Loan Guarantees by the Federal Government

Published by the Government Accountability Office on 1977-03-29.

Below is a raw (and likely hideous) rendition of the original report. (PDF)

                          DOCUMENT RESUME
00322 - [A1051843]

(Loan   Guarantees by the Federal Government]. March 29, 1977. 15
Testimcny before the House Committee on Banking, Finance and
Urban Affairs: Economic Stabilization Subcommittee; by Harry S.
Havens, Director. Prograr Analysis Div.
Issue Area: Accounting and Financial Reporting (2800).
Contact: Program Analysis Div.
Budget Function: General Government: Central Fiscal Operations
Organization Concerned: Office of Management and Budget.
Congressional Relevance: House Committee on Eanking, Finance and
    Urban Affairs: Economic Stabilization Subcommittee.

         Loan guarantees constitute a major portion of the
Federal credit assistance program, representirg an investment of
about $200 billion. The effects of loan guarantees are
far-reaching: (1) they confer large benefits on credit
recipients; (2) they impose large budgetary costs on the Federal
Government, and therefore, on taxpayers; and (3) they impose
indirect costs on firms and individuals who are not directly
involved with the programs. Like other subsidy programs, the
cost effectiveness of loan guarantees should be evaluated to
determine whether they are meeting their objectives in an
efficient manner. The preliminary findings of an ongoing study
at the General Accounting Office show that, although loan
guarantee programs have expanded rapidly in recent years, their
cost effectiveness has not been carefully evaluated. For 1975,
losses on guaraneed loan prcgramr were about $2.3. billion;
benefits to borrowers due to interest rate reduction were about
$2.6 billion. Evaluating a proposed loan guarantee program
requires answers to three questions: (1) Is it desirable to
stimulate investment in the sector tc which the guarantee would
be directed? (2) Is a loan guarantee the most appropriate form
of subsidy? and (3) How can the program be designed to operate
most efficiently? The budget authority needed to make good on
the guarantees should be explicit in the legislation and
adequate to provide the reserves necessary tc carry out the
planned level of activity. The budget should, of course, be
subject to full disclosure and to executive and congressional
reviews. (LDB)

                              WASHINGTON, D.C.   20548

                       Statement of Harry S. Havens, Director
                              Program Analysis Division
                           U.S. General Accounting Office

                   Before the Subcommittee on Economic Stabilization
                                        of the
      House of Rep-esentatives' Committee on Banking, 7inance, and Urban Affairs

                                                          For Release on Delivery
                                                          Expected at 10:00 a.m.
                                                          March 29, 1977
    Mr. Chairman and Members of the Committee.      Thank you for the

opportunity to testify before your Subcommittee on loan guarantees       v

the Federal Government.   Loan guarantees are a major portion of the wide

array of Federal credit assistance programs--the total of outstanding

guaranteed loans is about $200 billion.

     According to OMB's Special Analyses of the Budget, the principal

objective of Federal credit programs is to "encourage certain types of

economic activity by providing individuals, businesses, and government

bodies with credit at more favorable terms than would otherwise be availa-
ble in the private market."

     In fact, Federal loan guarantees do much more than this.        They:

     -- confer large benefits on credit recipients;

     -- impose large budgetary costs on the Federal Government

          and, therefore, taxpayers; and

          impose indirect costs on firms anr   individuals who are

          not directly involved with the program.

     Federal credit programs are subsidy programs.

     Like other subsidy programs, their cost-effectiveness should be

evaluated to determine whether they are meeting their objectives, how

much these objectives are worth to society, and how much the programs

cost.     Objective evaluation might, in some cases, show the advantages of

alternative policies--direct subsidy, price guarantees, or even no program

at all.

        My testimony today deals mainly with loan guarantees and is based on

 a study which is now underway in GAO.     Because the study is not yet complete,

 these remarks should be considered preliminary.
    Loan guarantee programs have expanded rapidly in recent years.

Nevertheless, the evaluation questions I have mentioned have not been

answered in any systematic fashion.

     In my remarks today, I will emphasize some of the problems of

congressional control over loan guarantees.   My object is not to describe

the many guarantees now outstanding, but rather to point out areas which

we believe should be of concern to the Congress.    These affect both

existing programs and new proposals which will undoubtedly be considered

during this session of Congress.

     There is great diversity among the numerous loan guarantee programs.

     --   They range from hundred-million dollar loans to large

          corporations down to loans of several hundred dollars

          to college students.

     -- Some guarantee the full amount of the loan, some only

          part of it.

     -- Some loans are backed by eollateral, some are not.

     -- Some loans are individually negotiated, some are almost

          entitlement programs.

     -- Some programs show slight profit- to the government, others

          lose money.

     -- The amount of control over the money by the government varies

          widely from program to program.

     In our analysis of loan guarantees, we have compared features of

the various    -rorams in order to develop a means for determining when

guarantees are the appropriate form of subsidy and when some other form

of aid might be more effecti7e.

                                   - 2-

     Different guarantee programs involve different degrees of subsidy,

and costs to the government.   The most effective programs are those that

correct some imperfection in the market or lead to the creation of a

market where one did not exist.    Such programs typically serve many

relatively small borrowers.    For example, FHA loans back in the 1930's

filled a certain gap in the credit market, perhaps caused by a reluctance

of banks to accept the risks of long-term home mortgages.     Whether these

banks overestimated the risks or were simply adverse to taking risks is

not clear.   The point was that by pooling the risks, the FHA created a

new credit market, and many homes were financed by a self-supporting

government program.

     In later years, credit assistance programs began to include a

greater degree of actual subsidy from the government.    These programs--in

the aggregate--are currently running at a loss.

     By our estimates, for 1975, losses on guaranteed loan programs were

about 1.4 percent of loans outstanding--about $2.3 billion.     For direct

loans, losses were about 2.3 percent of loans outstanding--about $1.7


     We have also estimated the benefits, or subsidy element, to borrowers

due to interest rate reduction.    For guaranteed loans in 1975, it was

about 1.6 percent, or $2.6 billion.    For direct loans, the benefit or

subsidy figure was about 5.4 percent, or about $4.0 billion.     These are

preliminary and rough estimates.

     The Office of Management and Budget, in its Special Analyses of the

Budget, also estimates the subsidy element, but by a different approach.

It calculates the present value of the subsidy element in new loans and

guarantees during a given year, i.e., how much the subsidy will be worth

over the life of the loan.   Our figures are for the subsidy element

during a year for all loans and guarantees then outstanding.    These two

approaches are not at all in conflict--they are different ways of looking

at the subsidy.   OMB's estimates of the subsidy element exceed ours,

because they are looking at the future of a growing list of programs,

while we look at only one particular year of exi    i!ng programs.

     These figures emphasize my point that credit assistance programs

are subsidy programs, and can be costly ones at that.

    Although our subsidy figures exceed our cost figures, it shou.id not

be concluded that these programs have delivered a "free lunch."      Our

figures do not include other costs of loan guarantees such as the cost

of using up, or "crowding out," credit that would otherwi3s have been

available on the private sector or increase interest costs elsewhere in

the economy.

     Loan guarantees have generally escaped the discipline of the budget

process, and they have, by and large, escaped the discipline of program

evaluation.    Regardless of how loan guarantees are scored in the budget,

there is a growing need for information on how they are working and what

they are accomplishing.

     Recently, loan guarantee proposals have moved away from the concept

of pooling risk for large numbers of relatively small borrowers.      Increas-

ingly, they have been used as a means to encourage single large ventures.

The subsidy element, due to risk reduction and interest rate reduction,

                                    - 4-
has been viewed as a policy instrument, quite apart from any consideration

of whether there is any particular imperfection in the market.

      How does the subsidy element come into being?   The rate of interest

charged by a private lender includes a "risk premium," which is the

excess interest over what would be charged to the least risky borrower.

The riskier the borrower, the larger the risk premium required by the

lender, and the higher the interest rate the borrower would have to pay

in the market.

      Loan guarantees lower the risk borne by the lender, thereby inducing

the lender to drop the risk premium.     The borrower benefits by the reduc-

tion in interest, as well as by other features of the guarantee which

may reduce the risk to the borrower.     Due to the reduced interest and

other advantages of a guarantee, the borrower zan be viewed as willing

t-   pay a certain amount of money for the privilege of having his loan

guaranteed.   The benefit to the recipient does not correspond directly

to a government cost.    In fact, the benefits (to which we refer as

 the "subsidy element") can exceed the cost borne by the government.

       The government may recover part of this subsidy by charging a

 loan guarantee fee to the borrower.   This fee may be high enough to cover

 expected losses, but still leave room for a subsidy.    Such was the case

 in early FHA mortgage guarantees, when the risk premium charged by the

 market may have been needlessly high.

       The magnitude of the subsidy clearly depends upon the riskiness of

 the borrower--this is the peculiarity of the loan guarantee as a subsidy.

 The riskier the borrower appears in the eyes rf the lender, the more

 valuable the guarantee.
    This feature makes loan guarantees a powerful policy instrument in

some circumstances, but ineffective in others.    For example, if the govern-

ment wished to induce General Motors to build smaller cars, offering a

loan guarantee would probably have no effect because the guarantee would

have little or no effect on the rate of interest at which GM could borrow.

     If the borrower ig in precarious financial condition, or if the

project is risky, then loan guarantees could be expected to have fairly

significant effects on the rate of interest.     If the program in question

is open to all applicants, then there will be a tendency for the more

risky firms or individuals to apply.

     This logic is applicable to such. programs as guaranteed student

loans.     The "riskier" studencs--insofar as they could be identified in

advance by lenders--received treater benefits through reduced interest

costs.     Needless to say, additional savings accrued to those who could

not or did not repay the loans.

     Despite the problem with existing programs, student loans are a prime

example of how loan guarantees can solve problems in certain credit

markets.     Education--viewed as investment in human capital-is difficult

to finance in existing capital markets, mainly because there is no tangi-

ble asset to be used as collateral.

     However convincing the rationale for a guaranteed loan program, it

must be administered properly.     Under the student loan program, the

Office of Education has paid out more than $280 million to lending insti-

tutions because about one of every six loans had been defaulted after

the student had completed or withdrew from school.     Of this, OE has

collected only about $25.1 million, and the collection progran is not

keeping pace with defaulted loans.

     Surely the degree of effort which is put into enforcing repayment

has some effect upon the number of defaults.     If an agency is known to

be lax in recovering its defaulted loans, there is little incentive to


     There is nothing inherently wrorg with loan guarantees.     They can

be applied in the wrong circumstances, they can be poorly designed, and

they can be poorly administered.     On the other hand, loan guarantees can

be useful, particularly if they are well designeq and well administered_

Good evaluation can help assure that they are used aprropriately.


     What is the future of loan guarantees?     Even though they owe much of

their popularity to-their off-budget status, it is clear they are here

to stay.   Regardless of how loan guarantees are treated in the budget, we

need to develop evaluation techniques for assuring that they are used

effectively when they are appropriate.     In fact, more rigorous budget

treatment would highlight the need for evaluation of these commitments

of public funds.

     Evaluating a proposed loan guarantee program should consist of

answering three questions.   First, is it desirable to stimulate investment

in the sector or to reach a group in society to which the guarantee would

be directed?   Second, is a loan guarantee an appropriate form of subsidy?

Third, how can the program be designed to operate most efficiently?        After

the program is established, periodic review and evaluation is necessary.

                                   - 7-
When zero-based budgeting is used, it should be applied just as rigorously

to loan guarantees as to on-budget programs.

Appropriateness of the Program

     The first questions that should be asked about any Federal program

are whether its objectives serve a useful function of government and

where those objectives rank in the hierarchy of public priorities.     Although

these questions involve value judgments, some economic considerations

are also involved.

     In particular, one should consider why the objectives of the program

are not being met through the market mechanism and whether any subsidy

could correct the problem.    That is, will a subsidy actually alter the

state of affairs?     It seems reasorable to conclude, for example, that

without a subsidy, few commercial ships would be built in the United

States because of high construction costs,     Other times the riskiness of

a loan means that lenders will not make funds available at a reasonable

price, such as to students and small busin, sas.     Or. the other hand, we

should be alert to the possibility that a subsidy program merely rewards

people for what they would do any way or, alternatively, is too small to

influence their behavior.

     We must also be alert to the side-effects of a program.     For example,

if the economy is near full capacity, one particular investment may be

obtainable only if another is given up.     Thus, there is a tradeoff between

the subsidized investment and the one %        is being sacrificed.   Unfor-

tunately, these side-effects are very difficult to quantify even though

we know they exist.

                                     - 8-
Type of Program

     Once it has been decided that a subsidy is appropriate to achieve a

particular objective, attention should focus on the way in which the sub-

sidy is to be provided.    Loan guarantees are one possible approach, but

there are many alternatives--direct subsidies, tax credits for investment,

price supports, direct government ownership, tariff protection, and so


     The basic question is how to achieve program objectives most effectively,

at the lowest cost to the government and with the least disruption to the
private sector.

     In evaluating loan guarantees, we must distinguish the mechanism from

the ultimate objective.    A loan guarantee will make funds available to the

borrower at lower interest rates.    But this is a means, not an end.     The

final desired impact might be to increase consumption of some good, such

as housing, or to increase domestic production of some goods, such as

ships or energy.   In some cases, the objective may be simply to transfer

money to recipients through interest rate reduction.     The underlying

quesuion is:   How effective and efficient is the mechanism of loan guaran-

 tees in achieving the ultimate goals of the program?

     Loan guarantees are at their best when there is some element of risk

 that has been overestimated in the private market.     Again, we come back

 to the old FHA experience, when loan guarantees helped make a new market

 in long-term mortgages.    But capital markets are much more sophisticated

 than they used to be, and it is not clear whether there are still many

 cases where financial markets overestimate riskiness.     If there were,

it is not clear that the government would necessarily evaluate riskiness

more accurately than the private market.

     Even if the market has properly assessed risk, loan guarantees can

he ar effective policy instrument.   They reduce interest costs to the

borrower, because they eliminate the risk premium, and may not cost the

government anything.   There would still be a kind of "crowding out" cost

born by other private borrowers who would face higher interest costs.

This cost is important, but here we are comparing loan guarantees with

other forms of subsidy, such as direct cash grants, which will have their

own "crowding out" effects.   The burden of loan guarantees probably falls

more heavily on the investment sector of the economy, whereas other

subsidies--financed by taxes--are likely to be more of a burden   on


     The effectiveness of loan guarantees in accomplishing their ultimate

objectives is likely to turn on the effect of interest rate reduction

on the investment behavior of the recipient--a much disputed question

among economists.   It is widely believed that lower interest rates

stimulate investment, but little is known about the magnitude of the

effect.   Incidentally, this question also applies to assessing the effect

of other forms of subsidy, such as the investment tax credit.

     Although loan guarantees are subsidies and do incur costs, they may

be, in some cases, relatively efficient forms of subsidy.   Consider a

guarantee-that is worth $100 to the recipient.   In other words, suppcse

that the guarantee had lowered the interest rate and r duced risk to the

extent theat the recipient would have been willing to pay $100 in order to

                                  - 10 -
                                                  It is quite possible that
have his loan guaranteed by the government.
                                          less than $100 in administration,
this guarantee would cost the government
                                         out this way, a subsidy worth
defaults, and other costs. If it works
                                          than $100. A direct subsidy
$100 has been provided at a cost of less
                                         cost the full $100. It is in
of the same value would presumably have
                                        e-irc.mstances, be more efficient
this sense that guarantees may, in some

than direct subsidies.
                                             adopt, the choice between
     In considering what type of program to
                                            discussion. Di,:ect loans are
direct loans and guaranteed loans deserves
                                       mainly because the government does
usually made at lox:er interest rates,
                                          Even though the guarantee removes
 not usually attempt to make a profit.
                                       pay more than the Federal Government
 risk, the private lender usually must
                                         and he also tries to make a
 to obtain the funds in the first place,
                                         interest rates, there are good
 profit. Though direct loans have lower
                                       compete too extensively in the
 reasons why the government should not

 banking business.
                                         is not determined by the market.
      The interest rate for direct loans
                                         more of a subsidy than is actually
 As a result, the government might grant
                                      low. With guarantees, the borrower
 needed by setting interest rates too

                                       terms he can fiud.
 has an incentive to seek out the best

  Design of Loan Guarantees
                                                  written in 1955, "The
       According to the hoover Commission report,
                                              careful the government must
  government is not a canny lender." However
                                            in guaranteeing loans, for the
  be in lending, it must be equally careful

  risk is essentially the same.

                                    -    11   -
     In the private sector, the lender decides upon the provisions of the

loan after judging both the borrower and the project.   In a guaranteed

loan program, however, the lender sometimes bears nont of the risk and

thus has very little incentive to guard against default.   When the

government assumes all the risk, it also must assume responsibility

for choosing among applicants and for judging the acceptability of the


     A fundamental decision must be mada at the outset--is the program

intended to be self-supporting or does congress intend an additional

subsidy in the form of losses on defau]ts?   For programs meant to be

self-suprorting, the loan guarantee fee theoretically can be set high

enough to cover expected losses.   If the program involves a large number

of relatively small loans, it is often possible to estimate future

losses through statistical analysis.   It is much more difficult to predict

the default rate for a program involving a small number of large loans.

Statistical estimating techniques generally cannot be applied in these

 situations and we must depend much more heavily on individual judgment.
                                                                 lie first
      There are two main mechanisms available to control risk.

 is to make the loan recipient subject to certain operating restrictions.
 For firms receiving loan guarantees, these could include such things

 limiting dividends and additional investments, requiring purchase of

 insurance or requiring operation within certain constraints on financial

 statement data.   For example, recipients of Maritime Administration loan

 guarantees are initially classiied as weak or strong firms depending

 on certain financial ratios.   As long as these ratios are maintained,

 firms must meet certain "positive covenants."   If the financial requirements

                                    - 12 -
                                                        on the firm.
are not met, additional "negative covenants" are placed
                                                          recourse to
     The second mechanism for controlling risk is through
                                                     Full recourse loans
the assets of the borrower in the event of default.
                                                       value of the loan
are those in which the government can recover the full
                                                      defaulted borrower.
guarantee (except in the case of bankruptcy) from the
                                                               loan. A
The government recovers nothing from default on a non-recourse
                                                            some part
partial recourse loan means that the government may recover
                                                   In the partial recourse
of the value of the loan in the event of default.
                                                     the loan was made.
case, the recourse is usually to the asset for which
                                                          value of the
     Non-recourse and partial recourse loans increase the
                                                       borrower's risks.
loan to the borrower, because limits are placed on the
                                                       costs) increase.
Simultaneously, of course, the governaent's risks (and
                                                          when the
Partial recourse loans might be appropriate, for example,
                                                       uncertain technology,
government wishes to encourage investment in a new and
                                                        the venture.      Of
but firms are unwilling to gamble their other assets on
                                                          this hurdle.
 course, loan guarantees are not Lhe only way to overcome
                                                          than depending
 Government might undertake the investment itself, rather

 on private industry to do so.   Since non-recourse or partial-recourse
                                                        government, consi-
 reduces risk to the borrower, and may be costly to the
 deration should be~g.ven to charging a fee for this extra
                                                                 If the
      How large a fraction of the asset should the loan be for?
                                                          will loan 95
 market would loan 80 percent, but the government program
                                                        government later on.
 percent, the subsidy may well show up as a loss to the
                                                         then the borrower
 If the value of the asset drops by more than 5 percent,
                                                         action, especially
 may find that default is a perfectly rational course of
  if the government has no recourse to the borrower's other

                                   - 13 -
Administration of Loan Guarantees

     In designing a loan guarantee program, careful thought should be

given to who will assess the merits of each loan, and how much flexibility

there will be in setting loan terms.    The budget authority needed to make

good on the guarantees should be explicit in the legislation and adequate

to provide the reserves necessary to carry out the planned level of

activity.   The budget, of course, should be subject to full disclosure

and to executive and congressional reviews.

     Plans should also be made for allocating credit when the demand for

guarantees exceeds the available authority.   From the standpoint of program

effectiveness, the best criteria would probably be compatibility with

long-run objectives of the program if this can be measured.    Alternatives

include allocating on the basis of need, risk, first-come-first-served,

random, or allocation through a market mechanism.     The market mechanism

often provides appropriate and efficient incentives, and we believe that

this approach should receive careful consideration.

     By a market mechanism, we mean allowing the price for the guarantee

to be established in a "market for guarantees", thus linking the cost of

the guarantees to their value to the recipient.   Let me give an example:

When it is feasible, guarantees could be granted on the basis of competi-

tive bidding on certain features of the guarantee, such as equity partici-

pation by the firm, the guarantee fee, the amount of the assets pledged

as collateral in case of default, or the portion of the loan that is

guaranteed.   By making prospective borrowers compete for terms, the

expected cost and risk to the government can be reduced.     Furthermore,

                                    - 14 -
by increasing the equity participation of the firm and by leaving part

of the loan uninsured, both the borrower and the lender have a greater

incentive to manage the venture efficiently so as to avoid default.   Such

competition would result in a higher equity/debt ratio for the borrower,

which reduces risks to the government as compared to a more highly

leveraged project.

     Such bidding procedures would not be appropriate in all cases, but

they might be useful, for example, when the government wants to encourage

some large project, wants to minimize the subsidy element of the guarantee,

is uncertain of how private borrowers and lenders view the risks, and

wants to draw the more efficient firms into the project.

     Mr. Chairman, this completes my prepared statement.   We would be

happy to respond to any questions.

                                     15 -