Their returns depend on the prices that their

21 The Financial Viability of the Government-Guaranteed Student Loan
Program, U.S. Treasury Department February 1998.

Chapter 1 23 purchasing lenders are willing to pay for such loans. The
returns to firms that sell loans depend on the purchase price offered and
the costs of financing, originating, and servicing loans until they are
sold. The sales price requested by the seller depends on the advantages of
selling versus holding the loans. The purchase price offered by the
secondary market will depend on the returns the purchaser can earn by
holding the loans or placing them into an ABS.

BUYING AND HOLDING LOANS Buying and holding loans is similar to originating
and holding loans. In an originating and holding strategy, the acquisition
cost of the loan depends on the costs of originations. In contrast, in a
buying and holding strategy, loan acquisition costs depend on the price paid
loan originators. The price the buyer is willing to pay depends on the
return that can be earned by buying and holding the loan. Thus, the
purchasing lenders? return fundamentally depends on the same economic
factors that affect originators.

SECURITIZING LOANS Large lenders, both for-profit and not-for-profit, can
fund loans by selling a pool of loans to a trust that in turn is funded by
securities sold to investors-otherwise known as asset backed securitizations
(ABS). Such sales can affect funding costs based on the equity capital
needed to create the ABS and the interest expenses associated with the
securities issued by the trust. As with other funding approaches, this
approach is used when the overall risk/returns tradeoff appears to be
advantageous. In late 1998 and early 1999, securitization of student loans
had decreased significantly compared to prior years as the interest paid on
ABS securities increased relative to the yield earned from student loans.

When securitizations are undertaken, the lender often can sell the loans to
the pool. Accounting principles 22 require lenders to record any gain or
loss as income at the time of the sale. Thus, the expected value of all
future net revenues from the student loans is placed on the balance sheet at
the time of the sale. This transaction boosts current income for the lender
and decreases future income since the income that would have been realized
over the life of the loan is realized immediately upon the sale. However, if
loans in the pool do not perform as expected, losses may accrue to the
lender even after the sale to the pool. In a securitization, the lender has
typically retained the residual cash flows from the loans as well as the
servicing responsibility.

22 Financial Accounting Standards Board 125.

Chapter 1 24 When the loans are placed in the pool, they can only be used to
support payments to the investors in the ABS. This insulates the investor in
the ABS from potential financial problems and bankruptcy of the lender and
thus increases the credit rating of the ABS. This permits the debt to carry
a relatively lower interest rate. The value of this decrease in interest
rates on the debt must be balanced against the amount of capital needed to
fund the ABS and other costs and revenues associated with creating and
operating the ABS.

Sallie Mae has been a leader in securitizing student loans. Sallie Mae
securitized $1.0 billion of student loans in 1995, $6.0 billion in 1996,
$9.4 billion in 1997, $6.0 billion in 1998, $4.0 billion in 1999, and $6.5
billion by May 2000.

At times, the rates that must be paid on securities sold by the pool are
considered economically infeasible for lenders. For example, in late 1998
securitization ceased and in 1999 securitization of student loans had not
completely recovered because the returns required by investors increased
substantially compared to the interest earnings available from student
loans. This occurred, in part, because student loan rates were tied to
T-bills and the ?flight to quality? in the fall of 1998 created relatively
low T-bill rates compared to other rates. In addition, disruption in the
markets for long-term private debt securities generally affected student
loan ABS. However in the first part of 2000, securitization of student loans
has accelerated to higher levels. (See chapter 4 for further discussions of
this point.)

RELATIVE IMPORTANCE OF REVENUES AND EXPENSES

The primary source of income for all lenders is interest revenue and special
allowance payments from student loans. For example, in the third quarter of
1999, the gross yield to lenders on Stafford loans in repayment was 7.62%,
and total interest costs varied from 5.33 to 5.94% based on the funding
option used by the lender as estimated in chapter 4. 23 Interest expenses
varied based on operating strategies, funding sources, hedging decisions,
creditworthiness of the student loan lender, and the extent to which
borrowed funds are used to hold student loans. The level and volatility of
the interest spread 24 (interest income and interest expenses) is a primary
determinant of the return and return volatility for lenders and their
stockholders. (For a further discussion of the volatility of interest rates
and the spread between T-bill and commercial interest rates, see chapters 2
and 4.). Other expenses as percentage of assets are around 100

23 Interest expenses are a percentage of borrowed funds, while interest
income and SAP are a percentage of the student loan assets. 24 To determine
the interest spread as a percentage of assets, the interest expense as a

percentage of assets must be determined. This requires an adjustment for
equity capital funding.

Chapter 1 25 basis points 25 and vary with the efficiency of general
overhead and servicing, with some providers experiencing higher expenses and
others lower expenses. 26

Administering, servicing loan expenses and other overhead costs can vary
with portfolio size, operating efficiency of the lender or servicer, the
size of loans and the composition of the student loan portfolio.
Administering and servicing expenses also vary across several dimensions,
such as loan activity (for example, originations, claim filling, or skip
tracing) and student?s loan repayment status (for example, whether the
student is in school, is in the loan grace period, or in a loan repayment
period). Lenders with larger loan balances from schools with historically
low delinquency rates often pay a relatively lower servicing fee, or cost
per dollar of loans outstanding, than lenders with smaller portfolios with
higher delinquency rates. The actual servicing fee can be related to basic
factors that affect costs, such as delinquency rates, type of school, and
size of the loans or accounts. Independent servicers may charge for
indemnification, computer tape processing, report programming and editing,
as will as other specific services. In addition, lenders with large
portfolios often obtain size-based break points in servicing fee schedules,
according to some servicers, since servicing costs do not vary directly with
loan size and have a fixed cost component as documented by CBO and CRS in
1998. 27

For-profit lenders are often judged in terms of their return on equity by
stockholders and other participants in capital markets. The stability of
returns on equity over time is valued by investors. The return on equity
depends on the return on assets, taxes 28 and the lender?s equity to asset
ratio. As taxes or the equity to asset ratio increases, the return on equity
decreases. However as the equity to asset ratio decreases, leverage
increases and risk increases.

25 For example, in a 1998 study by the Treasury, servicing costs were
estimated to be 95 basis points per dollar of loan. During interviews with
servicers, conducted by Study Group staff, a range of 100 to 135 basis
points was reported by one company providing servicing to lenders. 26 This
is an estimate based on interviews conducted by GAO and includes the cost
associated

with marketing student loans. 27 The Financial Viability of the
Government-Guaranteed Student Loan Program, Department of

the Treasury, 1998, and The Profitability of Federally Guaranteed Loans,
Congressional Budget Office in a March 30,2000, letter to the Honorable
Pete. V. Domenici Chairman of the Committee on the Budget, March 30,1998. 28
The effect tax rate on corporate returns can vary but the average effective
corporate tax rate

as reported in the June 1999 Federal Reserve Bulletin was around 33 basis
points.

26

CHAPTER 2: PROPERTIES OF RELEVANT INTEREST RATES AND THEIR DETERMINATION

Lender returns on new Stafford loans made under the Federal Family Education
Loan Program (FFELP) were indexed to the interest rate on Treasury bills
from 1977 until the year-end of 1999 1 . Treasury bill rates have long been
benchmark rates in the short- term sector of financial markets - money
markets -- and have some unique features. With the growing breadth,
sophistication and internationalization of money markets, however, some
private rates have become increasingly important. Two commercial rates, the
London Interbank Offer Rate (LIBOR) and the Commercial Paper (CP) rate have
been considered as alternatives to the Treasury bill rate for the purpose of
indexing lender returns on FFELP loans. These rates have been identified as
representative of private money market conditions, which may be more
indicative of lenders? costs of funds than T-bill rates. This chapter will
consider properties of these alternative rates and their determination. It
will show their historical levels and the spreads between them, their
historical volatility and the outlook for rates as projected by CBO and the
Administration. Thus, this chapter addresses major parts of questions (1)
and (2) in the congressional mandate for the study which deal with comparing
the volatility and liquidity in the markets for T-bills, LIBOR or commercial
paper.

TREASURY BILL RATES

Treasury bills are issued by the federal government in maturities up to one
year. The shortest-term Treasury securities regularly issued are bills with
initial maturities of 91 days, the maturity relevant for the largest part of
FFELP. 2 Investors can purchase 91- day Treasury bills either by submitting
a bid at a weekly auction or on the secondary market, which is quite active.
Because of the depth of the secondary market, Treasury bills are frequently
regarded as the most marketable of money market securities.

1 Lender formula returns for most guaranteed student loans were indexed to
auction rates for 91-day Treasury bills, to which attention of the Study
Group was directed by the Congressional mandate. However, the borrower rate
for PLUS loans issued between 1987 and 1998 is set annually based on the
auction rate for 52-week Treasury bills in the last auction before June 1,
and SAP payments are made quarterly and indexed to 91-day T-bill rates. 2
Shorter-term, cash management bills are used to help manage the federal
government?s cash flow, but

these are not issued on a regular schedule.

Chapter 2 27 Treasury securities are free from credit risk. This absence of
credit risk distinguishes these securities from other commercial debt
instruments, including those considered in this chapter.

The interest rates on these (and other) securities move as money and credit
market conditions change, such as changes in the relative supply and demand
for alternative financial instruments as the federal budget continues to
generate surpluses that reduce the supply of Treasury issuances. Because
Treasury securities are not subject to credit risk, movements in their
interest rates are dominated by general money and credit market conditions.
Private sector debt instruments differ from Treasury securities along two
main dimensions: credit risk and liquidity. Changes in perceived risk and
liquidity of private instruments, and investor attitudes toward risk and
liquidity, are additional factors that move private sector interest rates.
Usually such changes do not move Treasury yields. In exceptional
circumstances, a ?flight to quality? -- i.e., a sharply heightened demand
for safety -- may cause the level of Treasury yields to decline as the
demand for Treasuries increases. Various Treasury security rates have been
used as benchmarks to help price riskier or less liquid fixed-rate
securities as well. In addition, Treasury yields are one of the money market
rates used to index variable-rate debt instruments, such as some business
loans and mortgages.

Treasury bill interest rates tend to rise when expected inflation rises,
when the economy (and therefore credit demand) is strong, when monetary
policy is tight, and when fiscal policy is loose. Figure 2.1 below shows the
interest rate on 91-day Treasury bills since 1973. (Appendix V discusses
data sources and calculations used in this chapter.) The effect of inflation
on Treasury bill rates is apparent at first glance: the rates rose in the
late 1970s as inflation rose and fell in the first half of the 1980s as
inflation fell. The figure also includes vertical line to indicate the four
recessions that the U.S. economy experienced in this period. As can be seen
from the figure, Treasury bill rates tend to fall during recessions and rise
during booms.

Chapter 2 28 Figure 2.1 - U.S. Treasury Bill Rates (showing bill rates,
inflation & recessions)

Source: U.S. Treasury Department

U.S. Treasury Bill Rate

0 2

4 6

8 10

12 14

16 18

20 7301

7312 7411

7510 7609

7708 7807

7906 8005

8104 8203

8302 8401

8412 8511

8610 8709

8808 8907

9006 9105

9204 9303

9402 9501

9512 9611

9710 9809

9908

percent, bond equivalent

Bill Rate CPI Inflation Rate

Recessions: Nov.'73 to Mar.'75 Jan.'80 to Jul.'80 Jul.'81 to Nov.'82 Jul.'90
to Mar.'91

Chapter 2 29 More specifically, Treasury bill rates rose with rising
inflation in the late 1970s and peaked at around 17 percent in 1981 and
1982. Then bill rates and inflation both fell, with inflation falling more
rapidly. Rates rose in 1989-90, as monetary policy responded to perceived
dangers of a reacceleration of inflation. They fell as the economy
encountered a recession and as the Federal Reserve encouraged recovery
through low interest rates into the first part of 1994. Reflecting the
strong economy that had emerged by then, Treasury bill rates turned up once
again in 1994, but they remained considerably below their average levels in
the 1970s or the 1980s. A sharp shift of investors? preferences toward safe
and liquid assets -- ?flight to quality? - developed in the fall of 1998,
following several international and domestic financial disturbances. This
?flight? from other assets into Treasury bills lowered Treasury rates, and
the ensuing easing of monetary policy held them below mid-1998 levels
through October 1999. In November, anticipation of the final Federal Reserve
policy change, which fully reversed the fall 1998 easing, pushed bill rates
back to early 1998 levels. 3

Table 2.1 shows the average rate on a bond equivalent yield, 365 day basis,
on 91-day Treasury bills at auction for successive two-year periods since
1973 and also shows the coefficient of variation, a measure of the relative
volatility of these rates. 4

? 5 The coefficient of variations is the measure of volatility or dispersion
of a variable divided by the mean value of the variable. Thus, the larger
the coefficient of variation the greater is the relative volatility of the
rate. As indicated in the table, greater relative volatility in the Treasury
bill rate tends to be associated with periods of cyclical disturbance and
volatile monetary policy: the battle against inflation and its consequences
in 1977-82 and the cyclical stimulus and subsequent restraint in 1991-94.

Each 10-15 year period also reflects its own institutional changes in
conduct of debt management and monetary policies, as well as external events
that shape the economy. For example, the 1940-55 period is largely
irrelevant because the Treasury and the Federal Reserve were freezing
short-term rates during most of that time, to facilitate financing of World
War II. Early in the 1970s, the international financial architecture was
substantially modified as the Bretton Woods framework of fixed exchange
rates and the peg of the dollar to gold were dismantled. Ceilings on bank
deposit rates (Federal Reserve Regulation Q) were largely phased out in the
early 1980s, and reserve requirements on off-shore branches of banks and on
international banking facilities in the United States were relaxed. These
changes accelerated the

3 There were three one-quarter point reductions in the Federal Reserve?s
?target? federal funds rate - the rate sought through open market operations
and hence the index of monetary policy -- in the fall of 1998 (September,
October, November) followed by three one-quarter point increases in July,
August and November, 1999 with further increases in 2000. 4 The rates are
monthly averages of rates at auctions during the month, expressed as
bond-equivalent

yields. Treasury bill rates are commonly quoted as ?bank discount? rates,
because their entire return results from the difference between purchase and
sale price. The bond-equivalent basis makes yields comparable among
instruments of differing maturities, taking into account the cash payment of
principal and interest . 5 Volatility is measured here by the coefficient of
variation, the ratio of the standard deviation of the rate to

its mean. The standard deviation is a statistic that measures the
distribution of monthly deviations (positive or negative) of actual rates
from the mean (average) rate for the whole period being considered.

Chapter 2 30 integration of domestic and overseas financial markets that had
been taking place gradually with the growing role of multinational
corporations. Greater sophistication of small investors, the evolution of
money market and other mutual funds, heightened competition between
commercial bank and nonbank financial sources for business all led to more
efficient markets, culminating most recently in the emergence of large scale
markets for certain financial derivatives. Because of the dramatic changes
in the financial landscape, consideration of other rates will be limited in
the discussion to follow to the period since 1973.

Chapter 2 31 Note:

The coefficient of variation is the ratio of the standard deviation to the
mean. The standard deviation is a measure of the dispersion of the
distribution of observations around the mean or the. period being
considered.

Source: Calculated (-----------percentage points )

Mean Level Coefficient of Variation 1973-74 7.70 0.13 1975-76 5.56 0.11
1977-78 6.44 0.21 1979-80 11.30 0.22 1981-82 13.00 0.21 1983-84 9.44 0.09
1985-86 6.94 0.15 1987-88 6.44 0.12 1989-90 8.08 0.06 1991-92 4.55 0.25
1993-94 3.73 0.23 1995-96 5.41 0.06 1997-98 5.08 0.06 1999 4.64 0.04

1973-84 8.91 0.35 1985-99 (Sept.) 5.69 0.27

Table II.1: Levels and Volatility of Three-month Treasury Bill Rates

monthly data from beginning to end of indicated period

Chapter 2 32

COMMERCIAL PAPER AND LONDON INTERBANK OFFER RATES

Many large, well-known companies, both financial and nonfinancial, issue
short-term, unsecured debt, commonly known as commercial paper. Maturities
for commercial paper range up to 270 days. Longer maturities would require
registration of the debt with the Securities and Exchange Commission. Most
often, commercial paper is issued with maturities of less than one or two
months; the average maturity is estimated by the Federal Reserve currently
to be about 30 days. 6 The paper may be either directly placed with
investors or most importantly mutual funds, or issued through dealers.
Dealers participate in the majority of the issuance and help to provide
liquidity in a market where the short-term nature of the paper works against
secondary market trading. (See the following chapter for a further
discussion of the characteristics of this market.) Despite the important
role of dealers in the market, historically, data on financial commercial
paper - the relevant data for the FFELP index issue - came from large
issuers who placed their paper directly with investors. 7 (See appendix VI.)

High-grade commercial paper -- that issued by borrowers with the highest
credit ratings from rating agencies -- is generally considered to be a
relatively safe asset. There has been only one major default on commercial
paper in the last 40 years, which occurred in 1970 when Penn Central
Railroad failed and defaulted on $82 million in commercial paper. Since
then, almost all commercial paper has been rated by at least one of the
major rating agencies.

Because the paper is unsecured, these credit ratings and the standing of the
issuer are particularly important in determining the interest rate. For
example, for nonfinancial paper, for which the Federal Reserve tabulates
rates on commercial paper by alternative credit ratings, the difference
between interest rates on AA-rated paper and the lower rated A2/P2 paper is
about 15 to 25 basis points; a spread that may widen in periods of credit
stress. Corresponding data are not available for financial commercial paper
but it seems likely that similar sensitivity to credit ratings would apply.
With the widespread use of ratings from rating agencies, the experience of
individual firms is less likely than previously to rock the market as a
whole. The rate for a particular credit rating reported by the Federal
Reserve is, however, a statistical calculation and rates are sometimes
quoted in the press by individual issuing firm name.

6 See the Federal Reserve web site, www.bog.frb.fed.us/releases/, which
provides not only data on commercial paper rates but information about the
instrument and the statistical series (the latter is also included as
appendix VI to this report). Staff comments have indicated that the average
maturity ranges between 30 and 45 days. 7 Presently, these data come from
the trust company that handles almost all CP transactions and is

considered to be highly reliable. This means there is now no role in the
data collection process for subjective quotation of rates by dealers.
Previously, for financial paper, dealers were not involved in the data
collection because the quotes were on directly placed paper. Consequently,
the statistical series for financial paper was not subject to the series
break, in September 1997 (when the new collection methods were adopted),
that applied to nonfinancial paper. See discussion in the following chapter.

Chapter 2 33 LIBOR, the London Interbank Offer Rate, is the reference rate
on U.S. dollar- denominated inter-bank placements in London. Hence, it is
the interest rate on dollar- denominated offshore loans from banks with
temporary ?excess? funds to banks facing strong demands for funds, either
for dollar-denominated loans overseas or from their home offices. These
interbank loans are for a fixed term and are made in large denominations.
(See appendix VII.)

LIBOR also serves as a reference rate for a number of other transactions.
For example, the Federal Reserve reports on reference rates used for
?repricing? (periodic re-setting of the rate) on bank loans, based on one of
their official surveys. The survey shows that, among domestic banks and U.S.
branches and agencies of foreign banks, so- called ?foreign? money market
rates - almost assuredly LIBOR in some currency -- are the most frequently
used reference rates for loans that mature or are subject to repricing
within 31 to 365 days. 8 This re-pricing period is the most comparable to
student loans. In the securities market, the Securities Data Corporation has
reported that about 70 percent of publicly issued, taxable floating rate
securities were indexed to LIBOR in 1998, up from about 50 percent in 1990.
The second most common index in 1998 appeared to be bank prime, accounting
for about 7 percent; the Treasury bill rate was the index for 3 percent of
the issues and commercial paper for less than one percent of issues.
Floating-rate securities were 43 percent of total securities issuance.

The British Bankers? Association?s ?fixing ? of LIBOR is the standard
measure of ?the rate? for LIBOR used in financial markets. On a daily basis,
the Association polls the offer rates of 16 large banks that broadly reflect
the activity in the interbank market. The banks? responses, for each
specified maturity of contract, are ranked and the average of the middle
eight determines the LIBOR fixing. 9

8 Information is taken from the Federal Reserve release, ?Survey of Terms of
Bank Lending.? This survey also shows that the largest share of new loan
extensions -- but probably a smaller share of loans outstanding -- is
accounted for by loans that reprice daily, with federal funds as the most
common base for pricing. The third most common base for pricing is the prime
rate, used most commonly for loans that reprice at the lender?s option or
have more than a year to repricing or maturity.

9 The banks in the BBA sample are active institutions frequently engaged in
the interbank placement market. Rates quoted for the fixing that diverged
from their other market activity would be noticeable, so rates in this
highly competitive market are likely to be representative of current credit
conditions and assessments of risks.

Chapter 2 34 Figure 2.2 shows the rates on one-month and three-month
commercial paper and the rate on three-month LIBOR since 1973.

Chapter 2 35 Figure 2.2: Private Sector Money Market Rates

Source: U.S. Treasury Department

Private Sector Short-term Rates

0 5

10 15

20 25

73017401 7501

7601 7701

7801 79018001

8101820183018401850186018701 8801

8901 9001

91019201 9301940195019601970198019901

percent, bond equivalent

libor 3-mos CP 1-mo. CP

Chapter 2 36 All three rates move roughly together, although significant
differences have occasionally arisen. The figure shows that, like Treasury
bill rates, these rates were high early in the 1970s, fell somewhat in the
mid-1970s, and then peaked around the end of the decade at the time of high
inflation and stringent anti-inflationary monetary policy. These private
rates then fell with recession and with the slowing of inflation after 1982;
they rose again in the late 1980s before dropping sharply in the early-1990s
recession. With continued slow inflation, rates remained low until the first
quarter of 1994. Then, they all turned up, reflecting a stronger economy and
an increase in the federal funds rate engineered by the Federal Reserve.
Like Treasury bill rates, LIBOR and commercial paper rates were lower, on
average, in the 1990s than in the 1970s and 1980s. And like Treasury bill
rates, they dipped in late ?98 - early ?99 on the basis of the Federal
Reserve?s policy easing and turned up subsequently as the Federal Reserve
moved to greater restraint. These private sector rates, however, moved up
more quickly than Treasury rates to surpass their year-earlier levels and
jumped in October as contracts were made with maturities near the turn of
the millenium. It appears that liquidity concerns surrounding possible Y2K
problems had a disproportionate effect on private instruments, while the
Treasury market benefited more from steps taken and announced by both the
Federal Reserve and the Treasury to ensure liquidity.

Table 2.2 reports the average levels and volatilities of LIBOR, three-month
CP and one- month CP, over two-year intervals and over the longer periods
before and after 1984. The similarities among these rates that appear in the
chart are confirmed in the tables. Indeed, with the allowance for
differences in rate levels that is embedded in the coefficient of variation,
the volatilities of the various rates are strikingly similar. All the rates
show sensitivity to the inflation and activist anti-inflation policy of
1977-82 and to the countercyclical swings in policy in 1991-94.

Chapter 2 37 Table 2.2: Comparative Levels and Volatilities of Money Market
Interest Rates

91-day Treasury Bill 3-month LIBOR 1-month LIBOR 3-month CP 1-month CP Mean
Level Coef of

Var Mean Level Coef Of

Var Mean Level Coef of

Var Mean Level Coef of Var Mean

Level Coef of

Var 1973-74 7.70 0.13 10.30 0.17 10.14 0.16 8.32 0.14 8.83 0.17

1975-76 5.56 0.11 6.35 0.14 5.83 0.13 5.84 0.11 5.53 0.13 1977-78 6.44 0.21
7.55 0.25 7.24 0.23 6.86 0.23 6.68 0.24

1979-80 11.30 0.22 13.30 0.22 13.05 0.24 11.46 0.19 11.89 0.23 1981-82 13.00
0.21 15.33 0.19 15.08 0.21 13.27 0.19 13.82 0.21

1983-84 9.44 0.09 10.42 0.10 10.22 0.09 9.57 0.09 9.59 0.10 1985-86 6.94
0.15 7.72 0.13 7.69 0.12 7.31 0.13 7.39 0.12

1987-88 6.44 0.12 7.68 0.11 7.50 0.11 7.18 0.11 7.16 0.11 1989-90 8.08 0.06
8.90 0.08 8.89 0.07 8.58 0.08 8.67 0.07

1991-92 4.55 0.25 4.97 0.25 4.90 0.26 4.81 0.26 4.76 0.26 1993-94 3.73 0.23
4.08 0.24 3.88 0.23 3.95 0.25 3.79 0.23

1995-96 5.41 0.06 5.85 0.05 5.79 0.05 5.69 0.06 5.67 0.05 1997-98 5.08 0.06
5.73 0.03 5.68 0.03 5.60 0.03 5.56 0.03

1999 (through

September) 4.55 0.04 5.09 0.04 5.02 0.03 4.99 0.04 4.93 0.04

Average for period: 1973-84 8.91 0.35 10.54 0.35 10.26 0.37 9.22 0.33 9.39
0.37

1985-99 5.71 0.27 6.38 0.28 6.29 0.28 6.12 0.27 6.10 0.28 Source: U.S.
Treasury Department

Chapter 2 38 In general, these commercial rates are much more similar to
each other than any of them is to the Treasury bill rate. Differences appear
from time to time, most significantly before 1985. As previously mentioned,
changes in market participants? perceptions of the relative risk and
liquidity of the instruments, and in market participants? attitudes toward
that risk and liquidity, can lead to differences among these rates and
Treasury bill rates. Some of these differences are discussed below in
looking at spreads between these private sector rates and Treasury rates.

SPREADS BETWEEN RATES AND VOLATILITY OF SPREADS

Interest rates on commercial or private instruments reflect credit risk,
which is the critical difference between them and Treasury rates. That risk
may be perceived to change over time, which leads to movements in these
rates in addition to the effects from the changing forces of aggregate
credit demand and supply. Consequently, rates on private instruments are not
likely to move in lock step with Treasury rates of comparable maturity.
Particularly at times of stress in financial markets generally, private
rates such as LIBOR or commercial paper may diverge from Treasury rates to
reflect larger risk premiums.

In the 1970s and early 1980s, the banking system was especially vulnerable
to sharp upward movements in short-term rates because regulatory ceilings
kept banks from responding with their own deposit rates, thereby creating
risk of deposit loss. And, even if they could have raised their deposit
rates, the effects on their profits would have been adverse as rates on
their assets were generally fixed for longer periods than the term of their
liabilities. Additionally, during the period of evolving integration of
onshore and offshore financial markets, banks played a major role in linking
markets and transferring funds. but information about counterparties was
incomplete and facilities for clearing transactions were less fully
developed and slower to function than in the electronic age. Consequently,
banks have been perceived to be especially vulnerable to financial stress,
and LIBOR has tended to move more in absolute terms than the Treasury bill
rate. As shown in table 2.2 (discussed in the preceding section), however,
the difference is usually very slight when movements in rates are scaled by
respective rate levels. The bill rate was more volatile in the recent,
1997-98, period because of the ?flight to quality? into bills, which lowered
bill rates as investors became increasingly concerned about the credit and
liquidity risks associated with other instruments.

As shown in figure 2.3, movements in the spread between 3-month LIBOR and
Treasury bill rates show the tendency for LIBOR to move up relative to the
bill rate in periods of stress. The Herstatt bank crisis, in June 1974,
involved only a small bank. International counterparties, however, were
unable for a period of hours to clear transactions potentially affected by
its failure because of time differences within Europe and between

Chapter 2 39

Figure 2.3:

Source: U.S. Treasury Department

7301 7601 7901 8201 8501 8801 9101 9401 9701 0 1

2 3

4 5

6 Spread between the Treasury Bill Rate and LIBOR

3-month maturities; monthly average spread in percentage points. Herstatt
Bank Crisis

End of Bretton Woods System and Oil Cartel Price Increases

Inflation and Dollar Problem and new monetary restraint

New Monetary Policy Regime Banco Ambrosiano failure

Continental Bank

Gulf War Global

FInancial Problems and Long-term Capital Stock Market

Crash of '87

Chapter 2 40 Europe and the United States. This hiatus created extreme
uncertainties about counterparties? actual circumstances and generally
underscored the vulnerability of banks with offshore exposure in this early
stage of floating exchange rates and integration of financial markets. At
the same time, inflation rates and international financial flows around the
world were disrupted by OPEC?s hike in oil prices. Changes in the
international financial structure and oil price inflation both could be
expected to affect Treasury bill rates, as well as rates on private
instruments. The uncertainties injected into the financial environment early
in the 1970s, however, seemed to imply new risks for the banking system that
heightened private sector risk premiums, in addition to increasing the
volatility of Treasury rates. Hence, the spread between LIBOR and Treasury
bill rates increased erratically in this period.

The second round of OPEC oil price increases in the late 1970s, coming in a
period already characterized by rising inflation, a depreciating dollar and
higher-than-usual peace-time budget deficits again, created an atmosphere of
uncertainty that was perceived as entailing heightened risks for the banking
system. In 1979, a new monetary regime was established by the Federal
Reserve that clearly entailed the likelihood of greater volatility of
short-term interest rates. Rates would be allowed to float to clear markets,
including the market for bank reserves, while the Fed targeted growth rates
of the monetary aggregates much more tightly than had been the case
previously (or has been the case for the past 15 years). In this more
volatile environment, bank regulators changed interest ceilings, and more
floating-rate debt instruments were developed to shift some of the risks of
rate volatility to banks? loan customers. Nevertheless, maturity and basis
mismatches on the balance sheets of depository institutions were cause for
concerns exhibited in the sharp spikes of the Treasury-LIBOR (or TED, for
Treasury-Eurodollar) spread.

These systemic factors have not been the only ones to affect the TED spread.
The failure, near-failure or substantial stress of individual banks -- Banco
Ambrosiano early in the 1980s, Continental Illinois Bank in 1984 and the
known exposure of other major banks to problem debts of Latin American
economies early in the 1980s --- all reflected private management decisions
as well as global economic forces. Indeed, as financial markets have become
more open and internationally integrated, and as market participants have
become more sophisticated, it could appear that shocks can be taken more in
stride. Markets overall may have become less turbulent even as individual
institutions are subject to substantial variations in rates if their credit
standing changes. The major stock market crash of 1987 caused a noticeable
bump up in the TED spread, but it was short lived, and significantly smaller
than the surges of spreads in the early 1980s. Similarly, although the Gulf
war in 1990 could have threatened another round of oil price inflation, the
reaction of the TED spread was relatively small. On the other hand, the
events of 1998 provided a test of market tranquility and showed that the
managerial decisions about risk-taking at individual institutions can be
decisive factors in the behavior of the TED spread and other risk premia.
Although the background of the Russian debt default created substantial
unease, coming on top of losses for some institutions related to earlier
exchange crises in South East Asia, the problems of a single hedge fund
provided the sharpest upward jolt to the TED spread in the fall of

Chapter 2 41 1998. Table 2.3 presents levels and volatilities of the spread
between private sector rates and the Treasury bill rate. It shows, as the
chart suggests, that the level of the TED spread was lower in the
1985-to-present period than previously and even relative to this lower
level, movements were smaller, giving a lower measure of volatility. The
years 1995-96, however, were exceptionally tranquil, and volatility jumped
back in 1997- 98. For the commercial paper rate, there is a pronounced
reduction in volatility of its spread to Treasury bill rates, beginning in
the mid-1980s. The spread between the CP rate and LIBOR (both at three-month
maturities) also tightened considerably at that time. Relative to that lower
level, however, the volatility of the spread was about constant over the
entire period considered since 1973.

The commercial paper market appears to be slightly less sensitive to the
concerns about bank credit risks than LIBOR, which is exclusively a bank
rate. Financial commercial paper is issued by commercial bank holding
companies and investment banks, but is also issued by the financial arms of
nonfinancial businesses and by finance companies serving the household
sector. Consequently, the risks that impinge directly on banks may affect
some parts of the commercial paper market only indirectly, making investors
in paper less quick to become fearful about the instrument overall. This
difference in perceived riskiness of highly rated financial commercial
paper, compared to LIBOR, is likely the major factor in the smaller spread
between three- month commercial paper rates and three-month Treasury bill
rates than the TED spread. The average commercial paper spread was only 41
basis points in the period 1985-to-date, compared to the average of 67 basis
points for the TED spread (see table 2.3).

In summary, the spreads between the private money market rates and Treasury
bill rates reflect, on the one hand, the private sector credit risks. On the
other hand, the occasional swings in bill rates from flights to quality and
supply surprises, and the continuous importance for bills in particular of
expectations about monetary policy, can add to the volatility of the
Treasury bill rate. Relative to their average spread, the standard deviation
of the TED spread was about 50 percent of its mean, while the standard
deviation of the three-month commercial paper spread was about 60 percent of
its mean. In this sense, the commercial paper spread was slightly more
volatile.

Chapter 2 42 Table 2.3: Comparative Levels and Volatilities of Spreads
between Money Market Interest Rates

3-month.LIBOR less Treasury Bill

3-month CP less Treasury Bill

1-month CP less Treasury Bill

3-month LIBOR less 3- month CP

Mean Level Coef of Var Mean Level Coef of Var Mean Level Coef of Var Mean
Level Coef of Var 1973-74 2.60 0.45 0.62 0.82 1.13 0.60 1.99 0.48

1975-76 0.79 0.53 0.28 0.76 -0.03 -9.07 0.51 0.61 1977-78 1.11 0.54 0.42
0.66 0.25 1.20 0.69 0.48

1979-80 2.00 0.34 0.16 2.96 0.59 0.86 1.84 0.46 1981-82 2.33 0.25 0.27 1.95
0.82 0.80 2.06 0.27

1983-84 0.98 0.31 0.13 1.07 0.15 1.64 0.85 0.23 1985-86 0.78 0.17 0.36 0.43
0.44 0.49 0.42 0.29

1987-88 1.25 0.26 0.74 0.37 0.73 0.32 0.50 0.31 1989-90 0.82 0.34 0.50 0.49
0.59 0.43 0.32 0.30

1991-92 0.42 0.40 0.26 0.59 0.21 0.73 0.17 0.29 1993-94 0.35 0.36 0.21 0.55
0.06 1.59 0.13 0.46

1995-96 0.44 0.17 0.27 0.26 0.25 0.55 0.17 0.22 1997-98 0.65 0.24 0.52 0.32
0.48 0.42 0.13 0.15

1999 (through September)

0.59 0.19 0.48 0.20 0.40 0.22 0.12 0.31 Average for period:

1973-84 1.64 0.60 0.31 1.33 0.49 1.27 1.32 0.67 1985-99 0.67 0.51 0.41 0.59
0.39 0.71 0.25 0.64

Source: U.S. Treasury Department

Chapter 2 43

THE OUTLOOK FOR INTEREST RATES AND SPREADS AMONG THEM

At the time of the final study group meeting, the outlook for interest
rates, taking into account credible forecasts for government budget
surpluses and a generally favorable inflation outlook, was for little change
in money market conditions over the foreseeable future (once transitory
disturbances associated with potential Y2K problems were overcome). Interest
rates for Treasury securities, forecast at midyear by the Congressional
Budget Office, the Administration and private analysts, are shown in the
following table.

Table 2.4: Treasury Bill Rate Projections by CBO, the Administration and
Blue Chip Consensus (in percent) Calendar Year: 1998

actual 1999 2000 2001 2002 2003 2004

T-bill rate: CBO 4.8 4.6 5.0 4.6 4.5 4.5 4.5 Admin 4.8 4.5 4.5 4.5 4.5 4.6
4.6 Blue Chip* 4.8 4.5 4.6 4.7 4.6 4.3 4.3 *Note: from June 1, 1999, Blue
Chip Financial Forecasts, semiannual survey of long-term forecasts, p.14.
This survey also provided forecasts of LIBOR, which were 30 to 40 basis
points above the corresponding forecasts of Treasury bill rates.

As the table 2.4 shows, these rates are strikingly similar. Private
forecasts show a decline in rates in the more distant years that likely
reflects the falling inflation rate projected by these same forecasters for
that period. Such a slowing of inflation is not projected by CBO or the
Administration. All of these forecasters appear to expect that the favorable
budget and inflation situation outweigh the upward pressure on interest
rates that otherwise might result from continued strong economic expansion
in the context of low unemployment.

Both federal agencies or branches make forecasts of the broad macroeconomic
indicators twice a year as part of the process of projecting the federal
budget. In each case, the projections reflect an assessment of developing
trends in the overall economy, blending qualitative judgments of staff
analysts and econometric models. The CBO?s assessment of the economy is
described in greater detail in its Economic and Budget Outlook, Fiscal Years
2000-2009, published in January 1999. The

Economic and Budget Outlook, An Update presents their midyear revisions with
somewhat less detailed overall economic assessment. For the Administration,
the projections of key economic variables that underlie forecasts of budget
revenues and outlays are shown in the annual budget documents and the
midsession review. The

Economic Report of the President (and accompanying Annual Report of the
Council of Economic Advisers) presents more of the rationale for the
contours of the economic forecast.

Both the CBO and the Administration forecast Treasury security rates in
order to make projections of interest outlays in the budget. Interest rates
on private securities are not

Chapter 2 44 routinely forecast, but CBO undertook a special forecast
exercise for this study. In this exercise, forecasts of LIBOR and CP rates
were made by projecting the spreads of these rates above the Treasury bill
rates that were forecast in the economic outlook. These forecasts of
interest rate spreads were based on historical relationships over a 30-year
period, taking into account the effects on rate spreads of inflation, the
steepness of the Treasury yield curve, and the volatility of the federal
funds rate as a measure of the monetary policy environment. (See appendix IV
for a discussion of CBO?s methodology.) As noted earlier, spreads between
private sector rates and the Treasury bill rate, discussed above, are larger
and more volatile when inflation is high, and when cyclical disturbances
induce active and changing monetary policy. The CBO approach models these
observations more formally, using an adjustment that also has the effect of
giving a somewhat heavier weight in their analysis to recent experience in
the more tranquil 1990s.

In making the forecasts of the Commercial Paper rate and LIBOR, CBO used
their baseline forecasts of the Treasury bill rate and the spread between
the bill rate and 10- year note rate. It was also projected that inflation
would remain near the fairly tranquil experience of recent years and that
monetary policy would continue to be relatively tranquil, implying that the
federal funds rate volatility would remain low. Given these assumptions, the
spread of LIBOR over the Treasury bill rate (3-month maturity in both cases)
was forecast to decline from its actual level of 77 basis points in 1998 to
68 basis points in 2001 and thereafter. The spread of the CP rate over the
Treasury bill rate was expected to fall from its actual level of 58 basis
points in 1998 to 42 basis points in 2001 and thereafter. CBO projections
and the corresponding projections of OMB are shown in table 2.5

Table 2.5: CBO Projections of T-bill Rates, Spreads and Corresponding Levels
of CP Rates and LIBOR (percent or percentage point)

Calendar Year: 1998a 1999 2000 2001 2002 2003 2004

T-bill Rate (bey)* 4.91 4.72 5.13 4.72 4.61 4.61 4.61 3-mos. CP Spread 0.61
0.47 0.46 0.44 0.44 0.44 0.44 3-mos. LIBOR Spread 0.73 0.62 0.66 0.68 0.69
0.69 0.69 3-mos. CP Rate 5.52 5.19 5.59 5.16 5.05 5.05 5.05 3-mos. LIBOR
5.64 5.34 5.79 5.40 5.30 5.30 5.30 Note: The T-bill rates are the same as in
Table 4.4, converted to bond equivalent yields. The spreads were projected
in CBO Memorandum ?A Framework for Projecting Interest Rate Speads and
Volatilities,? January 2000, page 39.

In the Fall of 1999, OMB did not have significant differences with, or
objections to, CBO projections of CP rates and scoring of the proposal
before Congress to change the index of lender returns on FFELP loans to CP.
For the 2001 Budget, however, OMB projected a spread of the CP rate above
the T-bill rate of 0.57, 11 basis points above the spread projected by CBO.

CBO?s scoring of interest rate changes in the student loan program includes
an estimate of the probable government costs associated with the likelihood
that the differences in

Chapter 2 45 lender yield and borrower rate will widen more than the CBO
forecast. The estimate is based on CBO?s projections of possible forecast
error that would lead rates to move around -- rather than equal -- the
expected or forecast level. Both the projected levels of spreads and the
probabilities of divergences from forecast were in the general range of the
historical experience summarized in the tables on spreads and volatilities
of spreads in this chapter. As the CBO report indicates, they expect, based
on continued noninflationary economic expansion and fiscal restraint, that
there will be a benign monetary policy environment and that spreads and
volatilities will be favorably affected. Given the unusual terrain of an
historically long expansion with budget surpluses not seen for decades, the
underlying Treasury bill forecast is subject to considerable uncertainty.
Furthermore, given the financial shocks that have occasionally disturbed
markets in recent years, many resulting from increased global integration of
financial markets, it would be imprudent to assume that the wider spreads of
the past, which are included in historical averages, could never again
occur. Therefore, in making the budget scoring projections discussed
elsewhere in this report, the uncertainty of the interest rate projections
as well as the forecast levels are taken into account.

46

Chapter 3: Liquidity of Alternative Instruments for Setting Lender Yields

The liquidity of a financial instrument is a valid concern when deciding
whether its interest rate should be used as a reference rate for other
financial contacts. The interest rate of a liquid instrument -- one that is
issued and traded in a deep and resilient market - is more likely to reflect
fundamental credit market conditions. It will be less subject than the rate
of an illiquid instrument to spurious fluctuations, or excessive influence
or manipulation by a limited number of issuers or investors. The
congressional mandate for this study (in questions 1 and 3) requested an
evaluation of the liquidity of the instruments whose rates were under
consideration for use as reference rates for lenders? returns from Federal
Family Education Loan Program (FFELP) loans.

The analysis that follows responds to the congressional request. In doing
so, two things should be noted: first, the FFEL program had indexed both the
students? rates and lenders? formula yields to the auction rate for
three-month Treasury bills, which suggests that it is the primary market for
bills that is relevant. Quoted Commercial Paper (CP) and LIBOR rates are
also basically primary market rates. ?Liquidity,? on the other hand, is
often thought of as a property of markets for secondary trading in already
existing instruments. The distinction between primary and secondary markets
may be overdrawn, however. The efficiency of a primary market will be
enhanced substantially by the liquidity of the secondary market for the same
security. This is so simply because a potential purchaser in the new-issue,
or primary, market will pay only a limited premium over the secondary market
price for an already existing security with essentially the same attributes.
A liquid secondary market may not be essential for primary market liquidity,
however, if the instrument has a short maturity and an essentially
continuous supply of new issues. The following discussion will utilize
attributes of both primary and secondary markets.

Second, it should be noted that the liquidity of the market for an
instrument serving as a reference rate does not necessarily carry over to
the markets for derivative instruments tied to this reference rate. This
consideration is relevant for FFELP loans because securities issued by FFELP
lenders that were indexed to the FFELP reference rate would be more
attractive sources of finance for these lenders if markets for these
derivative instruments were liquid rather than illiquid. The next chapter
considers some market attributes of key sources of funding for FFELP lender;
it also addresses interest rate movements in key periods and ways in which
lenders deal with risks posed by receiving income indexed to one reference
rate while funding themselves at other rates. Consideration of the
functioning of markets for floating rate notes, asset-backed securities,
swap contracts or other derivatives tied to Treasury bill rates, CP or LIBOR
is deferred to the next chapter.

All three alternative debt instruments whose rates were under consideration
for use as an index for Federal Family Education Loans -- Treasury-bills,
commercial paper (CP), and Eurodollar (LIBOR) instruments -- enjoy a high
degree of liquidity by conventional measures. They all trade in large
markets (with varying degrees of activity) and, as

Chapter 3 47 shown in the preceding chapter, generally react in similar
fashions to the same macro economic events. They differ largely in credit
quality, secondary-market infrastructure, and, in some cases, types of
investors. 1 In general, the three instruments under consideration are
competing for investor support, which forces market makers to provide a
degree of liquidity for investors in these markets even if trading among
other holders is limited. Members of the Study Group generally agreed with
these conclusions from the following analysis of the past liquidity of the
T-bill, LIBOR and CP markets. There was no consensus on how the liquidity of
these markets might be changed in the future by possibly substantial changes
in relative supply of these instruments.

LIQUIDITY IS THE ABILITY TO BUY OR SELL AN ASSET APPROXIMATELY AT THE
CURRENT MARKET PRICE

A security or a commodity (or the market for that security or commodity) is
said to be liquid if the instruments can be readily bought and sold at
approximately current market prices. Liquid markets or liquid securities
have sufficiently large volume outstanding and sufficiently active trading
for large transactions to be made without a substantial movement in price.
Furthermore, because of active trading among investors and/or dealers,
isolated events and/or erratic behavior by a single market participant are
unlikely to have major effects on the market price. Liquidity is a desirable
characteristic of a security for investors because they can move in or out
quickly with relative capital certainty. As such, illiquidity is regarded by
investors as a risk; thus, investors in illiquid debt instruments require
additional compensation, usually in the form of a higher yield.

Factors that influence the liquidity of a security are the presence or
absence of a large number of active market makers, widely available pricing
information (transparency), and a large homogeneous pool (with respect to
credit quality, issuer, maturity date, optionality, security age, etc., of
securities).

Measures of Liquidity

A number of generally recognized measures of liquidity are considered in
this analysis. Direct measures of liquidity are observable market
characteristics, such as bid-ask spreads and trading volume. An inferential
measure of liquidity is a market characteristic, such as issue volume and
outstanding amounts, that allows the analysts to make indirect inferences
about liquidity. Qualitative secondary market characteristics, such as the
number of market makers, transparency, and homogeneity, also allow for
inferences to be made about liquidity

1 Some investors may be prohibited from investing in particular securities.
Also investors? motivations for purchasing particular instruments may
differ.

Chapter 3 48

Direct Measures

Bid-Ask Spread: This measure is the difference between the bid price (yield)
and ask price (yield) of a financial instrument. In general, instruments
trading in more liquid markets exhibit narrower bid-ask spreads. This is due
to the fact that the bid-ask spread represents the compensation to the
market-maker for taking a position in those instruments. In deep, liquid
markets, there are a lot of securities and a large volume of transactions
and correspondingly a number of different market makers. Competition among
these market makers narrows the bid-ask spreads. Bid-ask spreads are also
affected by other factors, such as age, credit quality, maturity, and price
transparency. Bid-ask spreads for specific securities are probably the best
single indicator of liquidity.

Market Trading Volume: Higher trading volume (daily, weekly, or monthly
dollar volume of transactions) is associated with greater liquidity because
more, and potentially larger, transactions can occur without materially
affecting the price. Inferential Measures

Amount Outstanding: The greater the total face value outstanding at any
point in time, the greater the liquidity of the instrument. This is because
large outstanding amounts imply a large volume of tradable supply. Even
though large amounts of outstanding securities are necessary for an active
liquid secondary market, they do not guarantee it. 2

Issue Volume: Issue volume is the gross dollar volume of new issues sold in
the primary market over some period, i.e., weekly, monthly, annually. Like
large outstanding amounts, large issue volumes are generally associated with
higher liquidity but, again, do not guarantee liquidity. Often, new issues
undergo a period of active trading in the secondary market as the new supply
is distributed to investors.

Qualitative Inferential Measures

Market Makers: Generally, markets with large numbers of market makers are
associated with greater liquidity, generally reflecting significant amounts
of tradable supply and ensuring competitive forces that keep prices moving
smoothly with fundamentals.

2 This can be seen in the ?off-the-run? or ?seasoned? Treasury note and bond
market, where sizable amounts of particular securities are outstanding but
very little day-to-day secondary-market trading activity occurs. This is
because these securities, over time, get placed into investor portfolios and
do not trade often. When trades of any sizable volume do occur, price
movements can be significant, indicating a lack of liquidity.

Chapter 3 49 Price Transparency: The availability of real-time information
to market participants on the price at which current transactions are taking
place is another characteristic associated with liquid markets. (Price
transparency does not imply public information about the size or other
characteristics of particular transactions.) The existence of such
transparency in a market (with all the infrastructure and resources needed
to provide it) suggests demand for such information from market participants
and implies deep, active, and liquid markets. Price transparency also
enhances liquidity in that every market participant (buyer, seller, and
market maker) can observe the latest price at which transactions can be
made.

Homogeneity: Homogeneous instruments are usually fungible. They have similar
credit quality, issuers, optionality, etc. Greater homogeneity in a class of
investment instruments is also associated with greater liquidity because the
market makers? pricing systems are usually more efficient. That is, they do
not have to spend time adjusting individual market prices based on a myriad
of product-specific factors. Established systems for ratings and
determination of ratings by rating agencies enhance the liquidity of major
classes of private instruments.

Any of these measures taken individually (with perhaps the exception of
bid-ask spreads) can sometimes be misleading because of large differences in
characteristics between primary and secondary markets for each instruments
as well as large differences among secondary markets for each of the
instruments.

LIQUIDITY OF VARIOUS MARKET INSTRUMENTS

This section presents an examination of liquidity measures for each of the
three instruments whose rates are under consideration as reference rates for
FFELP lenders? returns. Because this examination is focused on the markets
for the reference-rate instruments, it does not cover all the liquidity
issues that may be relevant to the potential volatility of lenders? total
interest spreads between interest returns and interest costs. The latter may
be affected also by the market characteristics of derivative instruments
used for funding. These issues are addressed in the following chapter. The
following analysis is also largely focussed on data and professional
observations of the development of T-bill, LIBOR, and CP Markets to date. It
is recognized that the relative supplies of securities in these markets may
be changed substantially in the future by federal budget developments. Such
prospects are uncertain and entail judgments on which the Study Group
members did not agree. The differences in judgment are noted at the end of
the chapter.

Chapter 3 50

3-Month Treasury Bill (T-Bill) Market Liquidity

The 3-month Treasury bill market is characterized by a large number of
market participants, both buyers and sellers. There are currently 30 primary
government securities dealers; these are large money center banks or
investment banks that are recognized by the Federal Reserve for bilateral
transactions, are subject to review, and are subject to requirements for
participation in the government securities market. They stand ready to make
a market in all government securities, including T-bills. 3 In addition,
T-bills are a popular short-term investment vehicle for a number of
corporate, institutional, and private investors, including money funds and
banks, as well as official institutional investors such as central banks
here and abroad. 4 There is generally a large tradable supply because the
Treasury conducts bill auctions on a weekly basis. Two- way price
transparency is excellent, with a number of vendors providing market
participants with real-time pricing and trading volume information on
T-bills. Unlike CP, where rates differ by credit rating, or the possible
tiering in the Eurodollar market, T-bills have a unique and homogeneous
credit quality.

Bid-Ask Spreads

Bid-ask spread data for 3-month T-bills comes from the Department of
Treasury?s Office of Market Finance. The data, which are obtained by the
Treasury from the trading desk at the Federal Reserve Bank of New York
(FRB-NY), are available since August 1988. For the period from August 1,
1988, to August 20, 1999, the data indicate that the bid- ask spread (BDR
basis) 5 has averaged 1.95 basis points. The bid-ask spreads on 3- month
T-bills remained steady at 2 basis points until late October 1998, when the
spreads started to narrow to an average spread of 1.38 basis points in 1999.
(See figure 3.1.)

3 The number of primary dealers has fluctuated, rising from 18 in 1960, when
the system started, to a peak of 46 in 1988, largely reflecting trends in
the structure of the securities market. Primary dealers tend to be multiline
firms dealing in many other securities, as well as Treasuries. The
securities business and the number of firms in it expanded sharply in the
1980s, followed by a period of mergers and consolidation in the investment
and commercial banking business. 4 Of the stock of Treasury securities
outside of the Federal Reserve and federal government accounts as

of the June 1999, 39 percent was held by foreign investors and foreign and
multinational official institutions; about 16 percent was held by
individuals, businesses, and local government operating accounts; and about
19 percent was held by depository institutions and mutual funds. U.S.
Treasury Bulletin, December 1999, page 50. 5 Bank Discount Rate basis.

Chapter 3 51 Source: U.S. Treasury Department

Fi gure 3.1: Average Annual Bi d-Ask Spreads For 3-Month T-Bi ll s

2.00 2.00 2.00 2.00 2.00 1.99 2.00 2.00 2.00 1.99 1.88

1.38 0.00 0.50

1.00 1.50

2.00 2.50

1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999

Years Basis Points

Chapter 3 52 Secondary Market Trading Volume Primary dealers? transaction
data (purchases and sales) for all Treasury bills are collected by the
FRB-NY and published in the Federal Reserve Bulletin. The data, which we are
using as a proxy for 3-month T-bill trading volume, are monthly averages of
daily transactions since August 1988. The data show that bill transactions
peaked in 1996 and have since fallen back to 1988-1991 levels. For the
period August 1988 through April 1999, daily trading for all bills averaged
$40.1 billion, with a standard deviation of $9.2 billion (see figure 3.2.).
For all Treasury securities, dealer transactions volume tends to be a bit
above 5 percent of the outstanding stock outside of official Federal Reserve
and U.S. government holdings.

Chapter 3 53 Source: U.S. Treasury Department

Figure 3.2: Dealer Transactions in Treasury Bills Monthly Average of Daily
Data 8/88 to 4/99

0 10000

20000 30000

40000 50000

60000 70000

Aug-88 Dec-88

Apr-89 Aug-89

Dec-89 Apr-90

Aug-90 Dec-90

Apr-91 Aug-91

Dec-91 Apr-92

Aug-92 Dec-92

Apr-93 Aug-93

Dec-93 Apr-94

Aug-94 Dec-94

Apr-95 Aug-95

Dec-95 Apr-96

Aug-96 Dec-96

Apr-97 Aug-97

Dec-97 Apr-98

Aug-98 Dec-98

Apr-99

Date Million $

Chapter 3 54 Outstanding Amounts The total amount of T-bills outstanding
with 3 months to maturity is greater than the weekly issue volume of 3-month
bills. This is because sales of 3-month bills are re- openings of 6-month
bills that have been outstanding for 3 months. Similarly, every fourth
6-month bill is a re-opening of 1-year bills that have been outstanding for
6 months. In addition, there may be some outstanding long-dated
cash-management bills that have coincident maturity dates to the 3- or
6-month bills issued in regular weekly auctions. Total private holdings of
outstanding 3-month bills from January 1, 1990, to August 16, 1999, averaged
$22.0 billion (private holdings exclude holdings of the Federal Reserve and
other federal agencies and trust funds). Private holdings of 3- month bills
had a standard deviation of $7.2 billion. (See figure 3.3.)

Chapter 3 55 Source: U.S. Treasury Department

Figure III.3: Total Private Outstandings for 3-Month Bills 1/1/90 through
8/16/99

0 5000

10000 15000

20000 25000

30000 35000

40000 45000

50000 01/02/1990

04/02/1990 07/02/1990

10/02/1990 01/02/1991

04/02/1991 07/02/1991

10/02/1991 01/02/1992

04/02/1992 07/02/1992

10/02/1992 01/02/1993

04/02/1993 07/02/1993

10/02/1993 01/02/1994

04/02/1994 07/02/1994

10/02/1994 01/02/1995

04/02/1995 07/02/1995

10/02/1995 01/02/1996

04/02/1996 07/02/1996

10/02/1996 01/02/1997

04/02/1997 07/02/1997

10/02/1997 01/02/1998

04/02/1998 07/02/1998

10/02/1998 01/02/1999

04/02/1999 07/02/1999

Date Millions of $

Chapter 3 56 As figure 3.3 shows, private holdings of bills peaked in 1996
and fell in 1997 and 1998, before turning up in 1999. A GAO report 6 has
noted that there were favorable revenue surprises in 1997 and 1998, leading
to the reductions in the amount of bills outstanding. The report says:
?According to Treasury and Federal Reserve officials, the amount of bills
reduced was sufficiently large to cause the market for bills to become less
liquid.? It adds that, immediately following the April 1998 surge in tax
receipts, the Treasury began to take actions so that it could reduce debt
while maintaining liquid markets for its securities. These actions, which
continued in fiscal year 1999, included reductions in the number of note
issues (coupon securities with maturities of 10 years or less) and in the
frequency of issue of remaining note maturities. The Treasury also increased
the size of its regular bill issues in months of negative cash flow and made
more active use of cash management bills (bills with irregular issue dates
and maturities). This policy enabled the Treasury to cover irregular funding
needs with lower average cash balances -- in contrast to the more costly
strategy of greater reliance on periodic longer-term note issues and larger
average operating cash balances. As a result of this strategy, while
outstanding notes decreased 9 percent over fiscal year 1999, the amount of
bills

increased 2.4 percent.

Issuance Volume

The Department of Treasury is the source for the issuance volume data. Since
the beginning of 1990, the weekly average private issue size has been $8.4
billion, with a standard deviation of $1.5 billion. (See figure 3.4)

6 (GAO-AIMD-99-279) September 1999, Federal Debt: Debt Management in a
Period of Budget Surplus, p. 6.

Chapter 3 57 Source: U.S. Treasury Department

Figure 3.4 :Weekly Private Issue Amounts of 13-Week T-Bills 1-2-90 to 8-5-99

0 2000

4000 6000

8000 10000

12000 14000

01/02/1990 07/02/1990

01/02/1991 07/02/1991

01/02/1992 07/02/1992

01/02/1993 07/02/1993

01/02/1994 07/02/1994

01/02/1995 07/02/1995

01/02/1996 07/02/1996

01/02/1997 07/02/1997

01/02/1998 07/02/1998

01/02/1999 07/02/1999

Auction Date $ Millions

Chapter 3 58

90-Day Eurodollar Time Deposits (LIBOR) Market Liquidity

The Eurodollar time deposit (LIBOR) market is an interbank funds market for
banks operating overseas, in much the same way the Federal Funds market is
an interbank market between domestic U.S. banks .7 8 The interbank market,
or more correctly the interbank placements market, is a market for
Eurodollar time deposits. Having accepted dollar deposits, Euro banks can do
one of two things to earn a return, either make loans or place the funds in
the interbank placements market.

In the interbank placements market, Eurodollar deposits are sold either
directly by Euro- banks to other Euro-banks or through brokers. The rate at
which a bank offers Eurodollar deposits in the placements market is its IBOR
(InterBank Offer Rate). Each bank that operates in the placements market has
an IBOR rate that it changes depending on its particular supply and demand
for Eurodollar time deposits. The credit quality of any bank in the
placements market also affects the rate for transactions with that bank.
Brokers operating in the placements market, while not revealing the names of
counter-parties, reveal enough information to allow sellers of Eurodollar
time deposits to assess credit quality. The implication of the credit risk
inherent in lending to each bank is that time deposits bought and sold in
the placements market are not homogeneous.

While rates differ across banks operating in the Eurodollar market, a
reference rate has evolved that distills bank rates for highly rated banks
into a single rate. The most common reference IBOR rate is the British
Bankers? Association (BBA) 11 AM London Interbank Offer Rate (LIBOR) fixing.
The BBA LIBOR is a reference rate calculated from a survey of 16 contributor
banks (including some U.S. banks) operating in the placements market. Thus,
the BBA LIBOR represents a subset of all banks that operate in the
placements market. (For details on the BBA and how the BBA LIBOR is
constructed. (See appendix VII.) Because of the risk differentials charged
to various banks in the Euro market and the absence of comprehensive
real-time quotes in this over-the-counter market, price transparency is
significantly less robust for Eurodollars than for T-bills. Nevertheless,
the banks involved in determining the BBA fixing rate are sufficiently
well-informed and committed as participants in this market that the fixing
rate is considered quite representative of market fundamentals and has
become a widely used reference rate.

7 Marcia Stigum, The Money Market, Dow, Jones, Irwin, third edition, 1990. 8
Banks that operate outside the United States (foreign banks and U.S.
branches) make up the Euro

banking market. Eurodollars are U.S. dollars deposited for a fixed time
period in a bank located outside the U.S. Eurodollar time deposits differ
from Eurodollar CDs. This study concerns itself with Eurodollar time
deposits for which LIBOR is the reference rate. Marcia Stigum, The Money
Market, Dow,Jones,Irwin, third edition, 1990, p. 228. The latter are
negotiable and evolved as a means of providing liquidity to investors.
Eurodollar CDs trade at rates (1/8% to 1/4%) below the comparable maturity
Eurodollar time deposits.

Chapter 3 59

Bid-Ask Spreads

We calculated bid-ask spreads by using the 3-month BBA LIBOR against the
3-month Eurodollar Deposit Bid Rates as reported by the Federal Reserve
Statistical Release

Selected Interest Rates, H.15. These data show that the spreads have
averaged 9.2 basis points. Other literature indicates that spreads between
bid and ask in the Eurodollar time deposits market range around 1/8
percentage point but may depend on maturity and be subject to tiering, with
a narrower spread for top credit quality 9 (See figure 3.5.) Posted spreads
on major screens are about 1/8 percentage point.

9 Stigum, ib. id, p.889

Chapter 3 60 Source: U.S. Treasury Department

Figure 3.5: 3-Month LIBID - 3-Month LIBOR Spread: Monthly Data 1/77 to 7/99
Average Spread: 9.2 bps

-1.5 -1

-0.5 0

0.5 1

1.5 2

01/01/1971 02/01/1972

03/01/1973 04/01/1974

05/01/1975 06/01/1976

07/01/1977 08/01/1978

09/01/1979 10/01/1980

11/01/1981 12/01/1982

01/01/1984 02/01/1985

03/01/1986 04/01/1987

05/01/1988 06/01/1989

07/01/1990 08/01/1991

09/01/1992 10/01/1993

11/01/1994 12/01/1995

01/01/1997 02/01/1998

03/01/1999

Date Percent

Chapter 3 61

Outstanding Amounts

Data from the Bank of International Settlements (BIS) Quarterly Review:
International Banking and Financial Market Developments, show cross-border
dollar liability positions of all BIS reporting banks. These data include
Eurodollars of all maturities, not just 3- month maturities. (See figure
3.6.) They give a total volume of $3.4 trillion in 1998, and an annual
average growth rate of 5.9 percent over the preceding decade.

Chapter 3 62 Source: U.S. Treasury Department

Figure 3.6: Dollar Liabilities from BIS Reporting Banks Quarterly Data,
12/83 - 12/98

0 500

1000 1500

2000 2500

3000 3500

4000 Dec-83Jun-84

Dec-84 Jun-85

Dec-85Jun-86 Dec-86Jun-87

Dec-87Jun-88 Dec-88Jun-89

Dec-89 Jun-90

Dec-90 Jun-91

Dec-91Jun-92 Dec-92Jun-93

Dec-93 Jun-94

Dec-94Jun-95 Dec-95

Jun-96 Dec-96Jun-97

Dec-97Jun-98 Dec-98

Dat e $ (billions)

Chapter 3 63 30- and 90-Day Commercial Paper Market Liquidity The commercial
paper (CP) market is more similar to LIBOR than to T-bills. Like LIBOR, CP
issuers have different credit ratings. Virtually all large issuers have
ratings from credit rating agencies, however, so their standing is readily
determined. Rates differ by credit rating. The larger and highly rated
issuers place their paper directly with investors. Other issuers,
particularly smaller ones, place their CP through dealers. The rates on
commercial paper are collected by the Federal Reserve from the clearing
house for virtually all CP transactions and largely reflect new issue rates.
The rates are sorted by credit rating and yield curves are estimated to
determine rates by maturity. 10 Rates for financial paper from highly rated
issuers, and rates for highly rated and intermediate rated issues of
nonfinancial paper, are posted by the Federal Reserve daily, so there is
substantial rate transparency. However, because these rates are read from an
estimated yield curve, the specific rates quoted for a specific maturity may
not represent actual transactions.

Dealer-placed financial commercial paper accounts for 70 to 75 percent of
total outstanding financial paper according Federal Reserve statistics, up
from around 50 percent at the beginning of the decade. The dealer market is
dominated by a handful of firms that control roughly 60 to 70 percent of the
dealer-placed CP market. 11 Consequently, the behavior of specific dealers
plays a key role in this market.

Investors in CP are also concentrated. They are predominately institutions
(most often mutual fund --see figure 3.7) attracted by the higher yields
available on CP than on T- Bills and they tend to be buy-and-hold investors,
leading to low dollar volume of transactions in the secondary CP market once
the initial distribution is accomplished. 12 Although offer rates are posted
on some proprietary systems, the absence of widely

available two-way pricing data and transaction volume data also indicates
thin 10 There is a statistical break in series in the data for CP rates that
occurs in September 1997, when the Federal Reserve first began to collect
the data from the clearing house. Prior to that time the data had been
collected, for financial commercial paper, from finance companies that
directly placed their paper and, for other paper, from the dealers who
distributed it. It has been thought that there is some discontinuity in the
nonfinancial commercial paper rate series because of the change in the
reporting source for the data. There does not appear to be such
discontinuity for the financial paper, however, because the issuers who
directly place the paper and the clearing house both have, and report, the
same information. It is the financial paper that is relevant to CP as an
index for FFELP lenders? returns. 11 Merrill Lynch, Goldman Sachs, and
Lehman.

12 Fabozzi, Frank, Handbook of Fixed Income Securities, 3 rd Edition, p.227.

Chapter 3 64 Source: U.S. Treasury Department

Figure 3.7: Portfolio Holdings of Taxable Money Market Funds, 3/90 - 12/99

0 100000

200000 300000

400000 500000

600000 Mar-90

Sep-90 1991

Jun-91 Dec-91

Mar-92 Sep-92

1993 Jun-93

Dec-93 Mar-94

Sep-94 1995

Jun-95 Dec-95

Mar-96 Sep-96

1997 Jun-97

Dec-97 Mar-98

Sep-98 1999

Jun-99 Dec-99

Period (Quarter/Year) $ Millions

U.S. T- bi l l s Commerci al Paper Eurodollar CDs

Chapter 3 65 secondary market activity. One would expect that the
non-homogeneity described above would have an adverse impact on CP
liquidity. However, these effects appear to be somewhat mitigated by credit
ratings for virtually all paper and other homogenizing features of the CP
market, i.e. credit enhancements, regulations (which results in certain
conforming characteristics in CP), and business practices. 13

Furthermore, CP generally has short maturity - although maturities range up
to 270 days, they average about 30 days according to the Federal Reserve
(see appendix VI, which is the Federal Reserve?s summary on its CP
statistics). This short maturity, together with the virtually continual
supply, augment liquidity of the market, serving some of the same functions
as secondary trading.

Commercial Paper Bid-Ask Spread Information Anecdotal information from
traders in the CP market was obtained on bid-ask spreads. 14 Bid-ask spreads
in the CP market have narrowed since 1987, when spreads averaged 12.5 basis
points, to closer to 5 basis points. 15 The bid-ask spreads no longer
exhibit differentiation related to credit tiering, but some sector tiering
is still reported. 16 CP underwriters make the 5 basis point spread on CP
sold to investors at initial offering. Should an investor seek to liquidate
his position, the bid-ask spreads tend to be fairly tight (on the order of 2
basis points).

This tight spread appears to be related to prevailing business practices of
issuers and dealers, practices designed to enhance distribution channels for
new issues. Underwriters repeatedly sell new offerings to the same group of
institutional investors. It is often in the interest of the underwriting
firms to ?keep these investors happy? by providing a tight bid-ask spread
when the investor needs to liquidate a position. This creates good will
between the investor and the underwriters and enhances the ability of the
underwriter to place new product with a reliable buyer in the future.

13 Credit enhancements: All issuers of CP back their outstanding paper with
bank lines of credit. The maximum maturity for commercial paper is 270 days
because longer dated paper requires SEC registration, which is a costly and
time consuming process. CP is often categorized according to registration
exemptions sections found in the Securities Act of 1933. Most of the ?plain
vanilla? CP in the market is 3(a)3 commercial paper; this section exempts
from registration any CP with a maturity of less than 9 months and used to
finance ?current transaction?. The second class of CP is 3(a)2 paper,
sometimes referred to as LOC (Letter of Credit) CP. Section 3(a)2 of the
Securities Act of 1933 exempts any CP from registration that is ?issued or
guaranteed by a bank?. Finally, there is section 4(2) CP or ?private
placement? CP, which can be only sold to ?accredited investors? in order to
be exempt and is negotiable only to other ?accredited investors.?
Secondary-market trading activity in CP adheres to a long- practiced trading
protocol whereby the CP investor must first obtain bids from the CP
underwriter who originally sold the CP. If the investor rejects that bid, he
is then free to go out and solicit other bids from other CP dealers. 14
Telecon with traders at JP Morgan.

15 . Stigum, op.cit. p. 1051 16 There is reported to be some sector tiering,
with the finance CP treated more favorably in the secondary

market than the industrial CP. Finance paper, according to the Federal
Reserve statistics, is also a much larger portion of the CP market,
accounting for about 80 percent of outstandings.

Chapter 3 66 Outstanding Amount Data for the amount of CP outstanding are
available from the Federal Reserve. (See figure 3.8.) This series is used as
a proxy for the liquidity of both the 30- and 90-day CP market. The Federal
Reserve does not release regular statistics on the average maturity of the
CP outstanding. Conversations with Federal Reserve analysts, however,
indicated that they estimate the average maturity of their CP outstanding
time series to be in the 30- to 45-day range. The literature suggests that
the average maturity of CP has been increasing somewhat as issuers become
more sophisticated in evaluating interest rate risks, but the change has
been incremental. 17

Issuance Volume Total monthly CP issuance for all CP can be estimated from
the total outstandings data collected by the Federal Reserve if an
assumption is made about the average maturity of CP in the index. Assuming
that the average maturity is approximately 30 days, the implication is that
each month, the entire amount of commercial paper outstanding rolls over.
With such an assumption, the monthly issuance can be approximated by the
monthly outstanding amount. Estimates of issuance volume were done assuming
both a 30-day average maturity and a 45-day average maturity. (See figure
3.9.) These two estimates can be viewed as the upper and lower bounds of a
range of issuance volume.

17 Stigum, op.cit. p. 1050.

Chapter 3 67 Source: U.S. Treasury Department

Figure 3.8: Commercial Paper Outstanding Monthly Data 1/91 to 7/99

0.0 200000.0

400000.0 600000.0

800000.0 1000000.0

1200000.0 1400000.0

Jan 91 May 91Sep 91

Jan 92 May 92

Sep 92 Jan 93

May 93Sep 93 Jan 94

May 94Sep 94 Jan 95

May 95Sep 95 Jan 96

May 96 Sep 96

Jan 97 May 97Sep 97

Jan 98 May 98Sep 98

Jan 99 May 99

Date $ Million

Total CP Outstanding NonFinancial CP Financial CP

Chapter 3 68 Source: U.S. Treasury Department

Figure3.9: MonthlyCP IssuanceAssumingVariousAverage Maturities Jan 91toJul99

0.0 200000.0

400000.0 600000.0

800000.0 1000000.0

1200000.0 1400000.0

Jan 91 May 91

Sep 91 Jan 92

May 92 Sep 92

Jan 93 May 93

Sep 93 Jan 94

May 94 Sep 94

Jan 95 May 95

Sep 95 Jan 96

May 96 Sep 96

Jan 97 May 97

Sep 97 Jan 98

May 98 Sep 98

Jan 99 May 99

Date $ Millions

30-DayAvgMat 45-DayAvgMat

Chapter 3 69 Liquidity Patterns Over Time Detailed analysis of liquidity
patterns over time is made difficult by lack of comparable liquidity
time-series data for consistent time periods for all three alternative
indexes and by the fact that some information is anecdotal rather than
quantitative and some measures are proxies. Sometimes inferential and direct
liquidity measures diverge, highlighting both differences between the
primary and secondary markets for these instruments and the need for caution
in accepting inferential liquidity measures prima facie.

Table 3.1 summarizes the trends seen in the available liquidity data, from
the date that the respective data is available to present.

Chapter 3 70 Table 3.1 Measures of Liquidity

Measure 3-Month T-bill 3-Month LIBOR 1- & 3-Month CP Bid-Ask Spreads (Direct
Measure)

Static @ 2 bps to Oct. 1998 has narrowed to 1.3 bps.

Highly volatile to 1977; static at 9.2 bps to present

Narrowed from 12.5 bps in late 80s to 2 bps currently

Secondary Market Trading Volume (Direct Measure)

Federal Reserve data show dealer transactions for all bills peaked in 1996

Not Available Not Available Total Amounts Outstanding (Inferential Measure)

Total outstandings privately held rose between 1990 and 1996, then trended
down toward the current level

BIS statistics show overall market has tripled since 1983

Federal Reserve Statistic show outstandings have doubled since 1994

Issue Amounts (Inferential Measure)

Amounts issued to private holders rose between 1990 and 1996 and declined
from 1996 to present

Not Available Flat until 1994- doubled since 1994

Price Transparency (Qualitative)

High-Real-time, two- way pricing, trade size and transaction volume data

Some-bid-ask data-No transaction volume data.

Very little -- offer rates are posted, no transaction volume

Market Makers (Qualitative)

30 primary dealers; other institutions active in the market

16 contributor banks in BBA LIBOR Fixing

3 major dealers control 60 to 70 percent of market; some large direct
issuers

Homogeneity (Qualitative)

Homogeneous; Single issuer, top credit quality ?a T-bill is a T- bill?

Less homogeneous than T-bills; variety of credit qualities, sovereign risks

Least homogeneous, credit qualities, various issuers, various sectors;
offset by credit enhancements

Source: U.S. Treasury Department

Chapter 3 71 The discussion that follows analyzes trends and/or patterns in
the liquidity measures for each of the three alternative indexes.

3-Month T-Bill Liquidity Pattern

The data show that for 3-month T-bills, while issue volume and amounts
outstanding increased from 1990 to 1996, bid-ask spreads remained steady.
Secondary market trading volume for 3- month bills, for which the
transactions data for all bills served as a proxy, was flat from 1988 until
1992 and then increased by 60 percent from 1992 to 1996.

After peaking in 1996, trading volume declined and is currently at levels
seen in the late 1980s and early 1990s. New issue and outstanding amounts
fell from September 1996, through the end of FY1998, reflecting a reduction
in government deficits and, more recently, the emergence of surpluses.
According to Treasury officials, these quantity measures may be stabilizing
at a new lower level as debt management policy has moved to augment the
volume of weekly bill issuance relative to year-bills, notes, and bonds.

Bid-ask spreads on 3-month T-bill began to tighten in October 1998.
Financial market turmoil erupted in Asia and Russia, causing capital to flow
into Treasury securities. A ?flight to quality? associated with financial
market turmoil would be expected to have an impact on bill rate levels and
may influence bill liquidity if the capital flows lead to a scarcity
premium.

3-Month LIBOR Liquidity Pattern

Eurodollar time deposits (BID/LIBOR) spreads demonstrate the maturing of
this market by the mid-1980s. Spreads for 3-month maturities exhibited
tremendous volatility in the early 1970s, corresponding to poor liquidity
due to the nascent state of the Eurodollar market and failures in 1974 of
banks with Euromarket exposure, most notably the Bankhaus I. D. Herstatt and
Franklin National Bank. Volatility in the bid-ask spread waned in the late
1970s and early 1980's as the Eurodollar market grew and matured, and since
the early 1980s, the 3-month Eurodollar bid-ask spread has been fairly
static. Over much of this time period, the total (all maturities) Eurodollar
deposit market has grown dramatically, nearly tripling in size since 1983
according to BIS data. The growth in outstandings of this market mirrored
the growth in economic activity in the United States.

The data suggest that the correlation between Eurodollar bid-ask spreads and
Eurodollar volume measures is weak, as was also the case for T-bills. The
bid-ask spread has been stable while outstandings have increased steadily
since 1983.

Chapter 3 72

1- & 3-Month CP Liquidity Pattern

In the commercial paper market, bid-ask spreads have tightened from 1/8
percent, or 12.5 basis points to about 2 basis points, since 1987 according
to anecdotal information from CP traders and financial literature.
Meanwhile, outstanding amounts and new issue amounts have doubled since
1994, after remaining flat for the first part of the 1990s. The greater
volume largely reflects a growing economy as well as the trend toward more
security market financing and less dependence by businesses on banks.

The trends in CP liquidity measures are consistent, a pattern that would be
expected in a market becoming increasingly deep and liquid. However, caution
is required in interpreting the liquidity measures of the CP markets because
of the major role of a few market makers who adhere to a long-standing
protocol on buying and selling. (See discussion above.)

The narrowing of the bid-ask spreads in the CP market over recent years
appears to reflect a strategy of CP underwriters to enhance distribution
channels for CP in the initial placements market. Since most CP investors
are buy-and-hold investors and continuously buy new CP, underwriters are
willing to provide liquidity to ?their? customers when, on occasion, the
customers/investors need it. Thus, CP liquidity is an ?administered? feature
of the primary market rather than representing a deep, active, and
competitive secondary market.

LIQUIDITY LOOKING FORWARD: IS THE PAST PROLOGUE?

The advances in information technology, electronic trading, and expanded
trading hours would seem to enhance liquidity in most of these markets.
Furthermore, the increased sophistication of financial market participants
can be expected to bring these markets closer to each other in
characteristics because, as noted above, issuers in each market are
competing for largely similar investors.

Expected Reactions of Direct and Inferential Liquidity Measures to Changes
in Government Surpluses.

The members of the Study Group did not reach consensus on the likely
evolution of liquidity for alternative money market instruments in the event
of accumulating federal government surpluses. On the one hand, liquidity is
important to Treasury debt management because a liquid Treasury market is
efficient, and an efficient market ultimately lowers the Treasury?s
financing costs. The liquidity of the T-bill market is especially important
because T-bills are a major cash management tool with high auction frequency
that permits debt managers to adjust financing to meet concurrent cash
balance objectives. Consequently, Treasury debt management is committed to
maintaining a deep and liquid T-bill market.

Chapter 3 73 Treasury officials cited recent debt management actions in the
face of surplus budgets highlight Treasury?s commitment to maintaining
market liquidity. They noted that the Treasury has changed auction sizes and
frequencies, and eliminated certain longer-term notes. The net effect will
be to increase the size of bill auctions, even as issuance of other
securities declines. Treasury also has developed a program for buying back
outstanding securities. Treasury sees buy-backs as a tool for lowering
interest costs by concentrating outstanding debt among fewer issues, while
enlarging the size and enhancing the liquidity of new Treasury issues. (See
appendix XV, which presents Treasury Under Secretary Gensler?s comments at
the time of the February 2000 refunding.)

If surpluses were to unfold as currently projected, the future supply of
bills and future transaction volumes would decline from current levels.
Treasury analysts note, however, that the bid-ask spread in this highly
developed market has been shown to be fairly insensitive to changes in
outstandings and trading volumes. There have been a number of changes in
issue amounts and amounts outstanding over the last 10 years, as deficits
have expanded and declined, without any resultant changes in bid-ask
spreads. Treasury analysts, therefore, find it premature to draw conclusions
about the response of T-bill liquidity to trends in T-bill volume that will
take a considerable time to penetrate boundaries of past experience.

FFELP financial community members of the study group noted that a GAO report
18 suggested that challenges to debt management will become more demanding
as currently projected budget surpluses actually accumulate and that there
had been several instances in the past few years where unexpected surpluses
in the Federal budget has adversely affected the liquidity of the short-term
T-bill market. According to Treasury officials, significant consequences
could be some distance away. Further, some analysts, including FFELP
Community members, see liquidity problems emerging and only getting worse in
the T-bill market. (See appendix XIV for a statement of the lenders?
position.)

There should not be any direct impact on either direct or inferential
liquidity measures for Eurodollar time deposits and CP liquidity measures in
response to changes in the Federal government surplus. Historical trends in
the growth of these markets suggest continued growth in as the economy
expands.

How bid-ask spreads would react to increased activity in these markets is a
matter of conjecture. It would largely be contingent on the reaction of
Eurodollar and CP market makers to increased activity. Since CP secondary
market activity is very light, and the practices and protocols in the
secondary CP market are long-standing, with bid-ask spreads being
effectively administered by the market makers, the essential question is
whether market making will continue to expand in pace with economic
activity. The

18 General Accounting Office, ?Federal Debt Management in a Period of Budget
Surpluses? September 1999 (GAO/AIMD-00-270, p. 5)

Chapter 3 74 availability of market-making capacity will be important in
accommodating the increasing number of smaller issuers who use dealers to
place their paper. If that capacity were not to respond effectively, then
interest rates in this market could become increasingly sensitive to
short-term fluctuations in supply. However, conclusions on this subject are
only speculative - and are independent of federal budget developments.

75

CHAPTER 4: HOW LENDERS HAVE MANAGED INTEREST RATE RISK

This chapter discusses FFELP lenders? basis risk, which was embedded in the
program prior to the recent formula change to a commercial paper reference
rate. Basis risk existed because the lenders? spread (lenders? yield minus
lenders? borrowing costs) was volatile. The yield was tied to a T-bill-based
formula while the borrowing costs were often tied to commercial interest
rates that did not necessarily move in tandem. It also reflects movements in
borrowing margins, which vary with credit risk, and in hedging costs, which
vary with basis risk. Thus, a simple comparison of 91-day Treasury bill
rates and commercial interest rates will not necessarily explain movements
in any particular lender?s spread. This chapter describes program features
and lender strategies that prevailed before the January 1, 2000 change in
the reference rate and provides an analytical framework for answering
questions 4, 5 and 6 in the congressional mandate for this study.

Lenders and others involved in the financing of student loans have expressed
concerns about the basis risk in student lending for a considerable time. In
early 1998, as the Higher Education Act was in the process of being
reauthorized, the lending industry advocated a switch of the reference rate
to the commercial paper based yield.

The 1998-99 experience illustrated the effects of basis risk under adverse
interest rate conditions. A few lenders hedged their risks, which meant they
paid a fee for someone else to bear the risks created by movements in the
T-bill rate and the rates they pay to fund their operations. Such hedging
limited risks while at the same time lowering expected returns for lenders.
However, lenders told us that their ability to hedge risks was limited in
some situations, because no one was willing to accept a fee to bear the
risks or required fees so high the lenders were unwilling to pay someone
else to bear the risks. As a result, the lenders bore all the risks
themselves.

Private lenders and secondary market officials, whom we surveyed or spoke
to, generally preferred a yield formula based on commercial rates such as
LIBOR or commercial paper. Furthermore, lenders in the study group proposed
a new reference rate and markup up for lenders? yields based on commercial
paper. (See appendices VIII and IX for summaries of interviews and surveys
of lenders.)

LENDERS? YIELDS USUALLY MOVE IN TANDEM WITH T-BILL RATES

Before the recent change to a commercial paper based index, the maximum rate
that lenders could earn on most FFELP loans-their formula gross yield -
generally

Chapter 4 76 tracked quarterly movements in T-bill yields. 1 The formula
yield would not necessarily track the Treasury reference rate exactly. The
lenders? yield resets quarterly during the year, but the borrowers? rate is
reset annually on July 1 and is a floor on the rate earned by lenders. Under
the 1998 formula, discussed in detail in chapter 1, the lenders? yield would
rise if the Treasury bill reference rate rose during the year but could fall
only to the rate paid by students, which is set at the beginning of the
year.

A hypothetical history of lender gross yields is shown in figure 4.1. In
this figure the formula yield is calculated as it would have been if the
1998-99 formula had been in place since 1985. The figure shows that the
floor would have affected lenders? gross yields in three periods: early
1990, late 1996 and after the Russian loan default and financial disruptions
in 1998. A total of 11 out of 58 quarters would have been affected.

1 Actual gross yields could fall short of the formula yield to the extent
that lenders absorbed fees or provided discounts to students that were often
tied to a borrower?s repayment performance. The Study Group did not obtain
or evaluate data on discounting or fee absorption.

Chapter 4 77 Source: federal statistics and calcuated values

0 2

4 6

8 10

12 14

85.1 85.3

86.1 86.3

87.1 87.3

88.1 88.3

89.1 89.3

90.1 90.3

91.1 91.3

92.1 92.3

93.1 93.3

94.1 94.3

95.1 95.3

96.1 96.3

97.1 97.3

98.1 98.3

99.1 99.3

percent Lenders' Formula Yield Yield, no floor

FIGURE 4.1: LENDERS' FORMULA YIELD and YIELD WITH NO FLOOR (hypothetical
returns under the 1998-99 T-Bill based formula, both series in percent)

Chapter 4 78

LENDERS? INTEREST EXPENSES ARE RELATED TO BUT DO NOT MOVE IN TANDEM WITH
THEIR MARKET REFERENCE RATES

Interest expenses for lenders have not always moved in tandem with their
gross formula yield or quoted commercial rates. Each lender borrows at a
rate that reflects market rates, its credit risk, and financing terms,
including any costs to hedge risk. Often a lender?s borrowing rate is quoted
as the sum of a market reference rate and borrowing margin. This margin can
be affected by the term to maturity of the loan, exact financing terms and
options in the contract, and the creditworthiness of the lender. Sallie Mae
also pays a special offset fee when it buys loans and holds them on the
books of its government-sponsored enterprise, which affects its spread. 2
The lenders? borrowing margin can vary across time as well as across
lenders, as economic conditions change.

Traditionally, nonprofit secondary markets, including state-designated
secondary markets and other state chartered not-for-profit secondary
markets, funded themselves with a mix of tax-exempt 3 and taxable debt
issues. According to several industry observers, Congress has limited the
amount of tax-exempt authority or states have shifted tax exempt authority
to other activities. In addition, the relative importance of tax-exempt
funding for secondary markets has declined.

Tax-exempt funding may also create volatility between changes in lenders'
costs of funds and movements in the commonly quoted reference rates such as
CP or LIBOR. Tax-exempt yields are almost always lower than rates on taxable
instruments and they tend to be somewhat more volatile than other money
market rates on a month- to-month basis. They have the same contours,
however, across business and interest rate cycles. 4

2 Sallie Mae pays the government a 30 bp offset fee on loans held in
portfolio, effectively an increase in borrowing costs, when it borrows money
to finance these loans under its GSE operations. Under terms of its
privatization, Sallie Mae is required to wind down its GSE operations by
2008. After winding down the GSE operation the 30 basis point fee will be
eliminated, but Sallie Mae will be paying more for its non-GSE debt. 3 In
1980, in a double-digit interest-rate environment, the special allowance for
FFELP loans funded

with tax-exempt obligations was reduced to one-half the special allowance
paid on loans funded with taxable obligations, subject to a floor of 9.5
percent. In the low interest rate environment of the early 1990s, this
special treatment for tax-exempt funding was repealed, but the repeal did
not apply to obligations "originally issued" before the effective date. This
"grandfather" clause has been interpreted to permit serial refinancings of
then-outstanding tax-exempt obligations to qualify for the floor of 9.5
percent. 4 Available data on frequently resetting tax-exempt rates suggest
that they are slightly more strongly

correlated with LIBOR than with T-bill rates, but the difference is not
significant.

Chapter 4 79

Most FFELP Lenders Fund Operations At Variable Commercial Money Market Rates

The use of variable rate financing has been increasing and this financing
has been increasingly based on or set with reference to commercial money
market rates such as LIBOR or commercial paper. 5 Most for-profit lenders
borrow on the basis of commercial rates and ABS (asset-backed securities)
issued for student loans are usually tied to LIBOR or auction rates, which
move with LIBOR rates. Some issues are tied to T-bill rates. However, when
given a clear choice in recent Sallie Mae offerings, investors seemed to
prefer LIBOR-based variable rate investments. Figure 4.2 illustrates the
extent to which LIBOR, not commercial paper, is the basis for current
variable rate funding for all securities (both short- and long-term) in U.S.
markets.

CREDIT RISK AND TERM TO MATURITY AFFECT BORROWING MARGINS PAID BY LENDERS

Lenders may fund using financial instruments with different terms to
maturity. Even if all funding interest rates adjust quarterly, the term of
the funding instrument may affect the margin over the reference rate. The
borrowing margin over the LIBOR or CP rate for a one-quarter instrument can
be lower than the margin over LIBOR or CP for an instrument maturing in 5
years with quarterly adjustments. Such differences in margins can be
observed in ABS funding by Sallie Mae, in which longer term securities pay
larger margins than shorter term securities over the same reference index
rate.

5 Commercial rates such as LIBOR or commercial paper are used because added
funding can not be obtained easily from small deposit funding sources. To
raise funds, lenders may enter several markets, such as transaction
deposits, large negotiable CDs, and very short-term sources such as federal
funds and repurchase agreements, as well as a variety of notes and
certificates whose rates may be expected to move with LIBOR or commercial
rates generally.

Chapter 4 80 Source: Sallie Mae and Securities Data Corporation

Fi gure 4:2 LIBOR Domi nates Fl oati ng Rate Debt Issued f or Al l Purposes
0 100,000

200,000 300,000

400,000 500,000

600,000 700,000

1990 1991 1992 1993 1994 1995 1996 1997 1998

Years Dollars in Millions

All Floating Issues LIBOR Issues

Chapter 4 81

Large Lenders and Small Lenders Have Different Sources of Funds and
Sensitivity to Changes in Market Interest Rates

Large and small banks have different sources of funds and sensitivity to
changes in market wide interest rates, as shown in table 4.1. 6 In addition.
nonbank holders of student loans fund differently than banks. Although the
proportion of interest-bearing liabilities that fund operations at all banks
are roughly similar, the mix of interest bearing liabilities differ. For
example, the largest 10 commercial banks have over 20 percent of their
deposits in foreign offices, compared to less than 0.1 percent for the
smallest banks. Rates on foreign deposits likely varied very closely with
LIBOR (probably about 12 basis points lower). Federal funds purchases and
repos range from nearly 10 percent at the largest 10 banks to less than 2
percent at the smaller banks, and rates for such funding vary with market
wide conditions. In contrast, small denomination deposits vary from about 6
percent of assets at the largest 10 banks to over 30 percent of assets at
the smaller banks. Rates on small time deposits are lower and considerably
less volatile than LIBOR or T-bill rates, tending to follow money market
rates with some lag and damping fluctuations. Large time deposits range from
less than 5 percent at the largest banks to 11 percent at the smaller banks.
However, interest rates on large time deposits (likely negotiable
certificates of deposit) move very closely with other money market rates.
Since 1985, they have averaged about 25 basis points less than LIBOR
(3-month maturity in each case) and have been slightly less volatile. In
addition, some small banks that offer student loans under FFELP have
specific sources of funds tied directly to student lending. For example, in
student lending, an institution might use an advance from another
institution. This advance would be used to originate the loan, and the
student loan would be used to secure the advance. Historically, Sallie Mae
has provided such advances to other FFELP lenders. 7 In general, funding
sources at larger banks tend to be from international sources and active
markets such, as federal funds and repos, while smaller banks have funding
sources that are less volatile or less tightly tied to market conditions.

6 ?Profits and Balance Sheet Developments at US Commercial Banks in 1998,?
Federal Reserve Bulletin, June 1999, pp.369-395. 7 Lenders also use repos to
fund student lending. A repo (repurchase agreement) is the sale of an

asset at a given price to another party with an agreement or obligation to
repurchase that asset at some later date at a higher price. The differences
in sales prices and purchase price is the interest paid for the use of the
other institution?s funds. The sale and repurchase is similar to a fully
collateralized loan. If the lender does not repurchase the asset, it stays
with the lender as if it were collateral on a loan.

Chapter 4 82

Table 4.1: Funding Sources for Commercial Banks By Asset Size Class

(% of assets)

Largest 10 banks

Banks ranked 11-100

Banks ranked 101-1000

Banks ranked over 1000

Liabilities 92.67% 91.63% 90.54% 89.53% Interest Bearing Liabilities 65.81
73.46 75.44 75.35

Deposits 47.65 51.52 62.45 71.76 In Foreign Offices 20.17 8.16 1.29 0.07 In
Domestic Offices 27.48 43.36 61.16 71.70

Other, Checkable Deposits 0.99 1.75 4.24 11.17 Savings 15.84 21.42 25.66
19.01 Small Denomination Time 6.03 12.83 21.25 30.42 Large Denomination Time
4.62 7.36 10.01 11.10 Gross Federal Funds purchased 9.79 9.48 6.16 1.50
Other 8.37 12.46 6.83 2.09 Non-interest Bearing Liabilities 26.37 18.17
15.10 14.18

Demand deposits in domestic offices 8.46 12.41 11.89 13.08 Revaluation
Losses on OBS items 7.66 0.76 0.01 0 Other 10.64 5.01 3.20 1.10
Miscellaneous Capital Accounts 7.33 8.37 9.46 10.47

Source: Federal Reserve Board

Chapter 4 83 Large and specialized FFELP lenders are even more sensitive to
changes in market rates than are banks because they do not have a deposit
base. Among the largest holders of FFELP loans -- Sallie Mae and the large
banks or specialized bank subsidiaries -- a range of financing options is
available such as T-bill, LIBOR, and commercial paper securities for both
long and short-funding. Large diversified banks and other holders of student
loans may also raise funds in the floating rate notes markets, typically
with maturities of one to three years. Sallie Mae in its government
sponsored enterprise operations, which must be wound down by the end of
2008, raises funds with floating rate notes indexed to Treasuries.

Large diversified banks and financial holding companies easily issue
floating rate notes, which are frequently placed with investors. These
investors include insurance companies and pension funds, which desire at
least some investments with returns based on short-term rates which are
higher than Treasury bill yields. Rates on these notes do not depend only on
the market reference rate to which they are tied. The rate also depends on
the credit rating of the issuer, the exact maturity of the note and
financial market conditions when the note is issued. Hence the cost of this
funding is subject to money market volatility but its availability tends to
be more reliable than long-term funding for securitizations.

BASIS-RISK, HEDGING, AND EXPECTED RETURNS

If lenders? yields move with T-bill rates while much of their interest
expense moves with commercial rates, they face basis risk. This risk exists
because commercial rates, including their borrowing margins, are not
perfectly correlated with T-bill rates, which creates volatility in the
interest rate spread earned by lenders.

Swaps 8 are a mechanism for hedging this risk that can be used by lenders to
ensure that interest income and interest expenses both move with the same
market interest rate. The swap eliminates some or all of the basis risk. In
a swap, one party receives payments based on one interest rate and pays the
other party based on another interest rate. In effect, a swap permits a
lender to change the interest rate on his expenses or revenues from the
underlying business. A lender, using a swap, earns a lower expected and less
volatile interest spread. If hedging or swap costs

8 A interest rate swap contract is one in which two parties exchange (swap)
interest rate payments based on a notional value for the principal for the
term of the contract. In a basis swap, both parties pay a variable rate
referenced to a market rate, such as T-bill, LIBOR or CP. One party pays
based on one rate while the other part y pays based on the alternative rate.
In student lending the T-bill payer has paid T-bill plus a margin and
typically received in exchange a LIBOR or CP. The margin over T- bill
reflects the spread between T-bills and the other rate as well as the risks
of movements in the spread over the life of the swap contract. In contrast,
a simple swap involves one party paying based on a variable rate, while the
other pays a fixed rate. A basis swap can be viewed as two simple swaps tied
together. One exchanges a variable T-bill rate for a fixed rate, and the
other exchanges a fixed rate for a variable LIBOR rate.

Chapter 4 84 increase significantly in comparison to the spread and its
volatility, the value of hedging to the lender may decline to the point
where the lender may hedge less or even not at all.

If swap dealers providing hedges anticipate higher spreads or greater spread
volatility in the future, the price they would charge for hedging would
increase to ensure that they can provide the hedge and still earn a profit.
If spreads are increasing and the future spread volatility is unpredictable,
swap dealers may abandon the market, charge high rates, or limit the amount
they would hedge, which would make the market less liquid and limit the
ability of FFELP lenders to hedge their basis risk.

Table 4.2 illustrates how a swap affects expected spreads for a simple
lender, earning a yield based on T-bill rates and paying interest based on a
commercial rate (CR). The spread for a lender who is not hedging depends on
the T-bill rate, the markup (280 BP), the commercial rate, and its borrowing
margin. Because the lender has not hedged, it could experience, in the
future, a decline or increase in its spread. 9 If the borrowing margin
increases due to general concerns about credit quality, the spread is
further squeezed.

Assuming the lender hedges with a swap, the spread for the hedged lender
depends on the markup (280 BP), borrowing margin, and swap spread. (Other
hedging mechanisms are available.) As the borrowing margin and swap spread
increase, the hedged lender?s locked-in spread decreases, but the hedged
lender?s spread on the loans is not affected by movements in the underlying
T-bill and commercial rates after the hedge is in place. Swap spreads and
borrowing margins can vary over time. If the original hedge by the lender is
not for the full term and amount of the student loans, the hedged spread can
change over time. The costs of replacing a hedge can vary as aggregate
interest spreads and swap spreads vary with market conditions, and as lender
and counterparty evaluate the probability of increased basis risks in the
future and each other?s creditworthiness. In addition, swaps for the right
amount or for the right term to maturity, which matches the life of the
underlying asset may be unavailable or be considered inordinately expensive
by lenders at certain times. In addition, the borrowing margin for credit
risk can vary over the life of the student loan assets as market conditions
and the lender?s own financial condition change.

9 The statutory formula yield includes a floor rate during each year. In
theory, this floor rate protects the lenders against large downward
movements in interest rates. In practice, the floor in the current T- bill
based formula has been invoked slightly less than 20% of the time since
1985.

Chapter 4 85 Table4.2: An Example of a Lender?s Interest Spread With and
Without Hedging

Without Hedging With Hedging Lender Yield T-bill + 280 BP T-bill + 280 BP
Lender Interest Expenses

Commercial Rate(CR) + Borrowing Margin(BM)

CR+ BM Gross interest Spread

(T-bill + 280 BP) - (CR+ BM)

(T-bill + 280 BP) - (CR + BM) Hedging Flows + CR

- (T-bill + swap spread (SS)) Net Interest Spread

(T-bill + 280 BP) - (CR + BM) (T-bill + 280 BP) - (CR + BM) + [+ CR -
(T-bill + SS)] = 280 BP - (BM +SS)

Source: Calculated

Hedging Costs Vary Over Time and Lenders

Lenders told us that swap spreads can be higher or lower than the cash or
spot market interest rate spread at any given time. 10 Lenders also noted
that the short- term cash interest rate spread is based on current
conditions, while swap spreads also depend on anticipated future rates and
their volatility. When future spreads are expected to be volatile, swap
spreads tend to be higher than when less volatility in the spread is
expected. In general, swap spreads tend to stay above cash market spreads
except during sudden upward movements of cash market rates. Swap spreads can
be below cash spreads if the market participants assume the upward movement
in the cash spread is temporary. However, if the market thinks the upward
movement will be sustained in the long term, the swap spreads will also turn
up.

The relationships among swap spreads, interest rates, and interest rate
spreads can vary over time. Swap contracts to hedge basis risk can be for
varying terms to maturity. For example, a T-bill/LIBOR swap can be for 1, 2,
5 or more years, and costs can vary across terms of the contract. These
differences in movements of the various rates that affect lenders? decisions
on hedging were discussed in chapters 2 and 3. Figure 4.3 shows that actual
swap spreads have, at times, been above and

10 The cash or spot market spread is the difference between two interest
rates on simple contracts that are settled immediately but may extend for
years, e.g., the spread between the T-bill and CP rates or between the
10-year Treasury and a corporate bond rate. In contrast, a basis swap spread
would be the add-on to the T-bill rate in the future in order to receive
another rate such as LIBOR on a contract that could last 10 years. As LIBOR
and T-bill rates vary in the future, so will realized spread between the
T-bill based payment and the payment based on the other rate.

Chapter 4 86 below the cash T-bill/LIBOR (TED) spread 11 for a 5-year basis
swap contract. (Similar figures could be generated for other swaps contract
terms.) However, after the Russia Default in 1998, swap spreads have
increased and until late 1999 stayed above the TED spread, increasing
hedging costs.

11 The TED spread is the rate on Treasury issues minus the rate on LIBOR
deposits for the same term to maturity.

Chapter 4 87 Source: PaineWebber

Figure 4.3: Movementsin the Cash T-bill/LIBOR(TED) Spread And 5 year
SwapSpread

1995- 1999

0.000 0.200

0.400 0.600

0.800 1.000

1.200 1.400

1.600 1.800

10/11/1995 12/11/1995

02/11/1996 04/11/1996

06/11/1996 08/11/1996

10/11/1996 12/11/1996

02/11/1997 04/11/1997

06/11/1997 08/11/1997

10/11/1997 12/11/1997

02/11/1998 04/11/1998

06/11/1998 08/11/1998

10/11/1998 12/11/1998

02/11/1999 04/11/1999

06/11/1999 08/11/1999

per cent

LIBOR/T-bill (TED) Spread SwapSpread

Chapter 4 88

Hedging Is Complicated Due To Prepayment Risk on Student Loans

Officials representing lenders, investment banks, and credit rating agencies
emphasized that hedging with swaps for a portfolio of student loans is
complicated by the fact that the remaining balance on a portfolio is not
retired on a fixed schedule, such as a normal set of bullet bonds, which pay
interest during their life and repay the full principal in the last payment.
In contrast, for student loans, payments do not occur on a fixed schedule.
12 If the student borrower returns to school, repayment can be deferred, or
the borrower can prepay the loan as it seasons and as the borrower?s income
increases.

The amortization problem is most severe for hedges related to pools of
student loans that underlie student loan asset backed securities because, in
that case, the hedge applies to a specific group of assets and prepayment
could require unwinding if prepayment assumptions built into the hedge prove
wrong. The problem is less severe for ongoing concerns with a portfolio of
student loans that rolls over because, in that case, prepayments may simply
be reinvested in new loans. Nevertheless, uneven cash flows create some
complications in risk management for all firms.

Swaps can be made on a balance-protected or a balance-guaranteed basis.
Balance-guaranteed swaps recompute the remaining balance each period, based
on actual payments and prepayments. Balance-protected swaps include an
agreed- upon schedule in advance of the deal as to how the remaining balance
will decline over the lifetime of the deal. Such protection increases the
cost of the hedge.

Diversified Lenders May Not Hedge Basis Risk in Student Lending

Some for-profit lenders fund themselves as stand-alone entities, but most we
talked to are funded through a central corporate treasury. When funded
through a central treasury, some student lending activity operations are
charged the same rate that is charged to other lines of business in the
diversified firm, while others are charged a rate adjusted for the risk,
term, and other characteristics of the student loan product. The rate
charged to student lending operations by the corporate treasury may not
reflect fully the near zero credit risk for student loans, since some
average cost of funding loans is charged to all lending units in the
corporation. In addition, if an average rate is charged to student lending
it may not reflect the basis risk inherent in student lending, since any
hedging of basis risk tends to be done on a corporate-wide level. Since
basis risk in different parts of a large diversified lender may differ in
different ways, corporate-wide hedging may not be able to hedge the basis
risk in student loans. In addition, the corporation, in general, may not
measure or manage basis risk.

12 Other loans, such as mortgages, car loans or credit card balance, also
introduce prepayment risks for lenders since they do not occur on a fixed
schedule.

Chapter 4 89

Hedging Instruments Are Not Always Available for Lenders Offering Student
Loans Indexed to T-bills

In practice, the LIBOR/T-bill basis swap market may be too small and
illiquid to permit investors or FFELP lenders to hedge basis risk at a
reasonable cost. The basis swap market may be too small and illiquid between
T-bills and commercial rates, especially for long-term contracts, because
there appear to be few counterparties for the swap desired by lenders. Few
institutions have liabilities that move sufficiently closely with treasury
rates to make them natural counterparties. For example, everyone knows when
Sallie Mae, an active user of swaps to hedge financings not issued directly
in T- bill indexed securities, enters the market, and this entry affects the
swap rate. Most FFELP lenders, as do most other lenders, fund in commercial
rates, such as LIBOR or commercial paper.

During the summer of 1999, industry members mentioned that the spot
difference between the T-bill and LIBOR had come back down since late 1998
but that swap spreads remained relatively high as the market remembered what
happened in the fall of 1998. One investment firm official said that the
T-bill swap with LIBOR historically was liquid in the 1-year to 10-year
term, although other lenders, noted above, did not agree. However, since the
October 1998 flight to quality, the market for this swap has been very
illiquid and ?gappy,? which has increased the costs of swapping out of
T-bills into LIBOR. Since the fall of 1998 swaps had been difficult to
arrange, especially at the volume needed to offset student loan holdings.
Such problems have happened periodically, we were told, and are apt to
happen when lenders most need the swaps to manage basis risk.

Lenders emphasized that the swap market had only a limited capacity to serve
their swap needs. Any price quote one observed for a swap, such as those
published on Bloomberg, 13 are generally for a $25 million transaction.
There is no guarantee one could get a posted rate for the size necessary to
hedge a portfolio. How much one could hedge at one time depends on market
conditions, but market participants suggested that it becomes problematic
after $150 million and probably impossible beyond $400-$450 million even in
the best of times. These limited levels can be contrasted with the
approximately $22 billion of new FFELP originations last year.

One swap dealer reported that the amount it would hedge on a 5-year
LIBOR/T-bill swap declined sharply during 1998. It offered to swap up to
$250 million through August of 1998. During the rest of 1998, this swap
dealer would only offer contracts up to $100 million. In the beginning of
the year, the bid-ask spread for this swap was around 10 BP; by the middle
of the year it had dropped to 6 BP; and by the end of the year, the bid-ask
spread had increased to 20 BP. This was another indication of the

13 Bloomberg is a service that publishes information on prices in financial
markets.

Chapter 4 90 increasing illiquidity of the LIBOR/T-bill basis swap. Other
swap dealers and lenders also reported persistent illiquidity and higher
swap rates during this time period. 14

One lender told us their biggest business risk is long-term funding beyond 1
year. They prefer to use swaps to stabilize earnings and fund out to 2-3
years, but using swaps affects current income. For example, if the current
LIBOR spot rate is 40 BP above the T-bill and the lender can borrow at
LIBOR, his spread would be (T-bill + 280BP) - (T-bill +40 BP)) or 240 BP.
However, the lenders? spread, in the future, would vary with relative
movements in the TED spread. Alternatively, if the swap rate is 70 BP, the
lender could lock-in a spread of (T-bill + 280 BP) - (T-bill +70BP) or 210
BP. This rate is less than the current spread of 240 BP, but it will not
decline for the term of the swap contract. Thus, the lender would pay 30 BP
to lock-in a spread. The spread would be lower, but the basis risk would be
eliminated during the term of the swap.

If the long-term swap spread beyond 2-3 years increased, lenders might fund
short to avoid swap costs. According to several lenders, given the swap
expenses, fully swapping for the life of the long-term student loan is too
expensive given their earnings targets. Another agreed they cannot afford to
be fully swapped. Thus, they accept the basis risk implied by partial
swapping. Beyond a term of 2-3 years, basis swap premiums become
prohibitive.

ASSET BACKED SECURITIES ARE ESPECIALLY SENSITIVE TO CHANGES IN INTEREST
RATES ON INSTRUMENTS WITH LONGER TERMS TO MATURITY

As discussed previously, Sallie Mae and other large institutions have funded
student loans through asset-backed securities (ABS), which have varying
reference rates, markups, terms to maturity, or other conditions. However,
since the middle of 1997, their use has been declining due to market
conditions. In ABS lending, the debt is structured to match the expected
life of the assets in the portfolio. 1516 Inherent in such funding is the
use of some longer term certificates or bonds because student loans can last
up to 30 years. If the rates paid on the notes or longer term certificates
increase, the net interest margin earned by the pool decreases. In late
1998, asset- backed securitization for all assets including student loans,
ground to a halt as financing costs on long-term ABS securities proved too
expensive to permit profitable securitizations. In late 1998 and early 1999,
given the rates available in the capital markets, issuers were unable or
unwilling to create ABS and postponed issuing ABS.

14 Electronic screens used by market participants to track the availability
and price swaps do not report on the T-bill/commercial paper swap. 15
Although credit risk is nearly eliminated due to federal reinsurance of
guaranteed student loans,

other risks remain, such as servicing and prepayment. 16 This pass-through
reflects the bankruptcy remote character of the ABS trust, the common over-

collateralization of the trust, and the evaluation of servicing by credit
rating agencies. All these create some cost for the issuer.

Chapter 4 91 By mid-1999, securitization reemerged as illustrated in table
4.3. By late 1999, loan securitizations were undertaken again in larger
volumes.

Table 4.3: Quarterly Issuance of Student Loan Asset Backed Securities
(assets in $ billions)

1997 1998 1999 1 2 3 41 2341234 $ 2.48 4.41 3.69 3.87 4.19 4.17 00 1.47 1.93
2.39 1.20 4.06 source: PaineWebber.

The growth, sudden absence, and recovery of ABS in student lending was a
consequence of the increased leverage possible with ABS, which increases the
sensitivity of ABS profits and risks to changes in interest rates, as well
as the late- 1998 disruptions in the market for underwriting long-term
securities. The leverage in student loan ABS, as with all leverage, provided
both higher returns and higher risk - increased volatility of the return on
equity for lenders who create ABS. When the spread for securitizing narrowed
in late 1998, large securitizers kept loans on their books because the
return from securitizing had declined. They were waiting for an improved
return, decreasing hedging costs, and a more liquid swap market that would
let them hedge their risks. Swap or hedging costs have remained higher than
usual in 1999 and 2000. 17 However, securitizers have brought new
securitizations of student loans to market.

Securitization Does Not Eliminate Basis Risk If the Lender Does Not Fund
With T-bill Indexed Securities

Student lending securitization may decrease the need for capital and using
credit enhancements may allow for more favorable credit ratings that may
lower funding costs. As a result, it may become the preferred method of
financing for some lenders. However, basis risk remains. Several officials
told us that securitizing student loans does not eliminate the effect of
basis risk for the securitizer, because investors are also concerned about
basis risk in their own investments. If the securities issued to investors
by a student loan ABS are indexed to T-bills, the investor may face the
basis risk between T-bills and their own cost of funds or income
requirements. To ensure T-bill indexed securities are acceptable to
investors, the securitizer may have to pay a rate premium to the investors
and this premium reflects the investors? hedging costs to convert T-bill
indexed rates into commercial rates that may be more acceptable to
investors. Another option available to securitizers is to hedge the costs on
its own book of business and pay the investor an index based on commercial
rates. In either case, hedging costs will affect the lender?s returns
because such costs are borne or absorbed, to a large extent, by the lender.

17 For example, on a 7-year LIBOR/T-bill swap, the rate in mid-October 1999,
was 94 BP; in December 1999 the rate was 79-80 BP; and by April 2000, the
rate was 95 BP.

Chapter 4 92 Since 1995, Sallie Mae has tried to create a market for T-bill
indexed asset-backed securities, which would eliminate its basis risk, but
has had little success. When Sallie Mae has issued both T-bill and LIBOR
indexed securities in an ABS, investors have shown a preference for
LIBOR-based securities.

In a June 1999 securitization, Sallie Mae offered the choice of a
Treasury-bill-indexed security and a LIBOR-indexed security, with both
securities maturing in 2008. The margin over the T-bill rate was 87 BP, and
the margin over the LIBOR rate was 8 BP. The difference between the two
rates, which can be viewed as an implicit swap spread, was 79 basis points.
18 This swap spread is a bit larger than rate quotes obtained from the
Bloomberg system (Tullet and Tokyo screen) of 75 basis points for a
three-year swap on the day of the announcement, but is in the range of
market quotes (using the bid side of the market). It is consistent with the
calculated swap spread for a 10-year amortizing swap implied by an
equivalent hedge and computed from quoted interest rates on Treasury
securities and quoted prices for Eurodollar futures contracts around the
time of the announcement. (See appendix X for a discussion of the
calculation.)

On the day of the June 1999 Sallie Mae offering, the TED spread between the
91 day Treasury bill rate (in bond equivalent terms) and LIBOR was 47 basis
points. 19 What accounts for the 32 basis point difference between the swap
spread and the TED spread? Equivalently, why would Sallie Mae be willing to
pay (and investors require) 79 basis points over Treasury bills in order to
obtain LIBOR-indexed funding when the TED spread was only 47 basis points?
The fundamental reason is that the TED spread is relevant for 3-month
investments while the swap spread applies to a longer term investment.
Between June 1999 and the maturity of the loans, the TED spread will
fluctuate. The difference between the current TED spread and the swap spread
reflects two factors: expectations about future TED spreads and a risk
premium. One could interpret the latter as a premium that Sallie Mae was
willing to pay (and that investors required) in order to lock-in LIBOR
indexed payments (rather than Treasury bill indexed payments) and thereby
avoid the effects of those fluctuations. A kind of insurance premium is
therefore embedded in the 79 basis point swap spread. 20 21

18 Although the securities have a stated maturity of 9 years, they are
effectively much shorter term. The weighted average life of the loans was
2.5 years. 19 LIBOR is used on a 360-day basis, as quoted by the British
Bankers Association, because the rate

is specified in this fashion for the asset-backed security. 20 In theory,
not all risk premia need to be positive.

21 The interest rates on adjustable-rate mortgages (ARMs) and conventional,
fixed-rate mortgages differ for similar reasons. Even though today?s rates
on ARMs are below today?s rates on conventional mortgages, a homeowner might
prefer a fixed rate for two reasons. First, homeowners might expect that the
interest rates on ARMs will rise so that, over the long term, the
conventional mortgage is expected to be less costly. Second, homeowners
might prefer the greater certainty associated with a conventional mortgage
even if interest rates on ARMs are not expected to rise enough to make the
conventional mortgage less costly. That is, homeowners might be willing to
pay a type of insurance premium to avoid the risk of an unexpectedly large
increase in the ARM rate.

Chapter 4 93 Sallie Mae has noted that it only issued $165 million of the
shortest term tranche on a T-bill basis in the June 1999 $1.0 billion
transaction. Sallie Mae?s intent in pricing, and issuing the T-bill security
was to maintain a small presence in the T-bill ABS market and this presence
does not indicate that Sallie Mae is willing to fund at these levels using
T-bills. Sallie Mae has stated that it would be a mistake to try to infer
too much about the premium that would exist in the market based on the small
size and intent of the transaction.

A similar asset-backed security was issued by Sallie Mae in August 1999 and
had an implied swap spread of 86 basis points, the difference between
LIBOR-indexed and Treasury-bill-indexed portions, while the corresponding
TED spread was 60 basis points. The underlying loans had an expected average
maturity of just over 1 year. The increase in the implied swap spread
appears to be within the range of market movements in swap rates between
June and August. Sallie Mae noted that this was a small tranche and an
attempt to maintain a small presence in the T-bill ABS market and should not
be interpreted too broadly.

The swap spread included both the expected TED spread and insurance premium
to cover unexpected movements in the spread. How large was the insurance
premium in the swap spreads? Unless we know investors? (and Sallie Mae?s)
expectations of future TED spreads, we cannot say for sure. If the TED
spread was not expected to change, the ?insurance premium? at the time of
the June ABS issue was about 32 basis points, (79 - 47 or T-bill markup
minus LIBOR markup minus TED spread in June). On the other hand, if the TED
spread projected by the CBO of 64 basis points is taken as the relevant
expected TED spread, then the insurance premium, which covers risks and
profits for the counter party, is only 15 basis points (79 - 64). But
decomposing the difference into expectations of future TED spreads and an
insurance premium is not really necessary. The main points are that the swap
spread and not the TED spread provides the relevant comparison between
interest rates for a long-lived investment, and that it reflects the market
price for avoiding or taking on basis risk for the life of the swap.

Only Larger Lenders Can Securitize

Although securitization has been a successful business strategy for some
large and sophisticated financial institutions, it cannot be used by most of
the institutions that originate or hold student loans. According to
financial market participants, securitization requires a level of expertise
that many institutions do not possess. Several of the larger institutions
that currently participate in the student loan program told us that they
were evaluating or initiating some securitizations to determine whether that
was an appropriate business strategy for them. One common theme in their
comments was that securitization only made sense if one could continue to
undertake securitizations, since there are large fixed costs in undertaking
them. This implied that the lender needed a large volume of securitizations
so that the large fixed costs could be spread out. This, in turn, required a
large volume of originations or

Chapter 4 94 purchases of student loans. Several lenders, thinking about
securitizing, said they did their first one to ?test it out? or ?get their
feet wet.?

Some lenders told us that their portfolios were not large enough to make
securitization efficient, and that securitization did not take care of the
main problem that they faced- the mismatch between LIBOR and the T-bill
(basis risk). Lenders would have to offer either T-bill securities (which
investors would only take if they could find a reasonable swap) or LIBOR
securities (which the lender could not afford to offer unless they found a
reasonable swap themselves). An official at one non- securitizer said that
some lenders securitize if they have a capital constraint (the respondent
did not) or to reduce the volatility of cash earnings by locking-in
long-term financing. Securitization makes sense only if the firm can take
advantage of some type of market leverage. An official at another
non-securitizer emphasized that creating and servicing an asset-backed
security creates a servicing risk and that these risks persuaded it not to
securitize in the future.

COMPARING VOLATILITY OF SPREADS UNDER DIFFERENT FUNDING OPTIONS

Table 4.4 presents calculations of lenders? interest spreads between gross
formula returns and interest costs, where estimated margins over reference
rates are taken into account on the cost side and swap spreads are
incorporated to allow for hedging costs. These calculations are based on the
T-bill-indexed formula for lenders? returns that was in effect from mid-1998
through 1999 (no data on discounts, etc., were available). They are
presented for four calendar quarters - a base period in the second quarter
of 1998, the fourth quarter of 1998, which was the peak of flight-to-
quality and financial market disruption following the Russian loan default,
and the recovery quarters-the first and second quarters in 1999.

The table illustrates the fact that across a period of substantial financial
market volatility, FFELP lenders were adversely affected and that the extent
of the effect was related to funding options used. The volatility in 1998
was followed, in 1999, by continued greater financial uncertainty and
continued higher risk premiums. Both swap spreads and lenders? margins over
the reference rates at which they borrow recovered somewhat by the second
quarter of 1999 and then rose again in the third quarter, reflecting these
risk concerns as well as the beginning of some Y2K related distortions. Over
the four quarters as a whole, the effects of this volatility were least for
lenders who had access to T-bill funding, and hence did not face basis risk.
The effects of the financial volatility, as presented in the table, were
worst for those borrowing at LIBOR and fully hedged. Although the effects on
the interest spread for a lender borrowing at LIBOR and not hedging appear
to be less, such a lender could face worse outcomes in the uncertain future
depending on the evolution of spreads. Spreads for securitization were
adversely affected, apparently responding to the same forces that drove up
swap spreads and hedging costs.

Chapter 4 95 No FFELP lender exactly fits these paradigms, because most
lenders use more than one funding option and may be neither totally hedged
nor totally unhedged. In addition, FFELP lenders will have student loan
portfolios that roll over with mixes of loans with different lender return
formulas and financed and hedged at different times with different margins
in borrowing costs and different hedging costs. For simplicity, the table
uses the lender return formula for Stafford loans in repayment (not PLUS
loans or consolidation loans). The table also uses, for hedging, the
long-term hedge for an amortizing loan based on the model discussed in
appendix X. Hence, the focus of the table is on the conditions facing a
lender acquiring a Stafford loan in repayment and planning funding for its
life at the date (calendar quarter) indicated. It is hoped that this
analysis focuses on the ?marginal? conditions affecting lenders? decisions
and will shed some light on forces that would be at work if those conditions
were to persist.

Chapter 4 96 Table 4.4.a: Illustrative Interest Rate Spreads for FFELP
Lenders Using Different Funding Options For Selected Quarters in 1998-99

Funding Option Date LIBOR, no

hedge LIBOR,

hedged T-bill indexed ABS-LIBOR ABS-T-bill

1998, quarter II 2.01 to 2.26% 2.07 to 2.32% 2.05 to 2.35% 2.04% 2.07% 1998,
quarter IV 1.75 to 2.00 1.80 to 2.05 2.03 to 2.33 1.44 None done 1999,
quarter II 2.14 to 2.39 1.82 to 2.07 2.13 to 2.43 2.10 1.99 1999, quarter
III (preliminary)

1.89 to 2.14 1.68 to 1.93 1.99 to 2.29 1.79 1.86 Note: The spreads were
constructed from information in tables 4.4.b and4.4.c, below. The interest
expenses in each quarter, for each strategy (from table4.4.c) are subtracted
from the yield in each quarter (from table4.4.b), resulting in the spreads
in this table.

Table 4.4.b: Lender Gross Formula Yield Calculations For Selected Quarters
in 1998-99

Quarter Calculation Rate Treasury bill auction rate, last in May 1997 (5.17)
+ 2.30 7.47% Average Treasury bill rate in quarter (5.13) + 2.80 7.93 1998,
quarter II

Lender yield = maximum of two rates above 7.93 Treasury bill auction rate,
last in May 1998 (5.16) + 2.30 7.46 Average Treasury bill rate in quarter
(4.42) + 2.80 7.22 1998, quarter IV

Lender yield = maximum of two rates above 7.46 Treasury bill action rate,
last in May 1998 (5.16) + 2.30 7.46 Average Treasury bill rate in quarter
(4.6) + 2.80 7.40 1999, quarter II

Lender yield = maximum of two rates above 7.46 Treasury bill auction rate,
last in May 1999 (4.63) + 2.30 6.93 Average Treasury bill rate in quarter
(4.82) + 2.80 7.62 1999, quarter III

(preliminary) Lender yield = maximum of two rates above 7.62

Chapter 4 97 Table 4.4.c: Calculated Spreads and Lender Interest Costs For
Selected Quarters in 1998-99

Funding Option Quarter LIBOR, no hedge LIBOR, hedged T-bill-indexed
ABS-LIBOR ABS-T-bill

1998, quarter II

Yield from table 4.4.b 7.93% 7.93% 7.93% 7.93% 7.93%

Interest Cost Components

Reference rate LIBOR = 5.77 T-Bill = 5.13 T-Bill = 5.13 LIBOR =5.77 T-Bill =
5.13 Added margin - 0.10 to +0.15 - 0.10 to +0.15 0.45 to 0.75 0.12 0.73
Swap spread -- 0.58 -- * -- Total Interest Costs 5.67 to 5.92 5.61 to 5.86
5.58 to 5.88 5.89 5.86

Spread (yield - interest costs)

2.01 to 2.26 2.07 to 2.32 2.05 to 2.35 2.04 2.07 1998, quarter IV

Yield from table 4.4.b 7.46 7.46 7.46 7.46 7.46

Interest Cost Components

Reference rate LIBOR = 5.36 T-Bill = 4.40 T-Bill = 4.40 LIBOR =5.36 T-Bill =
4.40 Added margin 0.10 to 0.35 0.10 to 0.35 0.73 to 1.03 0.66 None done Swap
spread -- 0.91 -- * --

Total Interest Costs 5.46 to 5.71 5.41 to 5.66 5.13 to 5.43 6.02 None done

Spread (yield - interest costs)

1.75 to 2.00 1.80 to 2.05 2.03 to 2.33 1.44 None done 1999, quarter II

Yield from table 4.4.b 7.46 7.46 7.46 7.46 7.46

Interest Cost Components

Reference rate LIBOR = 5.12 T-Bill = 4.60 T-Bill = 4.60 LIBOR =5.12 T-Bill =
4.60 Added margin - 0.05 to +0.20 - 0.05 to +0.20 0.43 to 0.73 0.24 0.87
Swap spread -- 0.84 -- * --

Total Interest Costs 5.07 to 5.32 5.39 to 5.64 5.03 to 5.33 5.36 5.47

Spread (yield - interest costs)

2.14 to 2.39 1.82 to 2.07 2.13 to 2.43 2.10 1.99 1999, quarter III
(preliminary estimate)

Yield from table 4.4.b 7.62 7.62 7.62 7.62 7.62

Interest Costs Components

Reference rate LIBOR = 5.53 T-Bill = 4.82 T-Bill = 4.82 LIBOR =5.53 T-Bill =
4.82 Added margin - 0.05 to +0.20 - 0.05 to +0.20 0.51 to 0.81 0.30 0.94
Swap spread -- 0.92 -- * --

Total Interest Costs 5.48 to 5.73 5.69 to 5.94 5.33 to 5.63 5.83 5.76

Spread (yield - interest costs)

1.89 to 2.14 1.69 to 1.93 1.99 to 2.29 1.79 1.86

Chapter 4 98

Table 4.4 sources:

Treasury yield data come from the Federal Reserve H.15 release, as
transmitted by Haver Analytics data service. Auction rates are converted to
bond equivalent as described in Chapter 2.

LIBOR is the 3-month dollar LIBOR fixing rate, from the London Times, as
reported by Haver. For comparability with the Treasury-based returns, LIBOR
has been converted from a 360-day year to a 365-day basis. It is on the
360-day basis, rather than the bond equivalent 365 day because the 360- day
basis is used for determining payments on floating rate notes and
asset-backed securities (see, for example, page S-39 in the prospectus for
SLM Student Loan Trust 1999-2).

Lenders? added margin over LIBOR on their financings was derived
econometrically from 78 observations of Floating Rate Notes, issued by two
large banks in the student loan business. The notes were all linked to
3-month LIBOR and had an average maturity, in 1998 and 1999, of 2 years.
These results were likely to be representative for large, well-known banks.
In order to show a range also relevant to smaller institutions, 15 basis
points were added to the upper end of the range.

Lenders? added margin over the T-bill rate on their financings was the
average for floating rate notes linked to the 3-month T-bill rate and issued
by major government sponsored institutions. A 30 basis point higher rate is
shown, to allow for dispersion among GSEs and to suggest the range that may
be available to few private borrowers. This does not include the offset fee
of 30 basis points that is paid to the government by Sallie Mae on its GSE
holdings of FFELP loans.

Added margin for LIBOR-based ABS in 1998:Q2 is from the SMS issue in that
period. For 1998:Q4, it is taken from the Student Loan Funding issue,
adjusting the 1-month LIBOR spread to a 3-month spread using the difference
in these rates in the cash market. For 1999:Q2 and Q3, the LIBOR-based added
margins and, for all periods, the T-bill-based ABS added margins, are for
Sallie Mae issues. The 1999 LIBOR-based added margins are averages across
three or four tranches. For the shorter tranches, where T-bill based ABS
were also offered, the added margins over LIBOR were 8 basis points for both
the 1999:Q2 and 1999:Q3 issues.

The swap spread used to calculate the cost of funds if borrowed in LIBOR and
swapped to a Treasury basis is the cost of a 10-year amortizing hedge
estimated at Treasury (see appendix X for methodology); a shorter swap could
cost a different amount. Also, market quotes for swaps of specific
maturities in the relevant range vary somewhat; one source quoted spreads
for 5-year swaps about 10 basis points above levels shown, while another
source found them close to 7-year swap spreads. Actual spreads may depend on
size of transactions and whether the balance is guaranteed. Spreads
presented here are not for balance-guaranteed swaps.

The lenders? interest spread is the difference between the formula yield and
lenders? interest expenses.

Chapter 4 99 The degree of difficulty associated with the temporary
financial market disruption in the fourth quarter of 1998 did not interrupt
extension of loans to students. Furthermore, it does not appear in the
analysis presented in table 4.4 to have been sufficient, by itself, to cause
lasting difficulties for FFELP lenders. However, such a judgment ultimately
must depend, of course, on the capitalization and other business
characteristics of specific lender firms and agencies; analysis of such data
was beyond the mandate of this study. The degree of recovery by the third
quarter of 1999 (for which data were not entirely complete at the time of
drafting this report) is even more difficult to assess. Market conditions
had barely recovered from the Fall 1998 disruptions when they began to be
affected by Y2K considerations, illustrating the array of circumstances that
can make each year or quarter ?special.? In addition, judgment about the
adequacy of a lender interest spread depends on evaluation of other costs,
particularly servicing costs, and market requirements for returns on capital
necessary for a going concern. The Study Group did not address these issues.
It does appear, however, that the lower end of the range of interest spreads
for FFELP lenders using LIBOR funding, if sustained over time, could press
acceptable limits. Therefore, continued evaluation of lenders? circumstances
would have been required, once the millennium date had passed, to determine
whether lender interest spreads were in a range consistent with targets for
lender returns when the 1998 reauthorization was enacted. 22

In the second quarter of 1998, an unhedged LIBOR-based funding strategy
would have generated a 2.01 to 2.26% gross spread, while a fully hedged
LIBOR based strategy would have produced a 2.07 to 2.32% gross spread. These
spreads were similar, as swap costs approximately offset the favorable
LIBOR/T-bill spread. (See table 4.2 for an analysis of the interaction
between hedging costs and interest spreads.) The spread for T-bill based
funding was 2.05 to 2.35% and similar to the LIBOR based spread, but only
some FFELP lenders can obtain T-bill indexed funding. ABS based funding
based on LIBOR produced a spread of 2.04%, while ABS based funding based on
T-bills produced 2.07%.

In the fourth quarter of 1998, T-bill rates fell relative to commercial
rates due to the flight to quality, and lender margins on their costs of
funds rose as the market disruption heightened concerns about risk.
Consequently, LIBOR-based financing became relatively more expensive - see
the first and fourth columns of the table - and lenders? net interest spread
fell. For example, the spread with unhedged LIBOR financing fell by about
0.26 percentage point from the second quarter to the fourth quarter of 1998.
In the second quarter, the range had been 2.01 to 2.26%, and by the fourth
quarter it was 1.75% to 2.00%. Meanwhile, nearly half the decline in T-bill
rates was offset by an increase in swap spreads between the second and
fourth quarters. As a result, a lender who financed with LIBOR but swapped
fully into a T-

22 In October 1999, Congress changed the index for lenders? yield to
commercial paper with a modified spread over the reference rate. This study
does not evaluate and compare the new yield formula and spread with the
T-bill based yield formula adopted in the 1998 reauthorization.

Chapter 4 100 bill liability was also faced with less decline in costs than
in gross yield and a narrowing of interest spread on new financing, by 0.27
percentage point (see second columns of tables). 23 In the second quarter,
the range on hedged LIBOR funding was 2.07% to 2.32% and by the fourth
quarter it had fallen to 1.80% to 2.05%. LIBOR - based ABS spreads fell
sharply from 2.04% in the second quarter to 1.44% in the fourth quarter) as
the greater risk sensitivity impacted the market and T-bill based ABS simply
became unavailable at any remotely plausible rate as corporate underwriting
dried up.

By the second quarter of 1999, spreads had improved. Spreads on unhedged
LIBOR-based funding had increased a range of 2.14 to 2.39%, which were
higher than the spreads in the second quarter of 1998. Hedged LIBOR-based
funding spreads had increased a range of 1.82 to 2.07%, but were still below
the spreads in the second quarter of 1998. The smaller improvement in the
hedged LIBOR based funding was associated with a continued elevation of the
swap spread. Lenders? margins in borrowing rate also had not quite fallen to
the second quarter 1998 levels. The swap spread in the second quarter of
1998 had been 0.58%, in the last quarter of 1998 the swap spread was 0.91%,
and by the second quarter of 1999 the swap spread was still 0.84%. In
addition, lenders margins were 0.05% higher. Lenders issued ABS funding
indexed to LIBOR and T-bills in the second quarter of 1999. The LIBOR based
ABS spread had increased to 2.10%, slightly better than the spread in the
second quarter of 1998. T-bill indexed ABS funding reappeared with a spread
of 1.99%, slightly smaller than the spread in the second quarter of 1998.

The third quarter of 1999 reflected ongoing changes in interest rates and
swap spreads. Unhedged LIBOR-funding spreads were 1.89 to 2.14%, lower than
the spread in the second quarters of either 1998 or 1999 but above the
fourth quarter of 1998. The change in the spread in part reflected the
resetting of the student rate since the floor on the gross lender yield was
no longer relevant, while the TED spread had risen again. Hedged LIBOR-based
funding spreads were 1.68 to 1.93%, the lowest of the four quarters
estimated, and still reflected unsettled financial conditions and elevated
swap spreads and hedging costs. T-bill-indexed funding spreads were 1.99 to
2.29%, also the lowest of the four quarters. ABS-based funding spreads fell
relative to second quarter spreads, which reduced the financial
attractiveness of ABS based funding. ABS funding spreads often move with
hedged funding spreads since both must reflect the costs of hedging to
either the lender or the investor in the ABS.

It should be noted that the funding comparisons presented here are only
examples of the more complex funding that may occur at many institutions.
The spreads and margins only apply to the financing of new loans in
repayment and do not represent any particular lender or group of lenders.
Different lenders likely experienced different changes in the spreads and
margins during the financial crisis in the fall of 1998 and ensuing
quarters. Movements in economic conditions that affect the spreads and
margins may be either transitory or long-term, and the consequences for

23 Of course, lenders whose funding had previously been fully hedged would
have benefited to the extent that the floor kept the gross formula yield
from falling as much as current market T-bill rates. This benefit
disappeared, of course, in 1999, when the student rate reset.

Chapter 4 101 student lending of any changes in conditions depend on the
level and persistence of such movements.

According to lenders and officials in other financial institutions If
adverse conditions persist, lenders would reconsider participation in the
FFELP program. Under such conditions, specific lenders could (1) continue to
operate without hedging risk; (2) remain in the market and continue hedging
to limit risk, while earning a lower return due to increase hedging costs;
(3) invest less in student lending; or (4) exit the market. In general,
lenders suggest that a continuation of lower spreads and increased risk and
hedging costs could decrease the incentives of other financial institutions
to enter student lending. If many firms exited student lending and few firms
entered student lending, there might be insufficient capital to sustain a
healthy and competitive FFELP program in the future.

On the other hand, the agility with which much of the FFELP industry
weathered the financial storms of 1998-99 impressed several nonfinancial
Study Group members. Signs of healthy innovation in the industry suggested
to some Study Group members that the industry is basically in good shape and
is able to cope with slightly higher persisting risk premiums in its
financing costs - in swap spreads and lender borrowing margins - and with
the occasional volatile episode. There was no overall consensus within the
Study Group on how to weigh these considerations.

102

Chapter 5: Summary and Analyses of Study Issues

When the Higher Education Act was reauthorized in 1998, the yield formula
was changed. Specifically the markup over the 91-day T-bill rate was
lowered. 1 The law left in place a shift to a long-term Treasury reference
rate in 2003. In 1999, Congress changed the reference rate for the lender
yield to commercial paper from January 2000 through June 2003. However, the
Congress did not modify the requirement to shift to a long-term Treasury
reference rate in 2003. 2

During the discussion of the 1998 change, lenders and others market
participants were concerned about the future of the T-bill in light of
projections of future surpluses in the federal budget. They believed such
surpluses could undermine the the use of the T-bill as an effective
benchmark for student loans. In addition, they were concerned with the basis
risk imposed on lenders who borrowed at commercial rates but received yields
on student loans reflecting Treasury bill rates.

FFELP financial community members of the Study Group 3 (lenders, state
designated secondary markets, Sallie Mae, guaranty agencies, credit rating
agencies, and investment banks) believed that a T-bill based reference rate
with lower markups introduced a level of basis risk that endangered their
ongoing participation in the FFELP program. In response to these lenders?
concerns, Congress asked the Comptroller General and the Secretary of
Education to analyze how rates, rate spreads, and liquidity have behaved in
the past and how they might behave in the future.

In accordance with section 802 of the 1998 reauthorization of student
lending, this study evaluates the 91-day Treasury bill, 30-day and 90-day CP
rate, and 90-day LIBOR in terms of the following:

1. The historical liquidity of the market for each, and a historical
comparison of the spread between (1) the 30-day and 90-day commercial paper
rate, respectively, and the 91-day Treasury bill rate; and (2) the spread
between the LIBOR and the 91-day Treasury bill rate. 2. The historical
volatility of the rates and projections of future volatility.

1 The 1993 reauthorization legislation had set a long-term Treasury rate as
the reference rate for lender yield effective in 1998, but the 1998
reauthorization mandated the continued used of the 91-day Treasury bill rate
as the reference rate. 2 Because long-term rates are usually higher than
short-term rates, a change in the reference

rate to a long-term instrument would require a lower markup to create the
same lender yield. Consistent with the study mandate from the Congress, the
Study Group did not evaluate the implication of changing to a long-term
Treasury reference rate. 3 Appendix XIV states their concerns and lists the
members of the Study Group who participated

in writing the appendix.

Chapter 5 103 3. Recent changes in the liquidity of the market for each such
instrument in a balanced

federal budget environment and a low-interest rate environment, and
projections of future liquidity assuming the federal budget remains in
balance. 4. The cost or savings to lenders with small, medium, and large
student loan portfolios

of basing lender yield on either the 30-day or 90-day commercial paper rate
or the LIBOR while continuing to base the borrower rate on the 91-day
Treasury bill, and the effect of such change on the diversity of lenders
participating in the program. 5. The cost or savings to the federal
government of basing lender yield on either the

30-day or 90-day commercial paper rate or the LIBOR while continuing to base
the borrower rate on the 91-day Treasury bill. 6. Any possible risks or
benefits to the student loan programs under the Higher

Education Act of 1965 and to student borrowers. This chapter summarizes the
study group?s analyses of these issues. Summaries of the first three issues
are drawn from analyses in chapters 2, 3, and 4 of this report. Analyses of
the last three issues are developed in this chapter. Briefly, changes in the
financial instrument used to set the reference rate along with changes in
the markup, would have different effects on lenders, the government, and
students and other FFELP participants, based on their ability to deal with
the consequences of interest rate changes.

Chapter 5 104

Issue 1: The Historical Liquidity of the Market for Each Instrument, and a
Historical Comparison of the Spreads

The analysis in chapter 3 demonstrated that all of the cash market
instruments-- the 91-day T-bill, 30- and 90-day CP, and LIBOR--are liquid.
Based on direct measures of liquidity (such as bid-ask spreads, yield
volatility surrounding large trades, and trading volume); inferential
measures (such as issue volume and outstanding amounts; and secondary market
characteristics (such as the number of market markers, transparency, and
homogeneity), T-bills appear to be the most liquid of the instruments
historically.

Spreads between T-bills and the commercial rates have fluctuated over time,
as seen in chapter 2, but they have generally narrowed and become less
volatile in the 1990s compared to earlier years. In the 1970s, for example,
the TED spread (3-month LIBOR - 3-month T-bill) was as high as 550 BP, and
in the early 1980s, it surpassed 300 BP on three different occasions. By
contrast, in the 1990s, the spread remained below 100 BP between the end of
the Gulf War in 1991 and the international financial crisis in the fall of
1998, and even in 1998 it remained above 100 BP for less than 30 days.

LIBOR and CP tend to move more closely with one another than either one does
with T-bills. However, they have tended to move increasingly more closely
together as financial markets have become more integrated since the
mid-1980s. Historically, in times of international stress, CP usually moved
more closely with the T-bill than with LIBOR--that is, domestic rates tended
to move together. In the fall 1998 crisis, however, CP and LIBOR moved
together, and both diverged from the T-bill as a domestic and international
?flight to quality? drove investors to the safest instrument. The movement
to Treasury paper lowered interest rates on Treasuries compared to interest
rates on commercial financial instruments.

The mandate?s language asked for an analysis of the liquidity of the market
for each instrument and a comparison of the historical spreads between the
instruments. However, several members of the study group noted that the
liquidity of instruments whose returns are derived from the base instruments
rates also affect the lenders. For example, if a lender issued T-bill based
instruments to fund student lending, its cost of funds would be a T-bill
rate plus some margin. Many lenders believe such T-bill based funding
requires a large margin over T-bill rates in order to encourage investors to
buy T-bill based securities, since investors have a preference for
commercial rate based securities.

Chapter 5 105

Issue 2: The Historical Volatility of the Rates and Projections of Future
Volatility

Analysis in chapter 2 demonstrated that each of the rates showed volatility
in past decades, but each also showed a decline in volatility over the past
10 -15 years compared to previous periods. For example, the coefficient of
variation or the relative volatility of the T-bill was 0.35 from 1973 to
1984 and was 0.27 from 1985 to 1999. The average T-bill rate was 8.91% from
1973 to 1984 and was 5.69% from 1985 to 1999. One factor that might explain
the difference is that greater volatility in the T-bill rate tends to be
associated with a higher level of the rate, and rates have been relatively
low during the 1990s. Thus, relative rate volatilities have been decreasing.
Like T-bill rates, LIBOR and CP rates have shown lower relative volatility
during the 1990s compared to previous periods. The 1999 CBO and
administration forecasts show that future volatility of each rate is
expected to remain low relative to the 1970s or 1980s, just as the rates
themselves are expected to remain low.

Volatility of spreads between the various rates has declined in recent
years. For example, the coefficient of variation of the TED spread was 0.6
from 1973 to 1984 and declined to 0.51 from 1985 to 1999. In the same time
periods the average spread went form 1.64% to 0.67%. Also, the coefficient
of variation for the spread between 3-month commercial paper and T-bills
went from 1.33 to 0.59 while the average spread increased from 0.31% to
0.41%.

Chapter 2 of this report notes that the spreads between private sector rates
and the Treasury bill rate are larger and more volatile when inflation is
high and when cyclical disturbances induce active monetary policy. CBO
projections done for the study group assumed a noninflationary economic
expansion, fiscal restraint and a benign monetary policy. Under such
conditions both the spreads themselves and the volatility of the spreads
between T-bill and commercial rates will be lower.

However, FFELP members of the Study Group suggested that there would be an
increase in the volatility of lender returns if the index continued to be
based on T- bills in the future as federal budget surpluses continue to
grow. They contended that such increased volatility, in conjunction with
narrowing spreads on student loans (T-bill based yield minus commercial
rates) would decrease lender participation in the market. (See appendices
XIV for a discussion by the FFELP financial community.)

Chapter 5 106

Issue 3: Recent Changes in the Liquidity of the Market for Each Such
Instrument, and Projections of Future Liquidity in a Balanced Budget
Environment With Low Interest Rates

The Study Group was unable to reach consensus on the impacts of potential
future changes in the liquidity of T-bills. The FFELP community members
contended that T-bills would be less liquid in the future due to expected
future budget surpluses. The FFELP members believe such surpluses would
depress t-bill rates and the lender?s net yield (the difference between the
formula yield and the lenders? cost of funds) if T-bill rates were the
reference rate. They also believe that changes in the liquidity of T-bills
would increase the volatility of the spread between T-bill and commercial
rates. Such changes would decrease the incentive for lenders to participate
in the guaranteed student loan program. Over the foreseeable future,
Treasury officials believe that appropriate management of the Treasury
offerings of Treasury instruments in terms of frequency of offering and term
to maturity of the offered securities would maintain the liquidity of
3-month T-bills. In addition, Treasury officials believed that the
cumulative budget surpluses are more likely to affect the general level of
interest rates rather than affect the spreads between short-term bills and
other money market rates.

There was no consensus about the future liquidity of the Treasury market and
its ability to reflect underlying market conditions. Chapter 3 concludes
that it is difficult to determine what liquidity will be in the future for
any of the alternative indices. According to the Treasury, the advances in
information technology, electronic trading, and expanded trading hours would
seem to enhance liquidity in most of these markets. Furthermore, the
increased sophistication of financial market participants can be expected to
bring these markets closer to each other in characteristics because issuers
in each market are competing for largely similar investors. In contrast,
lenders noted that the past liquidity of the T-bill market may not represent
the future liquidity of the T-bill market. In addition, they questioned the
ability of the T-bill rate to function as a valid or useful reference rate
for lender yield in the FFELP. Lenders believe that in an era of budget
surpluses, the likelihood that a T-bill reference rate and markup will be
able to ensure a lender spread that can encourage program participation is
in doubt. (See appendix XIV for comments on the report and liquidity
concerns of representatives of FFELP lenders and other financial
institutions who were in the Study Group.)

Treasury agreed that concerns have arisen about a possible decrease in
liquidity of T-bills because of declining future issuance of Treasury paper
as the United States enters what is projected to be an era of budget
surpluses. As chapter 3 points out, while inferential measures of liquidity,
such as issue size, may show decreases, some of these inferential measures
have risen and fallen in the past decade with little effect on more direct
measures of liquidity, such as bid-ask

Chapter 5 107 spreads. In addition, T-bills, which are auctioned at a much
greater frequency than other Treasury securities, are an important cash
management tool, because the frequent auctions permit debt managers to
adjust financing in the face of dynamic and volatile cash balances. As such,
Treasury is committed to a debt management policy that will maintain a deep
and liquid short-term T-bill market. 4 (See appendix XV for Treasury?s
statement on debt management policy.)

Lenders pointed out that recent Treasury announcements demonstrated that the
liquidity of the Treasury market was affected by the historic reductions in
public debt. The February 2000 Mid-Quarter Financing announcement reduced
the number of auctions of 52-week bills and mentioned the possible
elimination of that series. Treasury noted, however, that the reduction in
the number of 52- week bill auctions will permit increasing auction size and
liquidity of the 91-day Treasury bill market.

4 In addition, the Treasury is committed to preserving the liquidity of its
issues at other key benchmark maturities, which will permit continued
calculation of constant maturity yield curves with a range of representative
yields for private sector use.

Chapter 5 108

Issue 4: The Cost or Savings to Different-Sized Lenders of Basing Lender
Yield on an Alternative Instrument, and the Effect of Such Change on the
Diversity of Lenders

As described in chapters 1 and 4, FFELP lenders use a variety of different
business strategies regarding originating, holding, and selling or
securitizing student loans. The effects on lenders of a change in the
reference rate and markup would vary somewhat by the strategy that they use.
Changes in the yield formula will impact lenders differently according to
which funding option they are currently using. Under the
commercial-paper-based yield formula, lenders that borrow at commercial
rates will face reduced basis risk. If they have not hedged their basis risk
under the T-bill-based yield formula, the change to a commercial- paper
formula will reduce their risk and stabilize their interest margin. In
addition, lenders who hedged their basis risk under the T-bill-based yield
formula, will have a reduced need to bear the costs of hedging under the
commercial-paper- based formula.

Effects on Lenders Who Originate and Hold Through Repayment

Lenders who borrow to originate and hold loans face basis risk if their
interest income based on the lender?s yield differs from their interest
expenses based on commercial rates. Because their primary source of income
is the formula spread (the difference between the formula yield and their
cost of funds), lenders often prefer to earn income and pay interest based
on the same reference rate. In theory, lenders can use hedging strategies to
address their risks. But hedging increases interest rate expenses and lowers
expected profits even while it reduces the risk or volatility in those
profits. The choice of hedging, not hedging, or partially hedging depends on
management?s judgment about the consequences of the choice. Furthermore,
hedging costs often increase when markets become more volatile, and hedging
in such markets further decreases the spread. In other cases, in volatile
markets, counterparties needing to undertake hedging may not even be willing
to offer sufficient volume for large lenders trying to hedge.

Given the yields, interest expenses, and risks inherent in this basic
strategy of originating and holding, many lenders have explored or adopted
other strategies. Their ability to adopt other strategies, however, is often
tied to the basic size of their portfolio of student loans, their ability to
affect costs based on changing the size of the portfolio, and whether they
can adopt more sophisticated strategies.

Chapter 5 109

Effects on Lenders Who Originate, Hold Temporarily, and Sell

Many lenders with small portfolios do not hold loans on their balance sheets
permanently, since the costs of holding and managing the portfolio--such as
servicing costs (including the actual processing of the origination) or
hedging costs--are high when compared to the lenders? yield. Instead, many
small originators hold loans temporarily and subsequently sell them to
Sallie Mae, a state secondary market, or a large bank.

Lenders who originate and sell earn their profit based on the gain they can
book because of the sale. In some cases, they sell soon after origination
and, in other cases, they hold the loan while the student is in school.
Selling once the student graduates is common, since servicing costs, based
on a servicing contract, often increase substantially once the student
graduates.

The price obtained by the lenders selling loans depends on overall economic
conditions and the efficiency and prices of the lenders buying the loans. If
the fundamental economics of student lending become adverse for lenders who
buy the loans, the price received by the originator will decline and affect
profits and risks.

Effects on Lenders Who Buy and Hold Through Repayment

Lenders who buy loans face the same risk as other lenders. However, these
buyers are often large institutions that can take advantage of scale
economics in servicing and holding loans. Large for-profit institutions may
have access to better funding sources, such as internal funding by the
parent firm or funding from overseas. Larger for-profit institutions are
also more likely to be able to hedge risks by swaps or by diversification
across different components of the diversified firm. Some not-for-profit
institutions may have funding advantages based on access to tax exempt
funding and/or administrative savings since they do not need to create
asset-backed securities, with their attendant costs, to undertake
asset-backed borrowing.

If servicing and holding costs are lower for large buyers than they are for
small originators, competition among the buyers may lead them to offer
sellers higher prices as buyers compete with each other to acquire student
loan assets. However, large buyers may be able to add to their portfolios
more cheaply by originating than they can through buying. In particular,
this could be true if large buyers? origination costs are lower than the
origination costs of the sellers. Thus, the price offered to small
originators by large buyers will reflect their relative cost advantages in
originating, purchasing, servicing, and holding loans in portfolio.

Chapter 5 110

Effects on Lenders Who Securitize

Lenders who originate or buy can either hold on their balance sheet or
securitize. The choice between the two funding strategies reflects
management?s goals, constraints, and capacities. Both for-profit and
not-for-profit institutions face the constraints imposed by investors and
credit rating agencies, who will require higher interest rates and capital
levels if risks are higher. In general, lenders find asset-backed
securitizations more attractive to the extent they decrease risks and lower
the returns that must be paid to bond and stock holders. However, only the
largest institutions can adopt this approach, because the ability to
securitize depends on portfolio size and the flow of new loans.

Asset-backed securitizations for student loans started in the mid-1990s.
Asset backed securitization for student loans dried up in late 1998 when the
LIBOR and T-bill rates diverged. The market had reappeared by the middle of
1999, and volume continued to increase in the early part of 2000.

A securitization can offer investors a return based on T-bill or LIBOR. 5
Only Sallie Mae has tried to develop a market for T-bill-based securities,
and they have had little apparent success. With a T-bill-based security,
investors bear the interest rate risk if their funding costs or net income
needs diverge from the T-bill- based yield. Investors in T-bill-based
student loan asset-backed securities can bear the risk or swap out of it.
Historically, the investors have tended to require a relatively high return
on the T-bill-based securities due to basis risk bearing or hedging costs.
On the other hand, if the securitizer wants to issue LIBOR-based securities
to please investors, the securitizer must bear the interest rate risk within
the securitization or swap out of it. In either case, hedging costs will
affect the profits and risks for securitizers. FFELP community members noted
that, since the 1998 financial market disruptions, securitizers have found
it increasingly difficult to swap out of basis risk because the term of
affordable swaps has been decreasing.

In many senses, securitizations are just a special form of buying and
holding, with funding done via asset-backed securities. Consequently, any
analysis of how changes in the yield formula would affect lenders who buy
and hold would also apply to securitizers. However, securitizations are
unique since they usually need lower capital levels than other funding
sources of student loans due to the financial structure of the pools and the
securities issued. However, these lower levels of capital make them
especially sensitive to interest rates on funding or fund availability, as
happened in late 1998.

5 Some securitizations offer an auction rate, which resets periodically but
is not tied to a particular index. However, market analysts report that
auction rate notes have rates that track LIBOR.

Chapter 5 111

Effects on Different-Sized Lenders and the Diversity of Lenders

The preceding analysis--of the effect of a change in reference rate on
lenders using different strategies--also applies to the effects on
different-sized lenders, because different-sized lenders tend to use
different strategies. Most small lenders tend to originate loans and sell
them after origination, often when the borrower enters repayment. Those who
purchase loans tend to be larger lenders, and those who securitize are among
the very largest.

Small lenders often act as originators for large lenders who buy their
loans. For example, as of December 31, 1999, Sallie Mae had $1.0 billion in
advances to student loan lenders that actually originate loans, which they
later sell to Sallie Mae. The extent to which small lenders act as
originators for large lenders can also be illustrated by information on the
concentration of originating and holding. For example, SLMA and the
secondary markets jointly held 55.8% of outstanding loan balances at the end
of 1998. However, SLMA did not originate loans, while the secondary markets
originated about 12.1% of all loans in 1998. In contrast, banks originated
80.4% of all loans while they held 42.5% of loan balances at the end of
1998. In 1998, the top 50 originators were 67% of the market, while the top
50 holders were 88.2% of the market. Thus, holding is more concentrated than
originating in the student loan industry.

The future diversity among lenders does not depend only on the formula yield
and spreads. Some banks may offer student loans as a community service, or
to have access to new customers. Thus, profits created from providing new
services to students may warrant continued offering of student loans even if
the profit and risks on the student loan operation, in and of itself, do
not. Nonetheless, the rate at which firms may enter, remain in or exit from
student lending reflects expectations about profits and profit volatility on
student loans, which are only partly determined by the formula yield and
funding costs.

Diversity of lenders also depends on economies of scale. For example,
securitization only makes sense if the lender can deal with a large volume
of loans so that economies of scale can be realized and large fixed costs
can be spread out. In addition, according to financial market participants,
securitization requires a level of expertise that many institutions do not
posses. Consequently, small volume originators and holders may not be able
to enjoy the economic advantages created by securitizations.

A crucial component of costs, and thus of profits, is originating and
servicing costs, which often appear to vary by size and composition of the
portfolio (type of degree or of school attended by the student). Often,
servicing is contracted out by smaller lenders, while the largest lenders
service their own portfolios and provide servicing for smaller lenders. To
the extent spreads shrink or become more volatile, originating and servicing
costs may increasingly determine which lenders are willing to enter, remain
in, or exit from student lending. Thus,

Chapter 5 112 congressional decisions about yield formulae may indirectly
determine the market concentration or diversity among lenders.

However, congressional decisions that maintain higher yields, and encourage
smaller lenders to remain in the program, will also enhance the yield for
larger lenders who may not need the encouragement. Due to economies of
scale, such lenders may already earn a sufficient yield to keep them in
student lending. Recent trends in the concentration of originations and
holding are consistent with the economy of scale premise. Over time, fewer
firms are originating and holding student loans--originations and holding
are becoming more concentrated. This increasing concentration, especially in
holding, probably reflects, in part, the economies of scale in servicing,
holding, and securitizing loans.

Chapter 5 113

Issue 5: The Cost or Savings to the Federal Government of Basing Lender
Yield on an Alternative Instrument

Any change in the yield formula could impact budgetary costs, and thus the
effects of formula changes must be addressed in the federal budget process.
However, the Study Group was unable to reach consensus on the correct
measure of cost to the government. Two measures of cost were discussed:
budget-based, and economic- or risk-based costs. All members agreed that any
change in the reference rate and markup must be budget neutral, i.e., they
must not increase expected net outlays as determined by the Congressional
Budget Office. However, no consensus was reached on the appropriateness of
considering or measuring risk-based costs for the government created by any
change in the index formula.

As part of the late 1999 congressional deliberations on index changes, CBO
projected the budgetary costs of a change to a commercial paper index. CBO
used probabilistic scoring to evaluate changing the reference rate to the
3-month commercial paper rate and changing the markup to 1.74%, while a
student is in school, grace or deferment and to 2.34% while a student is in
repayment. Its analysis yielded a small budgetary savings. (See chapters 1
and 2 and appendix IV for further discussion of CBO?s approach.)

Some members of the Study Group, including executive branch representatives,
believe a change to a commercial rate for the lender index would transfer
basis risk to the government. The Executive Branch representatives believe
that such a transfer, without an appropriate adjustment of the markup, would
not let the government earn income for bearing the risk as would happen to a
private sector entity bearing the same risk. The executive branch members of
the study group believe that the government should not bear the risk or the
economic cost without earning the income for bearing such risk.

The executive branch believes that the costs of basis risk are represented
in financial markets by the basis-risk premiums built into lenders? costs
under T-bill indexing for their yields. The executive branch representatives
suggested that the private sector costs of hedging in the swap markets
illustrate these risk- premiums or costs borne by private lenders who deal
with basis risk under the T- bill-based lender yield. Executive branch
officials suggested that specific ways to measure these costs might include
swap spreads reported in the markets, the implicit hedging spread between
LIBOR and T-bill built in ABS, or the theoretical swap simulations developed
by and calculated at the Treasury. It was noted, by the executive branch,
that if a lender did not hedge (through swaps or some other mechanism), the
lender?s equity return would probably expect to earn an average higher
return to compensate for the volatility.

Chapter 5 114 FFELP community members on the Study Group disagreed with the
executive branch analysis. They pointed out that the swap markets that might
let lenders address basis risk were ?inefficient and flawed.? Such
inefficiencies were noted in surveys of lenders and during interviews of
financial institutions that participate in the program. Further, the FFELP
community members have emphasized that there were flaws in the swap market
and that this hedging market might not be available or could be too
expensive, under certain conditions, as happened in the fall of 1998. In
addition, the FFELP community members believed that government estimates of
basis risk based on their theoretical swap simulation model did not provide
an appropriate measure of value.

Portions of chapter 4 and appendices X and XII detail executive branch
concerns and analytical approaches toward economic or risk-based cost
calculations. Portions of chapter 4 and appendixes XI and XIV detail the
concerns of the FFELP community members and their explanation of the reasons
to switch to a commercial-rate-based index.

Chapter 5 115

Issue 6: Any Possible Risks or Benefits to the Student Loan Programs Under
the Higher Education Act of 1965 and to Student Borrowers

Since the borrower interest rate is not changed, there should be no impact
on borrowers in terms of the rates they pay. In addition, guarantors,
servicers, and schools should feel no direct impact. However, if changing
the yield formula leads to changes in lender use of different funding
options or business strategies, these non-lender participants in the
guaranteed student loan program may be affected indirectly. For example,
changes in the lender yield formula may affect different lenders differently
and thus indirectly create changes in market structures that could affect
service or investment by lenders and their relationships with other market
participants.

The risks or benefits to the students and the student loan program of any
change in the formula yield will depend on how any changes affect lender
participation. FFELP financial community members on the Study Group believed
that a formula change that leaves basis risk with the lenders might: (1)
decrease the availability of rate discounts for students by lenders, (2)
discourage new participants, (3) accelerate the loss of current
participants, and (4) discourage all participants from undertaking real
investment to create a functionally efficient program that could improve
services to schools and students. FFELP community members on the Study Group
believed that a formula change that lowers either the number of lender
participants or real investment in the program by lenders will not be
improved for students and schools. In general, higher and more stable lender
spreads are likely to encourage participation and real investment, while
lower and less stable spreads will discourage participation and investment.

Government officials noted that any change in the size of the spread and its
stability must also consider the costs to the government and the impact on
diversity among loan providers. Executive branch officials noted that
changes in the lender yield formula may affect different types of lenders
differently. It is not clear how each type of lender would ultimately
respond to the formula change as different business strategies picked by
each lender interact to determine whether continued participation makes
sense. Thus, the ultimate costs and risks of changes cannot be easily
predicted.

Non-government and non-FFELP community members of the Study Group proposed a
set of criteria for evaluating changes to the lenders? yield calculation.
Their criteria emphasized that:

budget neutrality should be maintained;

benefits of any change should be shared by all program participants
including the borrowers;

both the short-term and long-term aspects should be addressed;

lenders should have a fair rate of return in a competitive market;

Chapter 5 116

any change should be consistent with changes in volatility and liquidity in
the market; and finally

any change in the reference rate should not affect the second study on using
market mechanisms.

See appendix XIII for a copy of the full set of criteria and FFELP community
members? response to the criteria.

Appendix 1 117

APPENDIX 1: LEGISLATIVE MANDATE IN THE HIGHER EDUCATION ACT OF 1998 (P.L.
105-244)

SEC. 802, STUDY OF THE FEASIBILITY OF ALTERNATIVE FINANCIAL INSTRUMENTS FOR
DETERMINING LENDER YIELDS

(a) Study Required.--The Comptroller General and the Secretary of Education
shall convene a study group including the Secretary of the Treasury, the
Director of the Office of Management and Budget, the Director of the
Congressional Budget Office, representatives of entities making loans under
part B of title IV of the Higher Education Act of 1965, representatives of
other entities in the financial services community, representatives of other
participants in the student loan programs, and such other individuals as the
Comptroller General and the Secretary of Education may designate. The
Comptroller General and the Secretary of Education, in consultation with the
study group, shall evaluate the 91-day Treasury bill, 30-day and 90-day
commercial paper, and the 90-day London Interbank Offered Rate (in this
section referred to as ``LIBOR'') in terms of the following:

(1) The historical liquidity of the market for each, and a historical
comparison of the spread between (A) the 30-day and 90-day commercial paper
rate, respectively, and the 91-day Treasury bill rate, and (B) the spread
between the LIBOR and the 91-day Treasury bill rate. (2) The historical
volatility of the rates and projections of future volatility. (3) Recent
changes in the liquidity of the market for each such instrument in a
balanced Federal budget environment and a low-interest rate environment, and
projections of future liquidity assuming the Federal budget remains in
balance. (4) The cost or savings to lenders with small, medium, and large
student loan portfolios of basing lender yield on either the 30-day or
90-day commercial paper rate or the LIBOR while continuing to base the
borrower rate on the 91-day Treasury bill, and the effect of such change on
the diversity of lenders participating in the program. (5) The cost or
savings to the Federal Government of basing lender yield on either the
30-day or 90-day commercial paper rate or the LIBOR while continuing to base
the borrower rate on the 91-day Treasury bill. (6) Any possible risks or
benefits to the student loan programs under the Higher Education Act of 1965
and to student borrowers. (7) Any other areas the Comptroller General and
the Secretary of Education agree to include.

(b) Report Required.--Not later than 6 months after the date of enactment of
this Act, the Comptroller General and the Secretary shall submit a final
report regarding the findings of the study group to the Committee on
Education and the Workforce of the House of Representatives and the
Committee on Labor and Human Resources of the Senate.

Appendix 2 118

APPENDIX 2: GOVERNMENT AND NON-GOVERNMENT STUDY GROUP MEMBERS

The Comptroller General, the Secretary of Education, the Secretary of the
Treasury, the Director of the Office of Management and Budget and the
Director of the Congressional Budget Office were designated by statute as
members of the study group. Representing these agencies for the study group
were:

General Accounting Office: Thomas J. McCool, Managing Director, Financial
Markets and Community Investment

Department of Education: Donald Feuerstein, Special Assistant, Office of the
Deputy Secretary

Treasury Department: Robert Cumby, Deputy Assistant Secretary, Office of
Economic Policy

Office of Management and Budget: Lorenzo Rasetti, Program Examiner
Congressional Budget Office: Nabeel Alsalam, Principal Analyst Private
sector members of the group, as designated by the Comptroller General and
the Secretary of Education, were as follows:

Bill Beckmann President and CEO, Student Loan Corporation

Kathleen L. Cannon Senior Vice President, Bank of America

Rene R. Champagne Chairman, President, and CEO, ITT Educational Services,
Inc.

Jacqueline Daughtry-Miller Vice President, Independence Federal Savings Bank

Anthony P. Dolanski Executive Vice President: Systems and Finance, Sallie
Mae, Inc.

Ivan Frishberg Higher Education Project Director, U.S. Public Interest
Research Group

Richard D. George President and CEO, Great Lakes Higher Education
Corporation

Appendix 2 119 Prof. Jonathan Gruber MIT Department of Economics

Arthur M. Hauptman Independent Student Financial Aid Consultant

Michael H. Hershock President and CEO, Pennsylvania Higher Education
Assistance Agency

James C. Lintzenich President and CEO, USA Group

Claire J. Mezzanotte Senior Director: Structured Finance, Asset Backed
Securities, Fitch IBCA, Inc.

Barmak Nassirian American Association of Collegiate Registrars and
Admissions Officers

Chalmers Gail Norris Executive Director, Utah Higher Education Assistance
Authority

Richard H. Pierce President and CEO, Maine Education Services

Dr. Susan L. Pugh Director, Office of Student Financial Assistance, Indiana
University- Bloomington

Marilyn B. Quinn Executive Director, Delaware Higher Education Commission

Anthony Samu President, United States Student Association

Dr. Robert A. Scott President, Ramapo College of New Jersey

Paul S. Tone Senior Vice President: Industry & Government Relations, UNIPAC

Paul W. Wozniak Managing Director, PaineWebber Incorporated

Appendix 3 120

APPENDIX 3: A BASIC CHRONOLOGY OF BORROWER RATES; LENDER YIELDS; LOAN
PROGRAM ACTIVITY; AND CERTAIN QUARTERLY INTEREST RATES

Borrower Rates and Lender Yields

Effective date Borrower rate(s) Lender yield(s) 11/8/65 6% for Stafford; ED
subsidizes 5%

before repayment 3% in repayment based on need

Determined quarterly by committee; cap of 3% over borrower rate

10/31/68 7% for Stafford; repayment subsidy 1% and only in States with 6%
cap

10/1/77 91-day T-bill plus 3.5% rounded to nearest 1/8 th ; cap of 5% over
borrower rate

10/1/79 Cap eliminated 1/1/81 9% for Stafford and PLUS 1 Rounding
eliminated; halved for tax-exempt

funding with 9.5% floor 10/1/81 14% for PLUS 2 11/1/82 12% for PLUS 3
9/13/83 8% for Stafford 4 4/7/86 Weighted average rounded to

nearest whole % with 9% floor for consolidation

1/1/87 52-week T-bill plus 3.25% with 12% cap for PLUS 5 91-day T-bill plus
3.25% for Stafford

7/1/88 8% until 48 months in repayment, 10% afterwards for Stafford 6

10/1/92 91-day T-bill + 3.1% with 9% cap for Stafford; 52-week T-bill +
3.25% with 10% cap for PLUS

Lenders required to rebate excess interest on initial 8/10% loans

91-day T-bill + 3.1% for Stafford 1 Possible decrease to 8% upon finding by
Secretary that 91-day T-bill below 9% 2 Possible decrease to 12% upon
finding by Secretary that 52-week T-bill below 14% 3 Resulting from finding
by Secretary that 52-week T-bill below 12%. 4 Resulting from finding by
Secretary that 91-day T-bill below 9%. 5 Outstanding loans allowed to
convert; "Rule of 78s" calculation prohibited 6 Possible adjustment to
reflect differing rates on borrower's outstanding loans

Appendix 3 121 Effective date Borrower rate(s) Lender yield(s)

when they hit 10%; excess interest to be rebated on all subsequent loans
made at 8/10% or 8%, regardless of stage of repayment

10/1/93 Special treatment of tax-exempt funding repealed with grandfathering

7/1/94 8.25% cap for Stafford PLUS interest rate 52-week T-bill +

3.1% with cap of 9% Stafford variable rate applied to all

new loans, regardless of prior borrowing

Consolidation interest rate now determined by weighted average rounded
upward to nearest whole %; minimum interest rate eliminated

1/1/95 Stafford variable rate applied to old 8/10% loans

7/1/95 91-day T-bill + 2.5% in school, grace or deferment, + 3.1% on
repayment for Stafford

91-day T-bill + 2.5% in school, grace or deferment, + 3.1% on repayment for
Stafford

11/13/97 Stafford rates for consolidation 7/1/98 91-day T-bill + 1.7% in
school, grace

or deferment, + 2.3% on repayment for Stafford.

91-day T-bill + 3.1% for PLUS 91-day T-bill + 2.2% in school, grace or

deferment, + 2.8% on repayment for Stafford

10/1/98 Weighted average rounded up to nearest 1/8 % with 8.25% cap for
consolidation

1/1/2000 3-month CP + 1.74% in-school, grace or deferment; 3-month CP +
2.34% in repayment for Stafford.

3-month CP + 2.64% for PLUS Source: U.S. Department of Education

Appendix 3 122 New Loan Originations and Concentration, Selected Years

Fiscal year New FFELP

volume ($billion)

Share originated by top 10 originators (percent)

Share originated by top 50 originators (percent) 1988 10.2 26 53 1992 13.6
32 58 1995 20.8 49 77 1997 21.5 49 80 1998 22.4 49 79 Source: Department of
Education

Loan Holdings and Concentration, Selected Years Fiscal year end

Outstanding FFELP volume ($billion)

Share held by top 10 loanholders (percent)

Share held by top 50 loanholders (percent) 1988 45.1 42 61 1992 62.0 51 73
1995 92.9 59 85 1997 112.4 59 87 1998 121.7 59 88 Source: Department of
Education

New Student Loan ABS Issuances Year

New student loan ABS issuances ($ million) 1991 347 1992 0 1993 594 1994
3,578 1995 3,518 1996 9,502 1997 14,446 1998 9,830 1999 9,588 Source:
PaineWebber Incorporated. Includes all issuances by for-profit issuers and
floating-rates issuances by state or non-profit issuers. Does not include
certain other issuances, such as auction-rate securities, by state or
non-profit issuers.

Appendix 3 123

Quarterly Interest Rates for Relevant Series

3-month LIBOR 3-month CP Year Quarter

91-day T- Bill (B.e.y.) 360-day 365-day Rate B.E.Y. 1 8.04 8.40 8.52 7.97
8.25 2 8.03 8.47 8.59 8.09 8.37 3 7.74 8.17 8.28 7.77 8.04 1990

4 7.21 8.08 8.20 7.62 7.88 1 6.22 6.87 6.96 6.53 6.73 2 5.76 6.17 6.25 5.90
6.07 3 5.55 5.84 5.92 5.59 5.75 1991

4 4.66 5.05 5.12 4.82 4.95 1 4.02 4.25 4.31 4.07 4.17 2 3.78 4.04 4.10 3.83
3.92 3 3.14 3.42 3.46 3.23 3.30 1992

4 3.17 3.62 3.67 3.47 3.55 1 3.05 3.27 3.32 3.25 3.32 2 3.05 3.27 3.31 3.10
3.17 3 3.08 3.26 3.30 3.11 3.18 1993

4 3.14 3.42 3.47 3.20 3.27 1 3.34 3.57 3.62 3.41 3.49 2 4.15 4.47 4.53 4.27
4.38 3 4.63 4.97 5.04 4.75 4.87 1994

4 5.46 5.96 6.05 5.72 5.88 1 5.95 6.29 6.38 6.06 6.24 2 5.79 6.12 6.21 5.88
6.05 3 5.54 5.89 5.97 5.61 5.77 1995

4 5.43 5.86 5.94 5.57 5.73 1 5.08 5.40 5.48 5.14 5.28 2 5.17 5.52 5.59 5.31
5.46 3 5.26 5.59 5.67 5.37 5.52 1996

4 5.11 5.53 5.61 5.31 5.46 1 5.21 5.56 5.64 5.34 5.49 2 5.21 5.81 5.89 5.60
5.76 3 5.18 5.73 5.81 5.50 5.66 1997

4 5.24 5.84 5.92 5.63 5.79 1 5.19 5.66 5.74 5.46 5.61 2 5.13 5.69 5.77 5.49
5.64 3 4.97 5.62 5.70 5.44 5.59 1998

4 4.40 5.28 5.35 5.09 5.23 1 4.54 5.00 5.07 4.82 4.95 2 4.60 5.05 5.12 4.89
5.02 3 4.82 5.44 5.52 5.25 5.39 1999

4 5.22 6.14 6.23 5.90 6.07 2000 1 5.72 6.11 6.19 5.91 6.08 For notes on
conversions, see appendix 5.

Appendix 4 124

APPENDIX 4: CBO METHODOLOGY FOR BUDGET SCORING

Budget Scoring for the Federal Family Education Loan Program The
Congressional Budget Office (CBO) estimates costs of legislation that
affects federal spending and receipts and scores these costs against the CBO
budget baseline for spending and receipts in the absence of new legislation.
The baseline, or the base assumption against which changes to a program are
scored, is the projected cost of a program under current law. Because
student loans are classified as mandatory spending--an entitlement--changes
to terms of student loan programs are subject to pay-as-you-go provisions.
This means that changes must be cost-neutral--that is, a change that would
increase government costs must be accompanied by revenue increases or
offsetting spending cuts.

Costs for credit programs, such as student loans, are estimated under the
terms of the Federal Credit Reform Act of 1990. The budget records all the
costs and collections associated with a new loan on a present-value basis in
the year the loan is obligated. The costs of all changes affecting
outstanding loans are displayed in the year of enactment. Future years? cash
flows--both dollars flowing from the government and receipts accruing to the
government--are estimated over the life of the loans and discounted back to
the current year.

In estimating the expected federal costs of a program change, CBO uses a
model to simulate the variation in interest rates around the CBO's baseline
forecast. For example, the model provides probabilities of how often and by
how much the simulated rates might exceed the 8.25 percent interest rate cap
for borrowers. These probabilities are then used in CBO's model of the
student loan program to estimate changes in subsidy costs.

Example--1998 Reauthorization of the Higher Education Act (HEA) At the time
HEA reauthorization was being considered in 1998, the reference rate for
lender yield on FFELP loans was the 91-day T-bill. However, this was
scheduled to change to the rate on a 10-to-20-year Treasury bond on July 1,
1998. 1 The interest rate received by private lenders after that date would
be the interest rate on bonds of comparable maturity plus 1.0 percentage
point. 2 Borrowers would pay this same rate, but with a cap of 8.25 percent.
To the extent that the yield to lenders exceeds the rate paid by borrowers,
the federal government would pay lenders the difference, which is called a
special allowance. In addition, the federal government would pay the
interest for student borrowers with subsidized loans while they are in
school or in a period of grace

1 This change was part of legislation passed in 1993 that, among other
provisions, established the Direct Loan program. 2 The CBO baseline assumed
that the rate on bonds of comparable maturity is the 10-year bond

rate. The administration uses a blended rate of 10-year and 20-year
maturities.

Appendix 4 125 or deferment, as it does currently. The scoring of the 1998
reauthorization was computed relative to the cost of loans under the
then-current statutes.

When it was enacted in October 1998, the reauthorization act (P.L. 105-244)
set the rate paid by student borrowers (for loans disbursed during the
period October 1, 1998, until July 1, 2003) at the 91-day Treasury bill rate
plus 1.7 percentage points while the borrower was in school, grace, or
deferment and 2.3 percentage points when the borrower was in repayment.
Lenders received the 91-day Treasury bill rate plus 2.2 percentage points
while the borrower was in school, grace, or deferment and 2.8 percentage
points when the borrower was in repayment. The federal government paid
lenders the difference between these two rates, termed a special allowance
payment. Because the borrower rate was adjusted annually and the lender rate
quarterly, this difference was not always equal to the apparent 50 basis
point difference between the two markups over the Treasury bill. The cap of
8.25 percent on borrowers' rates was retained. The borrower interest rate
and lender yield are still scheduled to revert to the 10-to- 20-year
Treasury bond rate plus 1.0 percentage point in July 2003.

CBO estimated that the changes in borrower interest rates and lender yields
(from the 10-to-20-year Treasury bond rate that would have otherwise gone
into effect in July 1998 to the new formula) would increase federal costs
over the 1999-2003 period by about $3.3 billion relative to then-current
law. The increased cost was associated with the new special allowance
payment as well as the increased exposure of the federal government to
interest rate subsidies when rates rise sufficiently to cause the borrowers'
interest rates to be constrained by the statutory caps. Moreover, the 91-day
Treasury bill is a more volatile instrument than the 10-year bond rate.

Scoring of Proposals for Study Group Reauthorization extended the use of the
91-day T-bill as the reference rate, and the change to the 10-to-20 year
instrument was postponed until July 1, 2003. Thus, during 1999, changes for
proposed for the future were score differently based on the year the loan
was assumed to originate.

Loans made through June 30, 2003 were scored relative to the 91-day Treasury
bill, and loans made after July 1, 2003 were scored relative to the 10-to-20
year instrument.

In August 1999, CBO presented the Study Group with estimates of the cost of
changing to a formula based on either LIBOR or CP to determine lender yield.
The estimates were for the period beginning July 1, 2003, so the cost
estimates were made relative to the cost using the 10-to-20-year rate plus
1.0 percentage point, which will be in effect in those years under current
legislation. Also, the estimates included costs only for subsidized Stafford
loans. These are the

Appendix 4 126 largest component of FFELP, but results for unsubsidized
Stafford, consolidation, and parent loans might have been different.

The first estimate was simply the cost of retaining the 1998 HEA formula--a
lender yield of 91-day T-bill plus 2.8 percentage points for borrowers in
repayment. This entailed some additional budgetary costs relative to the
10-to- 20-year rate plus 1.0 percentage point. The other 4 estimates--based
on indexes of 1-month CP, 3-month CP, 1-month LIBOR, and 3-month LIBOR--set
the markup over each index rate so that the lender yield would be held
constant with the T-bill-based formula. Use of each of the four indexes
resulted in costs that were higher than the cost of retaining the
T-bill-based formula.

Of the four alternative indexes, the 3-month CP index had the lowest
additional cost. Although the estimates were for loans made after July 1,
2003, the magnitude of the results suggested the 3-month CP index would be
less expensive for any time period. Actual dollar costs would vary depending
on the time period chosen, because the estimates are sensitive to such
factors as loan volume, which tends to increase each year.

Change in Lender Yield on Student Loans due to H.R. 1180 In the fall of
1999, an amendment to H.R. 1180 3 changed the lender yield on new loans
issued between January 1, 2000, and July 1, 2003. Under this act, yields are
based on the 3-month commercial paper rate. For student loans, the yield
becomes the 3-month commercial paper rate plus 1.74 percentage points (while
the borrower is in school, grace, or deferment) or 2.34 percentage points
(while the borrower is repaying the loan). The Lender yield on parent and
consolidated loans are the 3-month commercial paper rate plus 2.64
percentage points. H.R. 1180 left the existing interest rate structure for
borrowers unchanged.

Under the then-current CBO forecast of interest rates, the yields set by
H.R. 1180 using the commercial paper rate were estimated to differ slightly
from the yields under then-current law, based on the 91-day Treasury bill
rate. CBO estimated that this change would have a negligible federal cost in
2000, but it would save $20 million over the 2001-2003 period. Over this
time period, approximately $80 billion in new loans will be issued by
private lenders.

3 Subsequently enacted as P.L. 106-170 on December 17, 1999.

Appendix 5 127

APPENDIX 5: DATA AND CALCULATIONS FOR THE VOLATILITY ANALYSIS IN CHAPTER 2

The data series used in chapter 2 (and appendix 5) were obtained from the
Haver Analytics data base using a monthly frequency. All interest rates were
converted to a bond (or coupon) equivalent basis (with a 365-day year).

Both Treasury bill rates and commercial paper are quoted on a bank discount
basis, requiring adjustment both for the assumed days of compounding in the
year and for the change from par (100) to discount price in the calculation
of the rate. The formula is

Bond equivalent rate = [365 x (discount rate/100)] / [360-(91 x (discount
rate/100))]

The three-month Treasury bill rate is the average of rates at the regular
weekly auctions that occurred during the month, converted as just noted.

The commercial paper rates, for one-month and three-month maturities, are
for AA-rated financial commercial paper, converted as noted above. These are
rates released by the Federal Reserve Board, based on data they receive from
the Depository Trust Corporation.

The three-month London Interbank Offer Rate is the rate on three-month
dollar interbank placements determined at the daily fixing by the British
Bankers? Association. The rate is converted from a 360-day basis to a
365-day basis.

Appendix 6 128

APPENDIX 6: DERIVATION OF COMMERCIAL PAPER INTEREST RATES (FROM FEDERAL
RESERVE BOARD OF GOVERNORS)

The following material comes from the Federal Reserve Board?s Board of
Governors web site:

Commercial paper consists of short-term, unsecured promissary notes issued
primarily by corporations. Maturities range up to 270 days but average about
30 days. Many companies use commercial paper to raise cash needed for
current transactions, and many find it to be a lower-cost alternative to
bank loans.

The Federal Reserve Board's information on commercial paper (CP) is derived
from data supplied by The Depository Trust Company (DTC), a national
clearinghouse for the settlement of securities trades and a custodian for
securities. DTC performs these functions for almost all activity in the
domestic CP market.

Data on rates for CP are updated daily with a one-day lag. Data on CP
outstanding are available as of the close of business each Wednesday and as
of the last business day of the month; these data are also posted with a
one-day lag.

The Federal Reserve Board disseminates its information on commercial paper
primarily through its World Wide Web site. In addition, the Board publishes
one-, two-, and three-month rates on AA nonfinancial and AA financial CP
weekly in its H.15 Statistical Release and monthly in its G.13 Statistical
Release. It also publishes some data on CP outstanding in the monthly
Federal Reserve Bulletin.

To calculate CP interest rate indexes, the Federal Reserve Board uses DTC's
data for certain trades to estimate a relation between interest rates on the
traded securities and their maturities. In this calculation, the trades
represent sales of CP by dealers or direct issuers to investors (that is,
the offer side) and are weighted according to the face value of the CP so
that larger trades have a greater effect on the resulting index. With the
relation between interest rates and maturities established, the reported
interest rates represent the estimated interest rates for the specified
maturities.

Interest rates calculated through the process described above are a
statistical aggregation of numerous data reflecting many trades for
different issuers, maturities, and so forth. Accordingly, the reported
interest rates purport to reflect activity in certain segments of the
market, but they may not equal interest rates for any specific trade. As
with other statistical processes, this one is designed to minimize the
difference between the interest rates at which actual trades occur and the
estimated interest rates.

Appendix 6 129 CP trades included in the calculation are chosen according to
the specifications listed in table 7.1 below. Data to assess CP trades
relative to these criteria are updated daily from numerous publicly
available sources. SIC code classifications are taken from the SEC Directory
of Companies Required to File Annual Reports with the Securities and
Exchange Commission. When an issuer's primary SIC code is not reported in
the SEC directory, the primary SIC code reported in the issuer's financial
reports is used; otherwise, SIC codes are determined upon consultation with
the Office of Management and Budget's Standard Industrial Classification
Manual or its Supplement.

For a discussion of econometric techniques for fitting the term structure of
interest rates, including bibliographic information, see, for example, Mark
Fisher, Douglas Nychka, and David Zervos, "Fitting the Term Structure of
Interest Rates with Smoothing Splines," Finance and Economics Discussion
Series 95-1 (Board of Governors of the Federal Reserve System, January
1995).

Appendix 6 130 Table 6.1: Criteria for Calculating CP Interest Rate Indexes
Item AA financial AA nonfinancial A2/P2 nonfinancial Short-term credit
rating

Programs with at least one "1" or "1+" rating but no ratings other than "1"

Programs with at least one "2" rating but no ratings other than "2"
Long-term credit rating

Programs with at least one "AA" rating, including split-rated issuers

Programs with at least one "A" or "BBB"/"Baa" rating, including split-rated
issuers, but none with any ratings outside the "A"- "BBB"/"Baa" range Credit
rating agencies considered

Duff & Phelps Credit Rating Co., Fitch Investors Service, Moody's Investors
Service, and Standard & Poor's

Credit rating reviews

Programs that would be included in an index calculation are excluded when
(1) the issuer's credit ratings are under review and (2) a one-notch or
downgrade would violate either credit rating criterion SEC registration
types

Both traditional programs (3(a)3) and private placements (4(2)) are included
Placement Both dealer-placed and directly placed programs are included
Industries included (primary SIC codes)

6000-6999, excluding 6189 (asset-backed CP) and 6200-6299 (security
broker/dealers)

100-5999, 7000-9999

100-5999, 7000-9999 Excluded trades Foreign and credit-enhanced programs;
secondary, repurchase

agreement/financing, and interest-at-maturity trades Weights Trades are
weighted by their face values

Source for entire appendix: http://www.bog.frb.fed.us/releases/CP/about.htm

Appendix 7 131

APPENDIX 7: THE BRITISH BANKERS ASSOCIATION (BBA) AND THE LIBOR FIXING

The BBA sets the most widely quoted LIBOR rate. Background on the BBA and
the methodology of the rate fixing is explained below. 1

Evolution of the British Bankers' Association The British Bankers'
Association was formed in 1919, but its current role and structure dates
from 1972 when membership was extended to the foreign banks in London. Until
then, membership had been restricted to British commercial banks in Great
Britain and the Commonwealth. The British accepting houses or merchant
banks, previously absent, also elected to join at this time.

These changes created a broadly based banking association, representative of
all banks in the UK and able, as such, to participate in the European
Bankers' Federation. Based in Brussels, the Federation is recognized by the
EC Commission, the Council and the European Parliament as the representative
body for the community's commercial banking sector.

The BBA is the trade association for the banking industry in the UK. Its
members are organizations authorized under the Banking Act by the Financial
Services Authority to take deposits from the public in the UK and to use a
banking name. The membership currently includes all major banks and numbers
some 330. Well over three-quarters of them are foreign owned or foreign
controlled although, in local asset terms, British-owned banks predominate.

As a trade association the BBA is the forum in which the banks in the UK
seek common ground to enable the banking industry as a whole to speak
collectively on matters of common interest and of public policy.

Being the voice of the banks in the UK the British Bankers' Association
communicates the industry's views to the British government, to the Bank of
England and City regulators, to the press, to opinion formers, to the
institutions of the EC and to governments and regulators around the world.

Since its formation the work of the BBA ran partly in parallel and shared a
secretariat jointly with the Committee of London Clearing Bankers (CLCB),
the Chairman of which was also the President of the BBA. The two
organizations diverged, however, in 1975. The BBA obtained its own Secretary
General and

1 This material comes from the web site of the British Bankers Association
(www.bankfacts.org.uk1).

Appendix 7 132 support staff. The office of President was no longer linked
to the Chairmanship of the CLCB.

Following the transfer of responsibility for the Bankers' Clearing House and
for other forms of money transmission matters from the CLCB to the newly
established Association for Payment Clearing Services (APACS) in 1985, the
dividing line between the work of the BBA and the CLCB (latterly re-named
the Committee of London and Scottish Bankers (CLSB)) had become blurred.

It was recognized that the BBA had become increasingly acknowledged as the
voice of the banking industry in the UK and that there was no longer any
justification for the continued existence of another body that spoke only
for a small number of banks. In 1991 it was decided that the CLSB should be
wound up and its work absorbed with that of the BBA.

The LIBOR System The British Bankers' Association (BBA) LIBOR is the primary
benchmark used by banks, securities houses and investors to fix the cost of
borrowing in the money, derivatives and capital markets around the world.

BBA LIBOR fixing evolved in the early 1980's with the growth of syndicated
lending and early developments in the derivatives markets. Since then it has
assumed an increasing importance as well over 20% of all international bank
lending and more than 30% of all FX transactions take place in London.

BBA LIBOR is now used to calculate the interest rates applying to a wide
range of contracts including OTC instruments such as swaps, loan agreements,
FRNs, FRAs and Exchange Traded Short Term Interest Rate contracts traded on
LIFFE, CME and DTB amongst others.

BBA LIBOR is fixed for the following currencies: GBP, CAD, NLG, XEU, USD,
AUD, ITL, YEN, DEM, PTE, CHF, FRF & ESP. All currencies are fixed on a spot
basis on each London Business Day apart from Sterling, which is fixed for
same day value.

LIBOR is provided as a free service to the market by the BBA. There is no
comprehensive list of all its users or uses, but it is generally
acknowledged as a truly international benchmark. BBA LIBOR is published
simultaneously on more than 300,000 screens throughout the world, being
distributed by, amongst others, the following major information vendors:
ADP, Datastream, Reuters, Bloomberg, Nomura Research, S&P Comstock, Bridge
Telerate, and Quick. Bridge Telerate manages the fixing process on behalf of
the BBA, collecting data from Contributor Panel Banks, applying quality
control tests to it and calculating the Fixing, releasing it just before
noon, London time.

Appendix 7 133 In the July 1998 The Banker survey of the top 1000 banks. 11
of the BBA LIBOR banks are in the top 20 world banks. Furthermore, 9 of the
BBA LIBOR banks are in the May 1998 Euromoney FX poll's top 10 indicating
that the euro BBA LIBOR Panel banks are amongst the most active in the world
in the wholesale interbank market.

The BBA LIBOR Fixing BBA LIBOR is the BBA fixing of the London Inter-bank
Offered Rate. It is based on offered inter-bank deposit rates contributed in
accordance with the Instructions to BBA LIBOR Contributor Banks.

The BBA will fix BBA LIBOR and its decision shall be final. The BBA consults
on the BBA LIBOR rate fixing process with the BBA LIBOR Steering Group. The
BBA LIBOR Steering Group comprises leading market practitioners active in
the inter-bank money markets in London.

BBA LIBOR is fixed on behalf of the BBA by the Designated Distributor and
the rates made available simultaneously via a number of different
information providers.

Contributor Panels shall comprise at least 8 Contributor Banks. Contributor
Panels will broadly reflect the balance of activity in the inter-bank
deposit market. Individual Contributor Banks are selected by the BBA's FX &
Money Markets Advisory Panel after private nomination and discussions with
the Steering Group, on the basis of reputation, scale of activity in the
London market and perceived expertise in the currency concerned, and giving
due consideration to credit standing.

The BBA, in consultation with the BBA LIBOR Steering Group, will review the
composition of the Contributor Panels at least annually.

Contributed rates will be ranked in order and only the middle two quartiles
averaged arithmetically. Such average rate will be the BBA LIBOR Fixing for
that particular currency, maturity and fixing date. Individual Contributor
Panel Bank rates will be released shortly after publication of the average
rate.

The BBA, in consultation with the BBA LIBOR Steering Group, will review the
BBA LIBOR Fixing process from time to time and may alter the calculation
methodology after due consideration and proper notification of the planned
changes.

In the event that it is not possible to conduct the BBA LIBOR Fixing in the
usual way, the BBA, in consultation with Contributor Banks, the BBA LIBOR
Steering

Appendix 7 134 Group and other market practitioners, will use its best
efforts to set a substitute rate. This will be the BBA LIBOR Fixing for the
currency, maturity and fixing date in question. Such substitute fixing will
be communicated to the market in a timely fashion.

If an individual Contributor Bank ceases to comply with the spirit of this
Definition or the Instructions to BBA LIBOR Contributor Banks, the BBA, in
consultation with the BBA LIBOR Steering Group, may issue a warning
requiring the Contributor Bank to remedy the situation or, at its sole
discretion, exclude the Bank from the Contributor Panel.

If an individual Contributor Bank ceases to qualify for Panel membership the
BBA, in consultation with the BBA LIBOR Steering Group, will select a
replacement as soon as possible and communicate the substitution to the
market in a timely fashion.

Instructions to BBA LIBOR Contributor Banks. An individual BBA LIBOR
Contributor Panel Bank will contribute the rate at which it could borrow
funds, were it to do so by asking for and then accepting inter-bank offers
in reasonable market size just prior to 1100.

Rates shall be contributed for currencies, maturities and fixing dates and
according to the quotation conventions.

Contributor Banks shall input their rate without reference to rates
contributed by other Contributor Banks.

Rates shall be for deposits: made in the London market in reasonable market
size; that are simple and unsecured; governed by the laws of England and
Wales; where the parties are subject to the jurisdiction of the courts of
England and Wales.

Maturity dates for the deposits shall be subject to the ISDA Modified
Following Business Day convention, which states that if the maturity date of
a deposit falls on a day that is not a Business Day the maturity date shall
be the first following day that is a Business Day, unless that day falls in
the next calendar month, in which case the maturity date will be the first
preceding day that is a Business Day.

Rates shall be contributed in decimal to at least two decimal places but no
more than five.

Contributors Banks will input their rates to the Designated Distributor
between 1100hrs and 1110hrs, London time.

Appendix 7 135 The Designated Distributor will endeavor to identify and
arrange for the correction of manifest errors in rates input by individual
Contributor Banks prior to 1130.

The Designated Distributor will publish the average rate and individual
Contributor Banks' rates at or around 1130hrs London time.

Remaining manifest errors may be corrected over the next 30 minutes. The
Designated Distributor then will make any necessary adjustments to the
average rate and publish it as the BBA LIBOR Fixing at 1200hr.

Appendix 8 136

APPENDIX 8: INTERVIEWS OF FFELP INDUSTRY REPRESENTATIVES

In order to learn more about trends in financing student loans made under
the FFELP, we interviewed representatives of lenders, secondary markets, and
investment firms (generally meaning firms involved in FFELP only indirectly
or as intermediaries, not lenders in the program) during the summer of 1999.
GAO staff conducted these, accompanied in some cases by staff from the other
government agencies that were part of the study group. This appendix
summarizes what the representatives told us and on what points they
disagreed. We also interviewed officials at several schools and trade groups
related to FFELP, but because their perspectives vary widely and we did not
conduct many such interviews, we use their comments for background
information rather than summarizing them here.

Overall Comments on Interest Rates Our interviews indicated that LIBOR is
the dominant index in most markets today. Many more instruments are traded
today based on LIBOR than on T-bills. Its use for student loans would result
in a more liquid market for any type of student loan transaction.

Student loan holders holding a T-bill-based instrument generally prefer to
swap to a LIBOR rate to match their funding costs, which are primarily
LIBOR-driven. However, some interviewees said that even before the October
1998 ?flight to quality,? swaps were difficult to arrange, especially at the
volume needed to offset student loan holdings. Several mentioned that the
spot difference between the T-bill and LIBOR has come back down since late
1998 but that swap rates have remained relatively high as ?the market
remembers what happened last fall.? One investment firm interviewee said
that the T-bill swap with LIBOR historically was liquid in the 1-year to
10-year term. It was one of several basis swaps available in the market.
However since October 1998, the market for this swap has been very illiquid
and ?gappy? as the flight to quality lowered T-bill rates and increased the
costs of swapping out of T-bills into LIBOR.

Lenders? Portfolio Strategies Several different strategies for holding and
hedging portfolios exist and each strategy can affect the risks undertaken
and returns earned by the lender.

Appendix 8 137 Hedging By Using Swaps Interviewees generally agreed that
there?s no natural counterparty for a T-bill- LIBOR swap (and no natural
investor for variable-rate T-bill-based assets). The market for T-bill-based
securities is thin because of the need to swap into LIBOR. The price that
would be required makes swaps unfeasible. Quoted swap rates can be higher or
lower than the Treasury-to-Eurodollar (TED) spread at any given time. The
TED spread is based on current conditions, but swap rates depend more on
anticipated future rates.

Any price quotes one observes for swaps, such as those published by
Bloomberg?s, are generally for a $25 million transaction. There?s no
guarantee you could get a posted rate for any particular large swap. How
much you could swap at one time depends on market conditions, but it becomes
problematic after $150 million and probably impossible beyond $400-$450
million in the best of times. One interviewee told us that if swaps were
widely available at the quoted rates, lenders would not be concerned about
the T-bill as an index.

T-bills are an unnatural index for lender yield because of special
characteristics of the instrument. For example, the supply is driven in part
by the Treasury?s borrowing needs. Some fear that a reduced supply of
T-bills because of future budget surpluses will drive T-bill prices up and
yields down--one specifically mentioned the fact that Treasury has begun
buying back securities in some maturity ranges. Also, when Sallie Mae, the
largest issuer of T-bill-indexed securities, enters the market, everyone
knows it, and this affects the swap price.

Swaps are complicated by the fact that the remaining balance on a pool of
student loans is not determined by a fixed schedule, such as a normal set of
bullet bonds. To deal with this variable amortization requires a swap that
includes adjustments of amortization of the student loan pool, introducing
another level of risk into the swapping arrangement. Swaps can be made on a
balance- guaranteed basis, balance-protected basis or on a fixed
amortization basis. The balance guaranteed approach involves recomputing the
notional balance each period, based on actual payments and prepayments. The
balance protected approach includes an agreed-upon schedule in advance of
the deal about how the notional amount will decline over the lifetime of the
deal.

One lender told us its biggest business risk is long term funding for 1
year. They prefer to use swaps and spreads to stabilize earnings and fund
out to 2-3 years via swaps. But using swaps affects current income. For
example, at the time of the interview, the current LIBOR spot rate was 40
b.p. above the T-bill, and the forward swap rate was 70 b.p. higher i.e.
they were paying 30 b.p. to lock in a spread. This means the lender fund
short although it would prefer to fund long. Given the high costs, fully
swapping to hedge basis risk is too expensive given the lenders? earnings
targets. Another agreed they cannot afford to be fully swapped. Thus they
accept the basis risk implied by partial swapping. Beyond a

Appendix 8 138 term of 2-3 years, basis swap premiums become prohibitive.
Another told us this has been especially true in recent years, compared to
1996, when lenders had better success getting swaps for the volume they
required. One lender that does not swap told us that it expected profits
were higher if they simply bear the basis risk.

Securitizing Asset-backed securitizations for student loans started in
mid-1990s; they have been used in mortgages and other loan industries, such
as credit cards, for much longer. The student loan securitization market
dried up in late 1998 when the LIBOR and T-bill rates diverged. New issues
have resumed in 1999.

If the creator of the securitization wants to issue LIBOR-based securities
to please investors, there will be a cost to swap out of variable T-bill
based rates. If the creator of the securitization doesn?t swap out of T-bill
based student loans, the final investor will bear the costs of swapping out
of T-bill based assets.

Two structures are possible for securitization trusts--a ?master (or
revolving) trust? and an ?amortizing trust.? Credit card securitizations
tend to be revolving trusts. At first, the trust might include $1 billion of
receivables. As credit card balances are paid down, new receivables are put
into the trust. In an amortizing trust, as balances are paid down, the trust
shrinks, because new loans are not put in to replace those paid off. With
student loan trusts, loan consolidations, deferments, and serialization of
loans (new loans for borrowers whose prior loans are already in the trust)
complicate matters. Another difficulty is really determining what is meant
by prepayment risk because it includes consolidations, voluntary
prepayments, guarantee payments due to default and extension risk (slower
than expected prepayments).

Only Sallie Mae has tried to develop a market for T-bill-based securities,
and they have had little success. With a T-bill-based security, the investor
bears the interest rate risk if funding costs diverge from the T-bill-based
return. Most other securitizers offer the investor LIBOR (or an auction
rate), which means the securitizer must bear that risk. The securitizer can
hedge either inside or outside of the securitization.

When a securitization is put together, credit rating agencies and investment
firms negotiate the credit rating, and lawyers ensure that a
bankruptcy-remote special purpose vehicle is created to ensure that the
rating is based on the quality of the pool and not the financial strength of
the creator of the pool.

Municipal agencies, in some sense, do something similar to a securitization
when they offer their bond issues, although bond repayments may or may not
be tied to repayments from a specific cohort of student loans.

Appendix 8 139 One nonsecuritizer said that lenders securitize if they have
a capital constraint (they don?t) or to reduce earnings volatility.
Securitization makes sense only if the firm can take advantage of some type
of market leverage. Another nonsecuritizer said it decided the big risk was
servicing when the trust department was the master servicer, which it would
have been in a securitization. A third said its portfolio is not large
enough to make securitization efficient, and in addition it does not take
care of the main problem it faces--the mismatch between LIBOR and the
T-bill. The lender would have to offer either T-bill securities (which
investors would only take if they could find a reasonable swap) or LIBOR
securities (which the lender could not afford to offer unless it found a
reasonable swap for itself).

Securitizations are most effective if done on a large scale and repeatedly.
Several who securitize said they did their first one to ?test it out? or
?get their feet wet.? Several also said they try to put all loans into
securitization, perhaps holding them on book while borrowers are in school
and then securitizing when loans go into repayment. (This is the same
strategy other lenders may use when selling loans on the secondary market.)

Other strategies--Lenders That Are Part of a Larger Holding Company Some
lenders fund themselves as stand-alone shops, but most we talked to are
funded through a central corporate treasury. When funded in this way, some
lender shops are charged some average rate that is the same as for other
lines of business within the holding company, while other holding companies
charge student lending shops a rate adjusted for the risk, term, and other
characteristics of the student loan product. Most are evaluated based on
earnings or net return (either return on assets or return on equity). This
is more pressure than state secondary market institutions face, although the
two types of institutions are both evaluated by credit ratings agencies when
they issue new securities to ensure their liabilities or asset-backed
securitizations are creditworthy. Most say that increasing market share is
either an explicit goal or at least a means of attaining another goal, such
as increased net returns.

For those we talked to, the student loan operations within banks do not
hedge within their own shops. Any hedging is done at the corporate level.
One lender, for whom student loans are a small part of a large organization,
mentioned that TED spread changes and the fall 1998 crisis did not affect it
directly. They sell most loans soon after origination rather than holding
large portfolio of loans over a long period; in addition, swapping and other
forms of hedging are done at the corporate level, not at the level of the
student loan group.

Marketing for student loans is aimed at schools, not borrowers. To generate
new volume, lenders try to get themselves included on a school?s preferred
lender list.

Appendix 8 140 Other Developments in the FFELP Industry Many fewer lenders
participate in the program compared to 5 or 10 years ago-- Department of
Education data show the number has fallen by several thousand. Some have
dropped out of the program; others have been merged or bought up by larger
institutions. However, only a few of the large lenders have dropped out.

Consolidation among student loan servicers has been even more pronounced.
Only a half-dozen or so servicers now dominate the market. Other lenders
generally contract with one (or more) of these dominant servicers for their
loans. Servicing requirements may vary by guarantor on the loan. One lender
told us it services its own loans that are guaranteed by the guarantor in
their state, with whom it deals most regularly, but it contracts out
servicing if the loan is with any other guarantor. Outsourcing servicing
with one of the well-known servicers also makes a lender?s loans look better
for either secondary market sale or securitization (ratings companies look
more favorably on the portfolio). One interviewee noted that because
securitization has made funding costs more equal across lenders, servicing
cost differences are now the main variable factor in profitability, rather
than funding cost differences.

The big risk (other than interest-rate risk) in student loans is not the
credit risk per se, which is the major risk for other types of loans.
Instead, it is servicing risk, since improper servicing can void the credit
guarantees provided by the guarantee agencies. Given the importance of
servicing, the servicer for student loans doesn?t usually change even if
ownership of the loan does. By keeping the same servicer during origination
and later servicing, the owner of the loans knows who made any mistake that
cost him his guarantee. One secondary market lender told us that it will bid
slightly more to purchase a portfolio of loans if it already have a
servicing contract in place with the portfolio?s servicer. If, on the other
hand, it would need to develop a new servicing relationship, it might build
that cost into the bid and offer a lower price. Implicit in this latter
strategy was the idea that purchasing the loans and then simply switching
them to a servicer it already worked with was not an option.

Interviewees were mixed on the possibility of servicing student loans in the
same centers in which lenders might service credit card or other operations.
Some thought economies of scale or scope could be exploited, while others
thought student loans are so unique that the differences between the
products overcame any potential advantage achieved by combining operations.

Servicing costs range widely, from somewhat under 1 percent to somewhat over
1 percent of a portfolio. One estimated range was 65 to 150 basis points;
another interviewee told us that 100 basis points, or slightly less, was the
usual ballpark estimate of the average; a third said 100 to 135 basis
points. Because funding costs are generally about 200 basis points below the
lender yield, a

Appendix 8 141 variation of 35 basis points in servicing costs has a high
relative impact on profitability.

Up-front discounts for borrowers on the 1-percent guaranty fee began several
years ago. The general trend has been driven by national guaranty agencies
looking to expand their portfolios and state-based entities looking to
protect portfolios from this expansion. (For both guaranty agencies and
lenders, volume is important, because fixed costs of participation are
relatively high and need to be spread over as large a pool as possible.)
Several interviewees mentioned that once a few agencies began discounting,
most of the others felt compelled to do so as well or face a big loss in
market share.

Interviewees also mentioned the effect of the 1998 Higher Education Act
reauthorization and the anticipated recall of more guaranty agency reserves.
In their view, agencies decided to give money back to students before they
lost it to the federal government. In one view, discounting is ?prepaying?
that money back to the government. (The 1998 reauthorization also
established two components of guaranty agency funds--an operating fund and a
federal guarantee reserve fund--and designated which dollars could go into
each and what payments could be made from each.) Several interviewees told
us this discounting would not be sustainable as guaranty agency managed
federal funds were depleted--some thought it might last for a matter of
months, while others thought that stronger agencies could hold out for a
year or two.

Although guaranty fee discounts have been around for a few years, discounts
on the 3-percent origination fee (sometimes 1 percentage point of this fee,
sometimes the full 3 percent) are more recent. Some of these discounts are
offered by state secondary markets, and banks that originate loans and have
forward purchase agreements with secondary markets will generally offer
their borrowers the terms that the secondary market offers. Back-end
discounts, whereby borrowers? interest rates might go down if they make
their first 36 or 48 payments in a timely manner, have also become common.

Competition from Direct Loans (DL) is a factor. The reduction of DL up-front
fees has not caused schools to switch to DL but may have partially arrested
a switch away from DL. One interviewee also mentioned that, although lenders
and other participants had a good sense of how default, FFELP consolidation,
and other events might affect a portfolio over time, the consolidation of
FFELP loans into a DL consolidation loan ?added a new wrinkle? and was less
predictable.

Tax-exempt funding has become less prevalent in recent years for several
reasons. One is that federal rules were put in place capping the amount
available in each state, and student loan authorities must compete with
others in the state (road and school construction, for example) for an
allocation under the cap. Another reason may be that state-designated
secondary markets have

Appendix 8 142 begun to operate more outside their home state, and they
cannot use tax-exempt funding for these operations.

There are some similarities between securitizations and bond sales by state-
designated secondary markets--cash flows and evaluations by credit ratings
agencies are similar.

Growth of Alternative (Private/Nonfederal) Loans Alternative loans have
grown rapidly beginning 6 to 8 years ago. They primarily go to students at
4-year schools and higher tuition schools; thus, they go to borrowers who
are attractive for a portfolio. (At first, they went primarily to graduate
students, but now they?re going to undergraduates as well.) Alternative
loans are not federally guaranteed, but lenders generally insure them using
private insurers. These loans can be included in securitizations, but most
lenders do not do so.

Appendix 9 143

APPENDIX 9: SURVEYS OF LENDERS AND SECONDARY MARKET INSTITUTIONS

During the Spring and Summer of 1999, we surveyed a limited number of
lenders and secondary market institutions, and this appendix discusses the
sample and results. Survey questions were reviewed by study group members
and modified to address their concerns; GAO staff sent out the survey and
compiled the results.

Sample for Surveys We asked the Education Finance Council (EFC) to
coordinate data gathering for state secondary markets (hereafter known as
the EFC sub-sample) and the Consumers Bankers Association (CBA) to
coordinate data gathering for other student loan lenders/holders (hereafter
known as the CBA sub-sample).

Each group sent out surveys to the institutions on the study group and a
small number of other institutions, of which we selected some and EFC and
CBA selected others. We tried to include both large and small institutions
in both groups, but the sample is not statistically representative of the
population represented by either group.

We received 24 usable survey responses. 1 The institutions that responded
held well over 50 percent of Federal Family Education Loan Program (FFELP)
loans held as of September 30, 1997. According to EFC, the 12 respondents in
the EFC subsample held over 45 percent of EFC-member state secondary market
loans. Based on Department of Education fiscal- year-end-1997 data, the 12
respondents in the CBA sub-sample held over 66 percent of other student loan
lender/holder loans.

Results for CBA Subsample Nine of the 12 respondents held more than $1
billion on book at FYE 1998. Eight of them originated more than $500 million
in loans in 1998.

Only one CBA institution, one of the smaller ones in the sample, was
primarily deposit-based (at 65%; no other had more than 14% of loans funded
by deposits). Of 11 respondents, 4 had 100% LIBOR-based funding, 2 others
were majority LIBOR-based, and 2 were 30%-50% LIBOR. Depending on the year,
3

1 We received 27 responses, but 3 of the 27 ?respondents? said they had
dropped out of the student loan business, so data are available on 24
respondents (or fewer, depending on the question).

Appendix 9 144 or 4 of the 11 had some T-bill-based funding, and only 1
respondent (and for that respondent, for only 1 of the 3 years) had more
than 50% T-bill-based funding.

Only four of 12 institutions hedge against their student loans, but they
were 4 of the 5 largest in our sample, based on reported 1998 holdings.
These lenders reported the term of the hedge as either less than 1 year or
1-3 years. From 1997 to 1998, the percentage hedged went up for one, down
for another, and remained constant for the other 2.

Five of the 12 respondents had done some type of securitization. Of these, 3
had some securitization that was T-bill-based, and 4 had some LIBOR-based
securities (with 2 using a mix of both). One of the 5 had a swap within
their securitization; the other 4 did not.

All 11 who responded said they would prefer a 1-month or 3-month LIBOR
basis, as opposed to the T-bill or commercial paper, for the index for their
yield (8 preferred 1-month LIBOR, 3 preferred 3-month LIBOR).

Responses for EFC Subsample EFC institutions we surveyed were smaller than
institutions in the CBA subsample: 8 of the 12 held less than $1 billion on
book at FYE 1998 (and 9 of 12 were under $1 billion in 1997). All 12
reported consistent growth of on-book holdings from 1996 to 1997 and 1997 to
1998--in the CBA subsample, on-book holdings sometimes fluctuated from year
to year.

Of the 12 EFC institutions, 5 had some LIBOR-based funding in 1998 (but 2 of
the 5 had only 1% LIBOR), and 3 had some T-bill-based funding. All 3
institutions with substantial LIBOR-based funding showed an increase in such
funding over the 3 years; however, another institution had some LIBOR
funding in 1996 and 1997 but none in 1998. All 12 had more than 50% ?other?
funding (non-T-bill, non-LIBOR; some tax-exempt for those that indicated
what it was).

Only 1 of the 12 hedged what they held on book, and 1998 was the first year
they hedged.

Only one had any type of securitization. This securitization was indexed to
LIBOR, and it had a swap for some portion. In 1997 and 1998, the majority of
this institution?s securitized loans were held on book, while others had
been moved off book.

Eight of the 12 said they preferred a LIBOR basis for the index (7 said
1-month, 1 said 3-month), two said T-bill, one said commercial paper, and
one was indifferent between LIBOR and T-bill.

Appendix 10 145

APPENDIX 10: TREASURY DEPARTMENT SYNTHETIC SWAP MODEL TO ESTIMATE SWAP RATES

The following is a brief description of the estimation of the swap spread
for a swap that could be used to hedge the basis risk of a holder of a
guaranteed student loan. 1 In the Federal Financial Education Loan Program
(FFELP), guaranteed loans pay the holder a return (comprised of payments
from the student borrower plus Special Assistance Payments from the
government) that until recently was indexed to the three-month Treasury
bill. As described in chapter IV of this report, holders of FFELP loans
whose cost of funds is LIBOR- based, or tracks closely with LIBOR, face
interest-rate risk (basis risk) because of the mismatch between their
returns and their cost of funds. This risk may be removed, at a cost, with a
basis-to-basis swap. The calculations show that the estimated swap cost
widened substantially during the ?flight to quality? in the later months of
1998 and has not returned to the earlier level.

For each date, the model first estimates two variable-for-fixed-rate swap
curves (swap rates by term to maturity) and then combines them to arrive at
a basis swap curve. Initially, the model assumes 100% efficiency in
executing trades (no cash reserves and no dealers? fees). The first step is
to calculate the implied yield curve for LIBOR from the LIBOR futures rates
quoted in the Wall Street Journal. Holders of FFELP loans who have a cost of
funds that is based on LIBOR may lock into a fixed rate by selling LIBOR
futures. The second step is to obtain the Treasury yield curve published by
the Federal Reserve (implied discount factors and Treasury forward rates are
calculated). Finally, the two curves are combined, and a 15 basis point
adjustment is made. 2 This methodology is applied repeatedly over time to
generate a five-and-half year history of estimated swap costs.

Figure 10.1 shows the estimated swap costs, and figure 10.2 presents the two
underlying variable-for-fixed-rate swap curves. For each
variable-for-fixed-rate swap curve, the swap rate calculation is based on
certain amortizing terms. The relevant swap is used to hedge a pool of
student loans that amortizes over 10 years. The amortization assumption over
the 10 years is based on amortization terms of an actual swap embedded in a
student-loan-asset-backed security that was marketed in March 1999.

1 This model was developed as part of the U.S. Treasury Department?s
oversight of Sallie Mae. 2 Holders of the Treasury-indexed FFELP loans may
lock into a fixed yield by separately

executing the additional trades of investing in long-term Treasuries funded
with short-term repurchase agreements (REPO) borrowings. The short-term
borrowing is rolled over until the long-term Treasury investments mature. By
providing Treasury securities as collateral, a borrower can borrow
short-term at only slightly higher rates than those presented in the
Treasury yield curves. Due to the premium over Treasury rates that is
charged to borrowers in the REPO market, the Treasury yield curve overstates
the effective fixed yield that REPO purchasers can lock into. The amount of
the overstatement was estimated to be 15 basis points. The four figures each
incorporate the 15 basis point adjustment to the treasury side swap. No such
adjustment was deemed necessary for the LIBOR side swap.

Appendix 10 146 As an alternative, figures 10.3 and 10.4 show the basis swap
between LIBOR and Treasury--and the underlying estimated swaps between each
of them and a fixed rate--for a 7-year bullet term (i.e. with no
amortization) over the period since March 1994. The results of these bullet
calculations may be more familiar to active participants in the swap market
than figures 10.1 and 10.2.

Appendix 10 147

Figure 10.1: Historical Basis Swap Rates (Amortizing)

91-Day T-Bill vs. 3-Month LIBOR - 10 Year Amortizing Notional Term (Notional
Amortizing Schedule Based on an Actual Student Loan Portfolio Swap)

Notes. The base swap rate measurement methodology assumes 100% efficiency in
executing trades involving the Treasury yield curve, excluding premiums
charged borrowers in the Repo market. To adjust for this, .15% has been
added to the calculated swap rate. The average rate for the period was .55%.
0.00%

0.10% 0.20%

0.30% 0.40%

0.50% 0.60%

0.70% 0.80%

0.90% 1.00%

Mar-94 Jun-94

Sep-94 Dec-94

Mar-95 Jun-95

Sep-95 Dec-95

Mar-96 Jun-96

Sep-96 Dec-96

Mar-97 Jun-97

Sep-97 Dec-97

Mar-98 Jun-98

Sep-98 Dec-98

Mar-99 Jun-99

Sep-99

Percent

Calculated swap rate

Appendix 10 148

Figure 10.2: Historical LIBOR and Treasury Swap Rates (Amortizing)

3-Month LIBOR vs. Qt. compounding fixed - 10 Year Amortizing Notional Term
91-Day T-Bill vs. Qt. compounding fixed - 10 Year Amortizing Notional Term

Notes. The base Treasury swap rate measurement methodology assumes 100%
efficiency in executing trades involving the Treasury yield curve, excluding
premiums charged borrowers in the Repo market. To adjust for this, .15% has
been deducted from the fixed rate. The average rates for the period, for the
LIBOR and Treasury swaps with a 10-year amortizing notional term, were 6.48%
and 5.93%, respectively. 0.00%

1.00% 2.00%

3.00% 4.00%

5.00% 6.00%

7.00% 8.00%

9.00% Mar-94

Jun-94 Sep-94

Dec-94 Mar-95

Jun-95 Sep-95

Dec-95 Mar-96

Jun-96 Sep-96

Dec-96 Mar-97

Jun-97 Sep-97

Dec-97 Mar-98

Jun-98 Sep-98

Dec-98 Mar-99

Jun-99 Sep-99

Percent

LIBOR-to-fixed Treasury-to-fixed

Appendix 10 149

Figure 10.3: Historical Basis Swap Rates

91-Day T-Bill vs. 3-Month LIBOR - 7 Year Bullet Term Notes. The base swap
rate measurement methodology assumes 100% efficiency in executing trades
involving the Treasury yield curve, excluding premiums charged borrowers in
the Repo market. To adjust for this, .15% has been added to the calculated
swap rate. The average rate for the period was .52%. 0.00%

0.10% 0.20%

0.30% 0.40%

0.50% 0.60%

0.70% 0.80%

0.90% Mar-94

Jun-94 Sep-94

Dec-94 Mar-95

Jun-95 Sep-95

Dec-95 Mar-96

Jun-96 Sep-96

Dec-96 Mar-97

Jun-97 Sep-97

Dec-97 Mar-98

Jun-98 Sep-98

Dec-98 Mar-99

Jun-99 Sep-99

Percent

Swap rate

Appendix 10 150

Figure 10.4: Historical LIBOR and Treasury SWAP RATES

3-Month LIBOR vs. Qt. compounding fixed - 7 Year Bullet Notional Term 91-Day
T-Bill vs. Qt. compounding fixed - 7 Year Bullet Notional Term

Notes. The base Treasury swap rate measurement methodology assumes 100%
efficiency in executing trades involving the Treasury yield curve, excluding
premiums charged borrowers in the Repo market. To adjust for this, .15% has
been deducted from the fixed rate. The average rates for the period, for the
LIBOR and Treasury swaps with a 7-year bullet notional term, were 6.51% and
5.99%, respectively. 0.00%

1.00% 2.00%

3.00% 4.00%

5.00% 6.00%

7.00% 8.00%

9.00% Ma

r- 94

Ju n- 94

Se p- 94

De c- 94

Ma r- 95

Ju n- 95

Se p- 95

De c- 95

Ma r- 96

Ju n- 96

Se p- 96

De c- 96

Ma r- 97

Ju n- 97

Se p- 97

De c- 97

Ma r- 98

Ju n- 98

Se p- 98

De c- 98

Ma r- 99

Ju n- 99

Se p- 99

LIBOR-to-fixed Treasury-to-fixed

Appendix 11 151

APPENDIX 11: RECOMMENDATION OF THE FFELP INDUSTRY MEMBERS OF THE STUDY GROUP
ON THE FEASIBILITY OF ALTERNATIVE FINANCIAL INSTRUMENTS FOR DETERMINING
LENDER YIELD

(Note--This recommendation was made before the Congress passed legislation
in November 1999 that changed lender yield to an index based on Commercial
Paper.)

The Study Group on the Feasibility of Alternative Financial Instruments for
Determining Lender Yields (the Group) has worked for several months to
fulfill its mandate as set forth in Section 802 of the Higher Education Act.
The Group?s work has taken place in the context of legislative budget
procedures that limit the scope of options that may be recommended by the
Group. These constraints include projections of LIBOR, Commercial Paper and
Treasury bill rates, estimates of Federal budget costs which include costs
associated with the volatility of such rates, and budget costs associated
with any change from the formula for lender yield scheduled to go into
effect on July 1, 2003.

Recommendation The FFELP industry members of the Group (FFELP Group)
carefully examined the historical relationships between the 91-day T-bill,
Commercial Paper and LIBOR. Based on this historical analysis and an outlook
for continued moderate economic growth, low short-term interest rates and
stable inflation, the FFELP Group concludes that the following alternative
lender return formulas produce no additional cost to the government:

FFELP Group Recommended Conversion Formulas

30-day Commercial Paper plus 2.45% 90-day Commercial Paper plus 2.40%
1-month LIBOR plus 2.35% 3-month LIBOR plus 2.30%

Note: Conversion assumes loans are in repayment (91-day T-bill plus 2.80%).
Based solely on capital market forces, the FFELP Group would have
recommended that the reference rate used in the formula for determining
lender yield should be based on the 1-month LIBOR rate. Providers of Federal
Family Education Loan Program loans support this recommendation. Using LIBOR
most closely corresponds to financing strategies now used in support of
student loans. By matching the lender index to funding strategies, liquidity
risks to the

Appendix 11 152 FFELP program associated with external events are greatly
reduced. Borrower interest costs would not be changed as a result of this
proposed change in the formula for lender return.

However, given the budget scoring process, the FFELP Group recommends that
the Congress adopt the 90-day Commercial Paper (CP) rate plus 2.40% (during
repayment), as the reference rate used to determine lender yield on FFELP
loans. Based on the analysis presented in the FFELP industry ?white paper?
prepared for the recent Reauthorization (see page 21), 1 the conversion to
CP plus 2.40% results in a slightly lower return to lenders than that
received under the T-bill based formula now in effect. However, 90-day CP is
the spread relationship that is the most consistent with those forecasted by
the Congressional Budget Office. Indeed, the conversion to 90-day Commercial
Paper plus 2.40% results in some budgetary savings over the period to July
1, 2003. In addition, commercial paper rates are highly correlated with
lenders? cost of funds, are published daily by the Federal Reserve and
vastly reduce the liquidity risks associated with the current index.

These recommendations will be supplemented with a more complete report that
will include analyses prepared for and used by the Study Group over the past
several months.

Analysis The FFELP industry members of the Alternative Indices Study Group
recommended the following lender return formulas together with the related
T-bill index conversion spreads:

Conversion Lender Formula Spread

30-day Commercial Paper plus 2.45% 0.35% 90-day Commercial Paper Plus 2.40%
0.40% 1-month LIBOR plus 2.35% 0.45% 3-month LIBOR plus 2.30% 0.50%

The FFELP industry participants? proposed 90-day commercial paper (CP) loan
formula incorporates a 91-day Treasury bill (T-bill) to 90-day CP spread
forecast of 0.40%. We believe that CBO?s forecast of 0.45%, which is quite
close to the industry?s forecast, needs to be lowered because of the
difference caused by the break in the CP data series as a result of the
September 1997 changes implemented by the Federal Reserve. Calvin Schnure,
the Federal Reserve economist responsible for the CP composite rate series,
has stated that analysis conducted by the Federal Reserve prior to the
implementation of the improved

1 This paper can be found at
http://www.salliemae.com/government-relations/whitepaper.pdf or at
http://www.nchelp.org/contents/elibrary/download/Document/FFELP_AltInd1999.pdf

Appendix 11 153 collection methodology showed that average rates in the old
data series were .10% higher than those in the new series. Financial market
news items and market data from that time support this conclusion. We would
expect that if CBO factors such effect into its historical data that the
90-day CP spread would be reduced from its current forecast of 0.45%. We
also point out that from 1990 to 1998 the T-Bill/CP spreads were equal to or
less than .40% in excess of 75% of the time when the CP data series are
conformed.

During the 1998 reauthorization, consideration was given to changing the
index to 90-day CP + 2.40%. Our fundamental long-term economic outlook has
not changed over the past year. Thus, we believe that a 90-day CP rate with
a margin of 2.40% is an appropriate level in order to retain similar
economics for market participants as well as for Federal budget scoring
purposes.

We believe that the 1990?s are the appropriate historical period to use when
forecasting spreads between LIBOR and T-bill rates. In the industry?s
briefing paper from March 1999 entitled ?The Federal Family Education Loan
Program: Alternative Indices for Determination of Lender Returns,? we
calculated that for the nine years ended December 31, 1998, the average
90-day LIBOR/T-bill spread was .507%. Our current forecast for the spread
between 90-day CP and 90-day LIBOR is .10%. Thus, the historical value of
LIBOR to T-Bill during the relevant period of the 1990?s, and the tight
relationship to CP leads us to conclude that a loan yield of 3-month LIBOR
plus 2.30% represents the economically neutral level for this alternative
index. Our conclusion is also substantiated by the OMB forecast through 2009
of a T-bill/LIBOR spread relationship of less than 0.50%.

Our long-term outlook for 30-day instruments indicates that rates will be
lower than those for 90-day instruments, consistent with the past 10 to 15
years. We do not believe that the LIBOR yield curve will be inverted during
the next ten years. The 5 basis point difference between the 30-day and
90-day instruments approximates the spreads from the 1990?s and also is the
same as CBO?s forecast for the CP.

Appendix 12 154

APPENDIX 12: ADMINISTRATION?S EVALUATION OF BASIS RISK FOR THE GOVERNMENT
UNDER ALTERNATIVE FINANCIAL INSTRUMENTS

The Administration recognizes that FFELP lenders confront basis risk, but we
note that dealing with this risk by changing the index for determining
payments to lenders simply shifts the risk to the government. We believe
that the second Congressionally mandated study group should search for
financial structures that could mitigate or eliminate basis risk, rather
than shift it. Changing the index for determining lenders? returns appears
to have large and varying implications for individual lender?s yields and
does not fully address lenders? ability to raise funds during general
disruptions in financial markets. Consequently, we believe such a change
should be accepted for the long-term of the FFEL program only if convincing
evidence emerges that basis risk must be addressed to maintain the program?s
health and after alternatives are carefully weighed and found to be less
acceptable.

Basis risk is not unique to student loans, and it may be dealt with in a
variety of ways including hedging, bearing the risk, or changing the index
for determining lender yield to a market rate more closely representing
lenders? costs of funds. Changing the index does not fully remove basis risk
since no one reference rate perfectly captures lenders? funding costs.
Furthermore, lenders pay an added margin over the reference rate when they
raise funds and variations in this margin with market conditions would not
be compensated by a change in reference rate. However, changing the index or
reference rate likely considerably reduces lenders? basis risk.

Changing the index or reference rate also shifts the basis risk to the
federal Government. It makes the Government?s interest payments to lenders
dependent on a private rate (LIBOR or the commercial paper rate) while
student interest payments and the Government?s cost of funds continue to be
linked to Treasury rates. The market value of this basis risk may be
approximated by what lenders pay to hedge the risk. One means for doing so
is the swap market. 1 Alternatively, the value may be read from parallel
transactions, alike in all respects but the reference rate, such as occurred
with the Sallie Mae student- loan asset-backed security sales in June and
August.

1 As described in Chapter 4, the hedging vehicle relevant for FFELP lenders
would be LIBOR/T- bill basis swaps. A swap involves the exchange between the
two parties of streams of future income. One party promises, in this case,
to make LIBOR-based payments (possibly because the firm has LIBOR-based
assets to provide the income); the other party makes T-bill based payments
(possibly because the firm has T-bill based assets, such as student loans)
and receives the LIBOR-based payments. These contracts usually require a
?swap spread? of the LIBOR rate above the T-bill rate that determines the
payments. The swap market may not be perfectly efficient for hedging
because, as also noted in Chapter 4, the swap spread (the cost of the hedge)
may change with the size of transaction. Hedging may be done using other
transactions as well, and the costs tend to be roughly the same.

Appendix 12 155 In the June sale, the first tranche of securities (with the
shortest maturity and first claim on payments) was sold with either a LIBOR
reference rate or a T-bill reference rate. The respective rates were LIBOR
plus 8 basis points or T-bill plus 87 basis points, implying a 79 point
spread. For the second such security sale by Sallie Mae, the spread at sale
between LIBOR and T-bill-based rates on the corresponding tranche was 86
basis points. Although definitive data on swap spreads can be difficult to
obtain for reasons discussed in Chapter 4, the general level of swap spreads
at the time was quite near the LIBOR-T-bill spread on these Sallie Mae
security offerings. While these measures indicated hedging costs of 80 to 85
basis points, the spread between the rate on T-bills and LIBOR, in the cash
markets, was 45 to 55 basis points. For commercial paper, the cash market
spread was 35 to 38 basis points.

Budget scoring of the cost of the program to the government is based on
projections of the cash market rates because this scoring is a forecast of
cash flows expected, on average, over the future. 2 An excess of the swap
spread over the cash spread implies that a subsidy is conferred by the
government when it assumes the basis risk, even if the mark-up in the
formula for lenders? returns is adjusted to achieve budget neutrality. The
swap spread, or corresponding measure of hedging cost, determines the
economic value of the shift of basis risk. The cash market spread roughly
determines how much return the lenders must give up in the formula for their
return in order to achieve budget neutrality while changing the index. When
swap spreads exceed spreads in the cash market, the difference is an
indication of the uncompensated gain, i.e., subsidy, to lenders who receive
more than they giving up.

Swap spreads typically differ from the spread between the rates in the cash
market, because of: (1) expectations about future movements in the two rates
over the life of the swap contract and (2) a risk (or insurance) premium. A
risk premium is built into longer-term interest rates and forward rates, and
correspondingly built into the swap spread. Risk premiums exist because most
participants in financial markets prefer more certainty about streams of
future income or costs. For example, fixed-rate mortgage rates usually
exceed floating rates, because the borrower is willing to pay for knowledge
of the amount of future payments and being protected against an upward
surprise in costs. The tendency for swap spreads or other hedging costs to
exceed the spread between alternative reference rates in the cash market
means that the adjustment in the

2 As indicated in Chapter 2, CBO projected measures of the likely error of
their forecast as well as the level of rates when scoring alternatives for
the FFELP index. These measures of likely error were used, as discussed in
Appendix ----, to allow for the fact that indexes with greater volatility
will create higher budget costs, on average, because of the lenders? return
formula is asymmetric - the federal cost of SAP payments rises if Treasury
rates exceed the 8.25 percent ceiling for setting student rates while the
amount that lenders give up when rates fall within a year is limited by the
student rate floor. This allowance for asymmetry in estimating costs that
will occur, on average, is not the same as charging for the increase in
volatility of federal costs that results from the government assuming basis
risk.

Appendix 12 156 lenders? return formula would need to exceed the amount
required for budget neutrality if lenders were not to receive a subsidy. 3

The Administration participants in the Study Group believe that an
additional subsidy is not appropriate for two reasons. First, the issue of
lenders? returns had been extensively considered by Congress in the
reauthorization of the Higher Education Act in 1998 and the mandate for the
current study did not appear to contemplate a reopening of that issue.
Second, as a matter of public policy principle, the government should not
absorb risk - particularly uncompensated risk - without assuring that no
other means is available to achieve the public policy purpose. Heedless
assumption of financial market risks by the government from the private
sector could easily mount into a misallocation of risks in financial
markets, distorting economic decisions and exposing taxpayers to added costs
in servicing outstanding federal debt.

On the other hand, some levels of compensation to the government for the
economic transfer of basis risk might effectively require some lenders to
pay for an amount of hedging -- from the change in the index -- that they
did not want. For example, some lenders might initially have less basis risk
than others. Consequently, it is understandable that Administration
representatives and lenders, as a group, approached the basis risk issue
quite differently.

Faced with this dilemma, Administration participants in the Study Group
determined that the issue of basis risk should be considered in the second
Congressionally mandated study. That study has a broader mandate, to explore
market-based means for determining lenders? returns. In the context of that
broader mandate, means might be considered for reducing basis risk rather
than simply shifting it from one party to another. Indeed, some individual
members of the Study Group mentioned such options as lender access to
federal funding, at a T-bill indexed rate, or changing the index for the
student rate. These options were viewed as outside of the mandate of the
current study and were not explored. Consideration of the basis risk issue
in the second study would permit

3 Compensation to the government for the transfer of basis risk could well
be below the actual cost of hedging for lenders because it would abstract
from market imperfections that cause the price of hedging to vary with the
size of the transaction, and it probably should be determined by the bid
rather than ask side of the swap market so as to exclude returns to swap
dealers. Additionally, the compensation should exclude the value of the
floor on lenders? returns (this floor at the student rate is equivalent to a
financial option, with market value as discussed in Chapter 4) since the
floor limits the needed amount of hedging to neutralize basis risk.

Appendix 12 157 exploration of other means -- such as, but not limited to,
these -- for dealing with the problem, before turning heedlessly to
government assumption of the risk.

Appendix 13 158

APPENDIX 13: PRINCIPLES OFFERED BY NON-LENDER, NON- GOVERNMENT MEMBERS, AND
RESPONSE TO PRINCIPLES, OFFERED BY FFLEP COMMUNITY MEMBERS

Principles Offered by Non-Lender, Non-Government Members

Herewith is a proposed set of principles with which to evaluate the
proposals put forth:

First, budget neutrality to protect the taxpayers' interests. Second,
borrowers should benefit equally in any changes, for example by a reduction
in fees or a reduced repayment rate.

Third, lenders deserve to earn a reasonable return on their investment and
risk. Fourth, there should be no increase in the complexity of the programs;
i.e., any changes should be systemic, not ad hoc.

Fifth, any changes should not detract from the goals of the second study.
Sixth, the recommendation should affect the long-term as well as the
short-term. Seventh, the recommendation should not detract from a healthy,
competitive market for student loans.

Eighth, any change recommended should be consistent with the structure of
the loan industry in term of liquidity, volatility, etc.

Ninth, the timing of changes to FFELP/DL should be made at a logical and
manageable time of the year for all parties involved (lenders, schools,
students) and the changes should be communicated in a systematic manner to
all parties involved.

Tenth, both FFELP and DL borrowers should have an equal opportunity to
benefit from any financial changes to FFELP. From the student's point of
view, both are federally subsidized student loan programs. It's not fair for
one student group to benefit from reduced costs and not the other.

Appendix 13 159

Response to Principles, Offered by FFELP Community Members

This memo is to respond to the suggest that we provide you with our analysis
of our most recent proposal, as put forward at the August 31st meeting of
the Study Group, using the principles which you shared with the Group on
September 1st.

We believe that when applied to any proposal, there will be some tension
among some of the individual principles. We have noted some examples.

Our responses are as follows: Principle 1 - Any change must be budget
neutral to protect the taxpayers? interest.

The FFELP community participants? approach to determining the correct
conversion from the existing T-bill based formula to one based on Commercial
Paper or LIBOR has always been that any recommended conversion would
maintain budget neutrality. The conversion formulas recommended by the FFELP
community can be implemented at no additional cost to the taxpayer through
the 2003 reauthorization and at no additional cost above the existing 91-day
T-bill formula after 2003. Budget neutrality should be based upon the final
CBO scoring analysis.

Principle 2 - Borrowers should benefit equally from any change. Borrowers
are currently benefiting from the low T-bill interest rate environment and
the 1998 legislated 0.80% borrower rate reduction. By changing the index
from the 91-day T-bill to one that reflects lenders? cost of funds, lenders
will be able to more closely match fund their student loan portfolios. This
will encourage lenders to remain in the program and take a long-term view on
investing in and improving the infrastructure of this student loan program.
Consistent lender participation and investment in the program will result in
increased competition among lenders to attract borrowers and schools.
Greater competition invariably leads to lower costs and better service for
borrowers. In the past, lender competition has resulted in reduction of fees
and repayment rates to students, a wide choice of lenders by schools and
needed investments in technology to develop systems that better meet school
financial aid administration needs. Investments in student loan servicing
technology will reduce default rates thereby reducing the overall cost of
the student loan program to the taxpayer. Students pursuing a higher
education in the future will be assured of the continued flow of private
sector capital to finance education through all economic cycles as well
continued investments in service delivery.

Appendix 13 160 Principle 3 - Lenders deserve to earn a reasonable return on
their investment and risk.

Lender return has been reduced due to the 1998 Reauthorization and market
conditions. Many lenders are not hedging against basis risk since it is not
affordable. The final index change will have different effects among
lenders. Under the lender formulas advanced by the FFELP community members,
lenders are expected to earn returns over the long-term that are slightly
lower than what they have earned in the past. Lenders are willing to accept
slightly lower future earnings in exchange for a decrease in volatility in
the spread between what they receive on their student loan assets and pay on
their funding. Vigorous competition will ensure that borrower interest rates
are kept to a minimum and services continually improve.

Principle 4 - There should be no increase in the complexity of the program.
Changing the index on which lender yields are based will not increase the
complexity of the program to students, schools or lenders. It will only make
raising capital more predictable for lenders. The change to a Commercial
Paper or LIBOR index can be implemented by lenders in the same manner as
past changes to the T-bill based formula.

Principle 5 - Any changes should not detract from the goals of the second
study. The goals of the second study are set forth in section 801 of the
Higher Education Act. Implementation of a market-based index is consistent
with the second study?s mandate to establish a market-based mechanism to
determine loan pricing. Lenders are committed and eager to create a stronger
and more rational industry structure.

Principle 6 - The recommendation should affect the long-term as well as the
short-term.

During the anticipated applicability of the revised reference rate formula
(through June 30, 2003), more than $55 billion in FFEL loans are expected to
be made. These loans have an average life of at least eight years, and some
will be outstanding longer. Even if a permanent change in the law is not
feasible at this time, a change through June 2003 would represent a positive
step towards improving the FFEL Program as we face the significant
challenges of the future.

Viewed more broadly, the adoption of a modern reference rate formula will
help assure the continued participation of a broad diversity of lenders and
other loan providers. This will set the stage for increased market
competition well into the future. This market competition will benefit
students in the of better customer service and rate competition.

Appendix 13 161 Principle 7 - The recommendation should not detract from a
healthy, competitive market for student loans.

A change in the lender index will reduce the uncertainty associated with
match funding student loan portfolios under the current T-bill indexed
formula. More stable and predictable match funding will encourage lenders to
remain in the program and take a long-term view when considering investments
to maintain and improve the infrastructure of the student loan program. From
1990 through 1997, the number of lenders participating in the FFELP student
loan program declined by over 50 percent. A change in the index will help
stem the decline in lender participation and increase the competition among
lenders.

Principle 8 - Any change recommended should be consistent with the structure
of the loan industry in terms of liquidity and volatility.

The relatively unstable relationship between T-bill rates and student loan
providers? funding costs has had a significantly negative impact on FFELP
participants? abilities to match fund their portfolios. By switching to a
market- based index and reducing the unnecessary capital markets uncertainty
that accompany the T-bill index, predictability and certainty in the FFEL
program can be enjoyed at no additional cost to students, schools or
taxpayers. A change to a lender index based on Commercial Paper will greatly
enhance liquidity to FFELP participants by reducing the funding volatility
associated with the legacy T-bill index.

Principle 9 - The timing of changes to FFELP/DL should be made at a logical
and manageable time of the year for all parties involved (lenders, schools,
students), and the changes should be communicated in a systematic manner to
all parties involved.

Changing the index on which lenders are compensated will not affect students
or schools. In fact, as competition between lenders increases, borrowers and
schools should benefit over time from program improvements.

Lenders will be required to make modifications to their loan servicing
systems and they along with their loan servicers have sufficient expertise
to make such modification within a relatively short time frame. Servicers
historically have made numerous changes due to revision in the laws and
regulations governing guaranteed student lending. Any changes, of course,
would be carefully coordinated with the Department of Education.

Principle 10 - Both FFELP and DL borrowers should have an equal opportunity
to benefit from any financial changes to FFELP.

Appendix 13 162 As indicated above, legislated student rates will not be
impacted by a change in the lender rate. Some differences in the FFELP and
DL programs exist but they are unrelated to the lender interest rate index.

Appendix 14 163

APPENDIX 14: FFELP COMMUNITY MEMBER COMMENTS ON THIS REPORT AND THE
IMPLICATIONS OF BUDGET SURPLUSES

This section represents the views of the representatives of private sector
participants in the Federal Family Education Loan Program (FFELP) on the
Alternative Indices Study Group. Section 802 of the Higher Education Act of
1998 required that the study group submit a report to Congress evaluating
the index on which lender returns are based in the student loan program.
Given the importance of the issue and the course of the study itself, we
felt compelled to put forth our views because the report, while replete with
details and analysis, does not address one of the most significant issues
facing the FFELP. The fact is that drastic changes in the Treasury
securities markets necessitate a change in the index on which lenders earn
returns in the student loan program. The FFELP community has advocated
proactively for some time to change to a new index, which was enacted at the
end of the last congressional session. Developments since then have further
demonstrated the necessity of that move.

While the Alternate Indices Study Group discussed changing Treasury markets
as a fundamental problem, the report does not address the issue. In
particular, chapter 3, which discusses the Treasury market specifically,
largely ignores the pending challenges to the Treasury Department in
managing the supply of Treasury securities in the time of budget surpluses.
The chapter looks backward (?Past as Prologue?) to provide assurances that
?significant consequences (of

budget surpluses on the Treasury bill market) seem quite far in the
distance.? However, there are plenty of sources that paint a less sanguine
picture. Both the Administration and the Congressional Budget Office project
the elimination of debt held by the public within twelve to thirteen years.
Over the past few years, with the advent of federal budget surpluses, each
year?s projection has moved up the date when the debt could be eliminated.
Under CBO?s most conservative forecast, there will be more surpluses than
debt available to retire by 2009. 1 In other words, CBO is forecasting that
the federal government will have excess cash on-hand by 2009. If that
forecast transpires, the short-term treasury bills would be virtually
eliminated. In the short term, CBO is forecasting that the public debt will
be reduced by 5 percent in 2000 and 2001, increasing to 6 percent in 2002, 7
percent in 2003, and 8 percent in 2004.

1 Congressional Budget Office, ?The Budget and Economic Outlook: Fiscal
Years 2001-2010,? January 2000, p. 20. Forecast assumes that discretionary
spending increases at the rate of inflation every year.

Appendix 14 164 CBO Projections of Debt Held by the Public Assuming
Discretionary Spending Increasing with Inflation

(in billions of dollars) 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008
2009 2010 Debt Held by Public 3,633 3,455 3,292 3,097 2,884 2,651 2,394
2,080 1,721 1,330 1,016 941 Reduction from py -89 -178 -163 -195 -213 -233
-257 -314 -359 -391 -314 -75 % reduction from py -2% -5% -5% -6% -7% -8%
-10% -13% -17% -23% -24% -7% Accumulated excess cash n.a. n.a. n.a. n.a.
n.a. n.a. n.a. n.a. n.a. n.a. 122 528

Similarly, the Administration?s fiscal 2001 budget puts forward a plan that
would eliminate the public debt by fiscal year 2013. Over the next five
years, the Administration?s budget projects that the debt held by the public
will be reduced by $852 billion, somewhat less than the $982 billion
forecast by CBO but considerable nonetheless.

The unexpectedly large surpluses in the past two years have already had an
impact on the market for Treasury securities, in general, and for the
Treasury bills, specifically. The General Accounting Office documented this
impact in its September 1999, report, ?Federal Debt: Debt Management in a
Period of Budget Surpluses.? The following excerpts from the report show the
impact of the unexpectedly large surpluses in 1997 and 1998 and the
challenge of managing debt reduction going forward.

The effect of the better-than-expected fiscal outcomes in 1997 and 1998
initially resulted in reductions in short-term debt... Since some bills
mature each week, the unexpected cash inflows were used to redeem bills.
However, according to a Treasury official, bills were redeemed at such high
levels that the liquidity of the bill market was adversely affected and the
average life of marketable debt increased modestly. 2

***

According to Treasury and Federal Reserve officials, the amount of bills
reduced was sufficiently large to cause the market for bills to become less
liquid. 3

The GAO report predicted that the large projected surpluses will pose a
significant challenge to the management of the Treasury securities markets
and that the goals of liquid markets and lowest cost to the government may
not be compatible.

Declining levels of debt prompt the need to make choices over how to
allocate debt reduction across the full maturity range of securities used.
The stakes associated with debt reduction strategies are considerable. As

2 General Accounting Office ?Federal Debt: Debt Management in a Period of
Budget Surpluses? September 1999 (GAO/AIMD-99-270), p. 5. 3 Ibid.

Appendix 14 165 debt declines, the Treasury faces more difficult trade-offs
in achieving broad and deep markets for its securities and the lowest cost
financing for the government 4 .

*** ? the level of debt held by the public projected by CBO for 2009 is less
than the dollar amount of federal securities owned by the Federal Reserve
and state and local governments combined at the end of fiscal year 1998? the
comparison above gives a sense of the size of the continuing and more
extensive adjustments by both the Treasury and market participants 5 .

*** These changes, for instance, may very well affect the use of Treasury
securities as benchmarks to price other financial transactions. Although
markets tend to adjust to these shifts over time, changes may not be
seamless or without cost 6 .

While markets may adjust to these shifts, the index for student loans is set
by statute and cannot be adjusted without legislation.

These challenges of managing Treasury securities in a time of federal
surplus are already apparent. In February, Gary Gensler, Under Secretary for
Domestic Finance for the U.S. Treasury, announced that, during the 3 rd
quarter of fiscal 2000, the government would have the largest reduction in
public debt in the nation?s history. Under the heading ?Debt Management
Challenges,? Mr. Gensler announced:

?as budget surpluses continue to diminish our borrowing needs, we now face
additional challenges going forward. First, debt held by the public is
forecast to shrink even further and faster than it has in the last two
years. As we announced on Monday, we estimate that we will paydown $17
billion in net market borrowing for the January-March quarter. This will be
followed next quarter with the largest reduction in publicly held debt in
our nation?s history, as we pay down approximately $152 billion. More
significantly, there is now a consensus among private sector and government
forecasters that these paydowns will grow in the future 7 .

4 Ibid, p. 17. 5 Ibid, p. 18. 6 Ibid. 7 February 2, 2000, Press Release from
the Office of Public Affairs, U.S. Treasury: Under

Secretary of the Treasury for Domestic Finance, Gary Gensler, Remarks at the
February 2000 Treasury Quarterly Refunding.

Appendix 14 166 The strategy that Treasury announced to manage this historic
paydown of public debt already has had huge ramifications in the Treasury
markets. While, in the short term, Treasury has taken actions to maintain
the short-term bill market, assurances such as those in chapter 3, ?Treasury
debt management is committed to maintaining a deep and liquid T-bill
market,? are unlikely to stand up in the long-term to the unprecedented
reduction in public debt. We are likely to see, over the average life of a
FFELP loan, even more dramatic changes in the Treasury markets.

It is important to keep in mind the long-term characteristics of student
loans. The repayment term of a FFELP loan is 10 years. Most of these loans
do not go into repayment until the borrower has left school-the average
in-school period is two years. The loan life can be extended through
deferments, forbearances, and may be extended for a period of up to 25 years
or 30 years, respectively in the case of extended repayment or
consolidation. As a result, there may be a large amount of FFELP loans
statutorily indexed to an instrument that no longer exists.

This potential risk is remarkably close to reality today for nearly 6
percent of outstanding FFELP loans: PLUS loans issued between 1992 and 1998
which are indexed to the 52-week Treasury bill. In February, Mr. Gensler
announced that the number of auctions of 52-week bills would be reduced from
13 per year to four per year. In addition, Mr. Gensler stated, ?It is likely
that, as further reductions in issuance become necessary, elimination of the
one year (bill) will be considered.? 8 Certainly, when these loans were
issued, it was inconceivable that the instrument on which the PLUS index was
based would be eliminated. Few anticipated the remarkable change in the
federal budget and projected surpluses. However, the government?s good news
is clearly a crisis for FFELP. Could there be any more compelling evidence
for the need for the new FFELP index?

Going forward, it is vital that new FFELP loans are being issued under a
more resilient index. The March 1999 report by the FFELP community, ?The
Federal Family Education Loan Program: Alternative Indices For Determination
of Lender Returns,?-which is included in Appendix xx of the
report-documented the global capital market?s overwhelming preference for
securities indexed to LIBOR or commercial paper rates and the need to base
the FFELP lender returns on one of these rates. As the Treasury debt
continues to be reduced, the potential need to find new benchmarks for
securities increases. Deputy Secretary Stuart Eizenstat acknowledged this
potential in his confirmation hearings before the Senate Finance Committee
last year. In responding to a question on the key uses of Treasury
securities and what impact the elimination of the public debt would have on
these uses, he stated:

As the supply of Treasuries dwindles in the future, as we gradually reduce
the debt held by the public, there would be a ready supply of other

8 Ibid.

Appendix 14 167 securities of other issuers including high quality
corporations and government sponsored enterprises that would likely become
benchmarks for the broader securities markets.

*** The Federal Reserve currently uses Treasury securities to conduct open
market operations, but it has not always been that way, nor would it have to
be in the future?. As with other market participants, the Federal Reserve
would adapt to such a changing environment by substituting other debt
securities for Treasuries 9 .

The developments in the Treasury markets over the past few years have
demonstrated the pressing need for a new index for FFELP loans.

However, FFELP could not make the transition to a new index unilaterally.
Legislation was enacted late last year to provide that, beginning on January
1, 2000, new FFELP loans be indexed to 90-day commercial paper rates. By
enacting this change, new FFELP loans are on a solid base for the future and
the provision of new private capital to the student loan program has been
greatly enhanced.

In opposing this change, the Administration argued that the change of the
lender rate to commercial paper while maintaining the student rate based on
the Treasury bill would result in unacceptable risk to the government (these
arguments are put forth in Appendix 11). At the same time the Administration
asserted there is no problem with the Treasury bill index (see chapter 3),
it claimed that the risk associated with a different index for lenders and
borrowers would result in significant costs to the U.S. Treasury.

The Administration pointed to the swap markets-markets where the rates for
LIBOR or CP can be exchanged with Treasury rates-to claim that the risk to
the government was unacceptably high. The Administration implies that there
is a significant economic cost to the government by using anecdotal evidence
from two Sallie Mae securitizations. However, it is the inefficiencies and
flaws in the swap markets documented in Chapter 4 and Appendix 8 that are
the very reason that this study group was convened. The findings in GAO?s
survey of lenders support why the swap market is an inappropriate measure
for expected government costs or lender ?value.? The survey found
significant problems with both sides of the swap market, stating that ?since
October 1998, the market for this [T-bill/LIBOR] swap has been very illiquid
and ?gappy.?? It found that there are few buyers in the T-bill/LIBOR swap
market and that there are no sellers in this market either. Appendix 8
states:

9 U.S. Senate Committee on Finance, Hearing on the Nomination of Stuart E.
Eizenstat for Deputy Secretary of the Treasury, Written Response to Question
from Chairman Roth.

Appendix 14 168 there?s no natural counterparty for a T-bill-LIBOR swap (and
no natural investor for variable-rate T-bill-based assets). The market for
T-bill-based securities is thin because of the need to swap into LIBOR. The
price that would be required makes swaps unfeasible?

*** T-bills are also an unnatural index because of special characteristics
of the instrument.

These flaws have only been exacerbated by the recent developments in the
Treasury markets.

As part of the debate, the FFELP community did acknowledge that there would
be some risk of additional costs to the government if the lender index were
changed. However, CBO?s probability scoring captures the expected government
cost associated with the likelihood that the difference between the Treasury
bill and the 90-day commercial paper rate will be larger than forecast.
CBO?s methodology is described at the end of chapter II, which states that,

CBO?s projections also provided measures of possible forecast error that
would lead rates to move around the expected level forecast? Therefore, in
making the budget scoring projections discussed elsewhere in this report,
the uncertainty of the interest rate projections as well as the forecast
levels are taken into account.

CBO forecast the difference between 91-day Treasury bill rates and 90-day
commercial paper rates would be about 0.44 percentage points. However, to
achieve budget neutrality and cover the likely government cost associated
with the volatility between the T-bill and CP rates, the legislation
changing the index last year adjusted the margin over the benchmark by 0.46
percentage points (the index was changed from T-bill plus 2.80 percentage
points to CP plus 2.34 percentage points). CBO estimated that this
legislation would reduce outlays by $25 million from fiscal 2000 to 2003.

Further, the FFELP community has asserted that the cost associated with that
risk of the new index was far less than the Administration asserted. To
demonstrate this, the FFELP community put forward a chart that showed the
potential costs and savings to the government from a lender index based on
90- day commercial paper plus 2.34 percent. This chart puts some perspective
on the costly risk that the Administration asserted the government would
undertake under the new index. The chart-included below-shows that if, in
any year, the difference between 3-month commercial paper and T-bill widen
more than CBO has projected, the federal government will incur some costs.
However, these costs are a mere thousandth of a percent of the costs that
Treasury manages.

Appendix 14 169 By the same measure, if spreads average their historic
levels (such as the average since 1990 or since 1985 or since 1980), lender
returns would be less and costs to the American taxpayer would be reduced.

Indeed, since the index change on January 1, it is the government that has
reaped the savings. Since January 1 to February 25, the CP/T-bill difference
has been 0.43 percentage points, below the 0.46 percentage point difference
which resulted in the new index of CP plus 2.34 percentage points. Further,
the CP/T- bill difference has been trending downward since the beginning of
the year, moving away from the impact of year 2000 issues on the financial
markets. Since mid-January, the average difference has been less than 0.40
percentage points, moving closer to historic averages.

Conclusion As members of the Alternative Indices Study Group, we appreciate
the time and effort put in by all members of the study group and by the
supporting staffs from the General Accounting Office, the Department of
Education, the Congressional Budget Office, the Congressional Research
Service, the Office of Management and Budget, and the Department of
Treasury. Clearly, the awareness of the issue has been heightened by the
strong debate during the study groups deliberations and the consideration of
legislation changing the index. We believe that the events over the past few
years, and increasingly over the past few months, continue to demonstrate
the necessity of changing the index in the student loan program.

Student Loan Lender Rate at CP plus 2.34% Potential Costs/Savings if
CP/T-bill Spread

Differs From CBO Forecast

-$100 -$80

-$60 -$40

-$20 $0

$20 $40

$60 $80

$100 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009

(in millions of dollars)

Government Cost if Spread at 1997-1998 Average Government Savings if Spread
is at 1990's Average

Government savings if spread is at CBO Forecast NOTE: Savings and costs
represent estimated cash impact in each year if difference between CP and
T-bill differ by less or more than CBO forecast in that year. CBO forecasts
the difference between CP and T-bill to be 0.44%; the spread is set at CP
plus 2.34, a difference of 0.46% to cover probability cost. The 1990's
average has been 0.35%; the 1997-98 average was 0.57%.

Appendix 14 170 The changing Treasury markets will continue to pose a
challenge in managing the more than $90 billion in outstanding student FFELP
loans that are still indexed to the 91-day Treasury bill. However, by
assuring the stability of new FFELP loans, the new index allows the FFELP
private sector participants to invest in the future of the student loan
program. The FFELP loan providers will continue to use the innovation and
technological capacity of the private sector to bring more options and new
opportunities to students and parents and schools relying upon FFELP.

Bill Beckmann President and CEO Student Loan Corporation

Kathleen L. Cannon Senior Vice President Bank of America

Jacqueline Daughtry-Miller Vice President Independence Federal Savings Bank

Anthony Dolanski Executive Vice President, Systems & Finance Sallie Mae,
Inc.

Michael H. Hershock President and CEO Pennsylvania Higher Education
Assistance Agency

James C. Lintzenich President and CEO USA Group

Claire J. Mezzanotte Senior Director, Structured Finance, Asset Backed
Securities Fitch IBCA, Inc.

Chalmers Gail Norris Executive Director Utah Higher Education Assistance
Authority

Dick H. Pierce President and CEO Maine Education Services

Appendix 14 171 Paul S. Tone Senior Vice President, Industry & Government
Relations UNIPAC

Paul Wozniak Managing Director PaineWebber Incorporated

Appendix 15 172

APPENDIX 15: TREASURY STATEMENT REGARDING BUDGET SURPLUSES AND IMPLICATIONS
FOR LIQUIDITY

FROM THE OFFICE OF PUBLIC AFFAIRS FOR IMMEDIATE RELEASE

February 2, 2000 LS-365

UNDER SECRETARY OF THE TREASURY FOR DOMESTIC FINANCE GARY GENSLER REMARKS AT
THE FEBRUARY 2000 TREASURY QUARTERLY REFUNDING

Good morning. I am pleased to be with you today to discuss the governments
refunding needs for the current quarter. This month marks the longest
running economic expansion in our nations history. The President announced
on Monday that, by the end of the year, we will have paid down approximately
$300 billion in debt over three years. As our nations debt takes up a
smaller portion of our economy and our financial markets, our continued
fiscal discipline contributes significantly to the health of the economy.

Debt Management Challenges To date, Treasury has managed the declining debt
by refunding our regularly maturing debt with smaller amounts of new debt.
We have accomplished this by two means. First, we have reduced the number of
longer-term debt issuances by one-third, from 39 to 26 auctions per year,
while keeping auction sizes relatively constant. Second, we have cut the
size of our short term bill auctions by almost a quarter, from an average of
almost $20 billion in 1996 to just over $15 billion in 1999, but have
maintained the number of issues.

Fortunately, as budget surpluses continue to diminish our borrowing needs,
we now face additional challenges going forward.

First, debt held by the public is forecast to shrink even further and faster
than it has in the last two years. As we announced on Monday, we estimate
that we will paydown $17 billion in net market borrowing for the
January-March quarter. This will be followed next quarter with the largest
reduction in publicly held debt in our nations history, as we pay down
approximately $152 billion. More significantly, there is now a consensus
among private sector and government forecasters that these paydowns will
grow in the future.

Second, the effect of seven years of fiscal discipline is already showing up
in our maturing debt. There will be a great deal less maturing debt to be
redeemed in the very near future. This fiscal year, $476 billion of coupon
debt will mature,

Appendix 15 173 down from a peak of $510 billion in 1998. Over the next 15
months, the last of the old 7-year and 3-year notes will mature. Thus, by
2002, debt maturing will decline significantly. Debt maturing in 2002 is
likely to be less than $400 billion.

Third, we face the challenge of how to continue to issue sufficient
longer-term debt without an unacceptable lengthening of our maturity
structure. For instance, if we maintain the current level of longer-term
financing (10-year and 30-year debt), the average maturity of Treasury debt
is forecast to lengthen from about 5 3/4 years currently to approximately 8
years by the end of 2004. Over the long term, this would impose an
unnecessary additional cost on the taxpayers to finance our debt.

We have several announcements to make today concerning adjustments we are
making across our debt management program to further address these
challenges.

Reducing Size of Long-Maturity Issues Our first announcement concerns
reductions in the issuance sizes of longer- maturity debt. This reduces our
funding, takes into consideration the longer-term fiscal forecasts, and
helps us manage the average maturity of our debt. In this regard, we plan to
reduce the issuance of 5-year, 10-year and 30-year debt, both fixed rate and
inflation-indexed securities.

At the last quarterly refunding, we announced new rules to facilitate
reopening of our benchmark securities within one year of issuance. We now
will be adopting a regular reopening schedule for our longer term
securities. Our current offering plans are as follows:

New 5-year notes will be offered in May and November, with smaller
reopenings in February and August. The February five-year note therefore
will be a smaller reopening of the November 5-year note.

New 10-year notes will be offered in February and August, with smaller
reopenings in May and November. The May offering of our 10-year notes
therefore will be a reopening of the 10-year notes we issue this quarter.

New 30-year bonds will be offered only in February, with significantly
smaller reopenings in August.

In line with the reductions we are making in our 5- and 10-year notes and
30-year bonds, we also intend to reduce the issuance size of our
inflation-indexed notes and bonds. We started this process last month, when
we reduced the auction size on the 10-year securities from $7 billion to $6
billion. We are now announcing that we plan to auction only one 30-year
inflation-indexed bond, which will be issued in October. There will be no
April issue. In addition, we most likely will make further modest reductions
in the size of the 10-year inflation- indexed note.

Appendix 15 174 Taken together, our aggregate issuance of 30-year debt for
this fiscal year will be less than half what it was in FY1999. We expect
that these changes to our auction schedule will preserve the liquidity of
our 5-, 10- and 30-year securities while reducing the overall size of our
longer term issuances. We will continue to assess the size, frequency, and
issuance of these securities in the future.

Debt Buybacks Last month, Treasury announced the adoption of a final rule
that permits us to conduct buybacks of outstanding Treasury securities prior
to maturity. We will begin using this new debt management tool promptly.

We plan to conduct up to $30 billion of debt buybacks this year, with the
first operations conducted in the next two months. Our initial buyback
operations will be approximately $1 billion each in size and will focus on
the longer-maturity sector. These initial operations will provide an
opportunity for both the market and the Treasury to gain experience with the
reverse auction process prior to more significant operations. After
evaluating our first buyback operations, we will refine our approach to
using buybacks going forward. The use of debt buybacks will help us best
maintain the liquidity of our remaining issues, while also managing the
average maturity of Treasury debt.

Reducing Number of Short Maturity Issues Lastly, we plan to reduce the
issuance of our shorter-maturity securities. Based on the Borrowing Advisory
Committees recommendations, we are reducing the auction frequency of our
one-year bills. These bills currently are auctioned every four weeks. We
will now auction one-year bills only four times each year. The last monthly
auction of the one-year bill will take place on March 2 and the next auction
will then be June 1. This change to our auction schedule will eliminate five
one-year bill issues this fiscal year.

Consistent with the Committees recommendations, we will maintain the regular
monthly auctions of our two-year notes at the present time. We plan,
however, to cut modestly the size of individual auctions of two-year notes.

These changes will enable us to increase the size of our three- and
six-month bill auctions, as well as respond to our reduced borrowing needs.
We will increase the size of weekly bills beginning with the regular auction
announcement tomorrow. It is likely that, as further reductions in issuance
becomes necessary, elimination of the one-year will be considered.

Appendix 15 175 Terms of the February Refunding I will now turn to the terms
of the quarterly refunding. We are offering $32 billion of notes and bonds
to refund $27.6 billion of privately held notes maturing on February 15,
raising approximately $4.4 billion.

The securities are: 1. A reopening of the 5 7/8 % note of November 1999,
maturing on November

15, 2004, in the amount of $12 billion; 2. A 10-year note in the amount of
$10 billion, maturing on February 15, 2010;

and 3. A 30 1/4-year bond in the amount of $10 billion, maturing on May 15,
2030.

These securities are scheduled to be auctioned on a yield basis at 1:00 p.m.
Eastern time on Tuesday, February 8, Wednesday, February 9, and Thursday,
February 10, respectively.

As announced on Monday, January 31, 2000, we estimate that we will have a
$40 billion cash balance on March 31, as well as on June 30. We expect to
issue cash management bills this quarter to bridge seasonal low points in
our cash position.

The next quarterly refunding press will be held May 3, 2000.

GLOSSARY 176

GLOSSARY

Asset-backed securities (ABS): an increasingly common and important source
of funding for many financial instruments such as mortgages, credit cards,
automobile loans, and, more recently, student loans. In a securitization of
loans, the holder of the loans transfers them to a single-purpose,
bankruptcy-remote trust or other entity. To finance the transfer, the trust
or other entity issues its own debt obligations (securities), primarily with
high credit ratings. The pools of loans are used as collateral to create new
securities of specified maturity and interest rate that can be bought and
sold. The creator of the pool guarantees that the interest and principal on
each security will be paid according to schedule. The original lender or
another institution services the loans by collecting loan payments from the
borrowers and passing them on to the creator of the loan pool. In turn, the
creator of the loan pool pays the interest and principal on the asset-backed
securities to the investor who holds the securities.

Basis Point: one one-hundredth of a percent.

Basis risk: the potential for financial loss that can arise through interest
rate changes when the variable rate on the asset is tied to a financial
instrument that is different from the variable rate on the liability that
funds it. In student lending the asset is tied to T-bill while the funding
is often tied to LIBOR or CP rates.

Basis swap: a swap contract has two parties who exchange (swap) interest
rate payments based on a notional value for the principal for the term of
the contract. Both parties pay a variable rate referenced to a market rate,
such as T-bill, LIBOR or CP. One party pays according to one rate while the
other party pays according to a different or rate. Normally the T-bill payer
pays T-bill plus a margin and receives in exchange LIBOR or CP. The margin
over T-bill reflects the spread between T-bills and the other rate as well
as the risks of movements in the spread over the life of the swap contract.

Bid-ask spreads: the difference between the bid price (yield) and ask price
(yield) of a financial instrument. Bid-ask spreads are an indicator of
market liquidity, with narrower bid-ask spreads indicating greater
liquidity.

Commercial paper: short-term unsecured debt issued by many large, well-
known companies, both financial and nonfinancial. Maturities for commercial
paper range up to 270 days. Most often, commercial paper is issued with
maturities of less than 1 or 2 months.

Credit risk: the risk of loss if the borrower does not make payments of
interest and principal on a timely basis. In the case of FFELP loans, this
risk is covered to a great extent by guarantee agencies and federal
reinsurance for student loans, provided that the loans are serviced
properly. See servicing risk.

GLOSSARY 177

Formula spread: the difference between a lender?s income and expenses in its
FFELP activities, which is affected by the formula for lender yield and
interest rates paid for funds. In student loan lending the yield on the
assets usually moves with T-bills but the funding costs move with LIBOR or
CP rates. Movements in the formula spread create basis risk or interest rate
risk for FFELP lenders.

Guaranty agencies: agencies that guarantee FFELP loans up to 98 percent of
principal and accrued interest (100 percent in the case of death,
disability, bankruptcy discharge, closed schools and loans-of-last-resort).
The government reinsures up to 95 percent of the guaranty agencies' risk and
provides them with income streams (loan-processing/issuance,
account-maintenance and default- aversion fees, and collection retention)
and federal reserve funds, including any borrower-paid guaranty fees, to
cover their guaranty obligations and operational expenses. (The government
also directly guarantees FFELP lenders against the inability of guaranty
agencies to fulfill their guarantees because of insolvency.) Guaranty
agencies ensure that servicers? collection efforts are undertaken in
accordance with federal regulations. Use of appropriate collection efforts
protects the federal re-guarantee of student loan values in the event of
default. As of September 1999, The Secretary of Education had designated on
a state or national basis 36 state and private, nonprofit guaranty agencies.

Hedging: a financial strategy that reduces risk by entering contracts with
other financial institutions that result in a reduction in the basis risk,
or the volatility of the formula spread. Normally a hedging strategy
simultaneously lowers the risk and the expected risk at the same time
because the lender must pay the other financial institution to accept the
risk it is shedding.

Lender?s interest expense: a market rate plus a margin.

Lender?s spread: the difference between its yield and interest expense
(interest income minus interest expense).

Lender?s yield: for FFELP lenders, this equals a reference (prior to January
1, 2000, the 91-day Treasury bill rate; now, the 3-month commercial paper
rate) plus a markup, as long as this yield does not exceed the student?s
borrowing rate. The yield is adjusted quarterly. The difference between the
lender yield and the borrower rate is paid by the federal government and
referred to as the "special allowance payment" (SAP).

Liquidity: as a characteristic of a security or a commodity or the market
for that security or commodity, liquidity means that buying and selling can
occur readily at current market prices. Liquid markets or liquid securities
have sufficiently large volume outstanding and sufficiently active trading
that large transactions

GLOSSARY 178 can be made without a substantial movement in price.
Furthermore, because of active trading among investors and/or dealers,
isolated events and or erratic behavior by a single market participant are
unlikely to have major effects on the market price.

Loan servicers: entities that ensure that cash flows of FFELP loans are
recorded and transferred to lenders, guarantee agencies, and themselves in
accordance with loan requirements and other federal regulations. If
servicing is done incorrectly, the lender may not be reimbursed if the
borrower defaults. Except for Sallie Mae and a few of the largest student
loan lenders, legal holders of student loans tend to arrange for some or all
of the FFELP program loan origination and/or maintenance functions to be
performed by a third-party servicer. Some large holders of student loans
provide loan servicing to other holders that may or may not be selling loans
to the large holders.

London interbank offer rate (LIBOR): the reference rate on U.S. dollar-
denominated interbank lending in London. Hence, it is the interest rate on
dollar- denominated offshore loans from banks with temporary ?excess? funds
to banks facing strong demands for funds, either for dollar-denominated
loans overseas or from their home offices. These interbank loans are for a
fixed term and are made in large denominations. LIBOR also serves as a
reference rate for a number of other transactions.

Management risk: the risk that management will be unable to manage the firm
or the state guarantee agency. The mismanagement of the firm can lead to
bankruptcy and missed payments to investors. The mismanagement of a state
secondary market entity can lead to a state takeover or even closure of a
firm.

Margin: a market-based add-on to a market rate (a margin can be negative and
it changes based on market conditions and the creditworthiness of the
borrower)

Markup: for purposes of this report, an amount, determined by the law, added
to an established reference rate (such as a Treasury bill rate) to determine
a new rate.

Market rate: interest rates that are used by market participants to
determine contract requirements because these rates are considered to be on
liquid instruments and therefore truly reflect market conditions. In this
report we often refers to three market rates: T-bills, LIBOR, and CP.

Prepayment risk: the risk that student will pay off loans more quickly than
expected.

Reference rate: an interest rate used, or referred to, in a formula for
calculating another rate. For example, the interest rate on 91-day Treasury
bills serves as a

GLOSSARY 179 reference rate in determining the interest paid by borrowers of
FFELP loans and in the formula for determining the lender yield on a FFELP
loan.

Reinsurance: insurance provided to the guaranty agency by Treasury.

Regulatory risk: the potential for financial loss due to changes in law or
regulations, also known as political risk. Regulatory or political risk
exists because the returns and risk in the student loan program depend not
only on the fundamental economic and market risks but also on legal and
regulatory changes created by the federal and state governments. Over the
years, the FFELP has undergone a variety of such changes, including changes
in the interest rate subsidies and regulatory rules dealing with guarantee
fees and payments in the event of student borrower default.

Secondary market: for purposes of this report, an institution that buys
loans from originators.

Servicing risk: the risk of financial loss due to poor servicing of a loan.
Improper servicing can void the federal guarantee of student loans.

Special allowance payment (SAP): a payment that the federal government makes
to FFELP lenders, which is designed to compensate lenders for within- year
fluctuations in interest rates. This payment, calculated quarterly, is the
lender yield minus the student borrower?s rate. If this difference is
negative, the SAP is zero.

Spot rate: the rate of interest or price being currently charged.

Spread risk: for purposes of this report, the potential for a FFELP lender?s
financial loss resulting from differences between the interest rate the
lender pays for funding FFELP loans and the legislated special-allowance
reference rate of 91-day Treasury bills. In such cases, the lender's net
interest income will vary depending on the current spreads between the
91-day Treasury bill rate and the interest rates in which it actually funds.
For example, if LIBOR was 2.8 percent above the 91-day Treasury bill rate,
and a lender financed exclusively at LIBOR, the lender would not have any
net interest income.

State designated secondary market: institutions that purchase loans from
originators and must originate student loans for borrowers who cannot get a
loan elsewhere.

Student borrower?s rate: in the FFELP, equals a reference rate (91-day
Treasury bill) plus a markup (currently 1.7 percentage points while the
borrower is in school or in certain post-schooling periods not requiring
repayment, and 2.3

GLOSSARY 180 percentage points while the borrower is in repayment), with a
cap of 8.25 percent. This rate is adjusted annually.

Swap: an agreement between counterparties to make periodic payments to each
other for a specified period. In a simple interest rate swap, one party
makes payments based on a fixed interest rate, and the counterparty makes
payments based on a variable rate. A FFELP lender can arrange for another
party to assume the spread risk (and opportunity) associated with the
student loans by purchasing an interest rate swap. However, such a purchase
will reduce the overall profitability of the lender?s student loan business.

Swap Spread: the markup over the T-bill rate paid by firms swapping aT-bill
rate for a commercial rate

Treasury-Eurodollar (TED) spread: the difference between the rate on 91-day
Treasury bills and the London interbank offer rate (LIBOR). This difference
is generally determined through a comparison of the monthly average of
91-day Treasury bill rates and the monthly averages of LIBOR. See formula
spread and

lenders? spread in this glossary.

Treasury Bill (T-bill): A noninterest-bearing obligation of the U.S.
government, payable to the bearer, maturing in less than a year from the
date of issue. The shortest-term, regularly issued Treasury securities are
bills with initial maturity of 91 days, the maturity relevant for FFELP.
Treasury bill rates serve as the reference rate in determinations of
interest paid by borrowers and in the formula for lender yield.
*** End of document. ***