: a’GAO Testimony I llwlwl 141105 For Release Need for Improved Budgetary Controls 06 Delhery Over Federal Credit' Programs Expected at 10:00 a.m. Wednesday April 11, 1990 .. Statement of Charles A. Bowsher Comptroller General of the United States Before the Task Force on Urgent Fiscal Issues Committee on the Budget House of Representatives CWU5,\ ILffi’405 GAO/T-AFMD-90-18 I GAO Form 160 (12/W) . Mr. Chairman' and Members'of the Task Force: I am pleased to be here to present our views on the need for improved budgetary controls over federal credit programs,. I would like to emphasize at the outset that the need for credit budgeting reform is a matter on which the Budget Committees, the Office of Management and Budget (OMB), the Congressional Budget Office (CBO), and the General Accounting Office .(GAO) are in substantial agreement. The issues have been debated and analyzed for many years. Rarely will you find a matter on which there is such broad agreement among those who have studied it. I believe the time has come to move beyond discussion to the enactment of legislation needed to implement reform. There are some differences in the.details of the reform proposals of these organizations, and in my statement today I will outline these variations as well as the common features. There is another point I will make, and that is the need to have a broad perspective on the problem--one that is not restricted simpl,y to direct loans and loan guarantees. 'The government 'also needs to consider the implications for the budget of federal insurance programs and the activities of government-sponsored enterprises (GSEs). Our testimony last week on the Federal Savings and Loan Insurance Corporation (FSLIC) underscored this point.1 b 1Resolving the Savings and Loan Crisis: Billions More and "1. Additiona Re orms Nee e . But first, let me put the issue in context, drawing upon a report we recently issued.2 GROWTH IN FEDERAL CREDIT AND INSURANCE PROGRAMS Recently, FSLIC, the Federal Housing Administration (FHA), and the Department of Education suffered enormous losses due to financial institution insolvencies, foreclosures and related losses on federal housing loans and, loan guarantees. These losses are adding billions of dollars to'the deficit. yet' these are only three examples.of the extensive credit assistance an-d insurance programs that the federal government, the nation's largest financial institution, is i,nvolved,with. These programs can be divided into four broad categories: direct loans, loan guarantees, insurance commitments, and GSE loans. As seen in the table,below, there has been a dramatic growth in all of the categories since fiscal year 1965. The total of outstanding amounts 'has'increased over 1,000 percent and currently stands at almost $6 trillion. Even a'llowing for the ,growth in the size of the economy, this percentage increase in outstanding amounts has been enormous. It has been largely driven, however, by the insurance and GSE activities. 2Fedsal Credit and Insurance: Programs May Require Increased Federal Assistance in the Future (GAO/AFMD-90-11, November 16, n89). 2 Reported Outstanding Amounts of Individual Credit and Insurance Programs Since Fiscal Year 1965 Dollars in billions Percent increase Program 1965 1985' '1989 1965 to 1989 Direct loans $ 33 $ 257 $ 207 527 Loan guarantees 91 410 588 546 Insurance commitments 299 2,852 4,213a 1,309 Government-Sponsored ' Enterprise loans 15 ,370 763 4,987 Total $438 $3,889 $5,771 1,218 aOMB redefined insurance commitments for fiscal year 1989. This makes the 1989 figure not strictly comparable to the figures for 1965 and 1985. However, the change appears to affect relatively small amounts. I note that direct loan amounts declined from 1985 to 1989, partly reflecting a recent shift away from such loans to loan guarantees because of the latter's lower short-term outlays effects. Loan guarantees do not result in budget outlays until a borrower defaults, while direct loan disbursements (net of repayments) result in outlays when the loan's are made, immediately increasing the reported budget deficit. Therefore, shifting from direct loans to guaranteed loans reduces current outlays and, correspondingly, the reported deficit. However, this shift does not necessarily represent a savings. If the guaranteed loans default in the future, the government will have u 3 to pay for the cost of the defaults,,which, in turn, will increase the deficit at that time. The table also shows a sharp increase in the insurance commitments of the government. Federal insurance activities include such programs as those of FSLIC, the .Federal Deposit Insurance Corporation (FDIC), and the Pension 'Benefit GuaraMy Corporation ,(PBGC). The 1965 to 1989 increase in insurance commitments grew by a little over 1,000 percent. I mention these insurance amounts even though such programs are usually not covered by credit budgeting reform proposals. The proposals normally cover only the direct loans and loan guarantees of federal agencies. ; I think that the Congress should not overlook the similarities between the credit and insurance programs of federal agencies. Insurance programs represent formal contingent liabilities of the government,, like loan guarantees, and therefore pose the same basic kinds of budget reporting and control issues. ,The most dramatic grqwth'since 1965 has been in theeloans extended by GSEs. GSEs are government-chartered, privately-owned entities whose activities include making loans to expand credit availability to students, farmers, prospective home buyers, and other segments of our society. They are off-budget. Their loans outstanding increased almost 5,000 percent over the fiscal years 1965*to 1989 period. I mention these ‘GSE amounts because it is 4 time that we start considering more seriously their potential budgetary implications. The loans 'of a GSE, if not repaid, could trigger a federal financial bailout under a federal "moral obligation" commitment. FEDERAL CREDIT AND INSURANCE PROGRAMS EXPOSE THE GOVERNMENTTO SIGNIFICANT LOSSES The significance of these g,rowth trends in credit and insurance amounts is that they increase substantially the risk of large budgetary losses to the federal government. A nearly $6 trillion exposure is a huge amount. ,While it should be emphasized that the government will probably experience losses on only a small percentage of this total exposure, the risk of very substantial losses is real. It should therefore be a matter of concern that there are some disturbing developments or trends in the loss patterns of these loan, guarantee, insurance, and GSE programs. Direct Loans There has been a significant increase in direct loan write- offs. Between fiscal years 1985 and 1989 loans receivable annual write-offs increased from about $1 billion to about $4 billion. Despite the write-offs, loan delinquencies continued to incrtase in the same 4-year period, going from about $15 billion to about $27 billion, an 80 percent increase. 5 These write-offs plus the interest payments made by the Treasury in financing the loans are the main components of the government's net costs in making the loans. A key point is that such costs are not properly.reflected in budget estimates. Therefore, when the Congress considers new loan levels in its budget process, it does not routinely consider the ult.imate costs of the programs involved. As mentioned earlier, there has been a shift away from new loans to loan guarantees', where there are also matters of concern. Loan Guarantees For one thing, the newer loan guarantees pose greater risks than the ones made in earlier years. Many new guarantee Programs, such as those for education and energy loans, involve loans with little'or no marketable property as security. This means higher net costs to the government when defaults occur under such guarantees, This is a change from the midl1960s when about 93 percent of guarantees were for housing related loans backed by liens on marketable property. This makes particularly disturbing the recent trends in guarantees and terminations for default. The government's outstanding guaranteed loans increased 62 percent, from $364 .billion to $588 billion from 1983 to 1989. Duriffg the same time,, guaranteed loan terminations for default 6 increased even more sharply, from about.$S, billion to .about $11 billion, a 138 percent increase. Even on the loan guarantee programs backed by marketable property, such as the federal housing program, the g,overnment has experienced significant losses. In recent testimony3 we stated that in fiscal year 1988 FHA's four mortgage funds, which make up the largest federal housing credit program, incurred a loss of $4.2 billion and together have an equity deficit of $2.9 billion. We noted that substantial appropriations would be required to - restore solvency to the'Genera1 'Insurance Fund component. It is important to remember that the projected costs of these guarantees are not systematically considered in the budget process and not appropriated for. What about the insurance area? The risks to the government today are certainly more evident than in years past. Only 3 years ago, some savings and loan industry representatives spoke of $5 billion being needed to recapitalize,the weakened FSLIC. Today, there is a much grimmer picture. Last week, when we testified on our audit of,FSLIC's final 'financial statements, we stated that at least $325 billion would be needed from all sources to pay off FSLIC's obligations, much of which will come 3198% Financial Audit: Federal Housing Administration (GAO/T- AFMD-89-17, September 27, 1989). 7 from the U.S. Treasury. The $325 billion could easily go to $400 billion, or even to a half trillion dollars if the economy turns against us. Furthermore, most of the taxpayer amounts that will be paid out b,y Treasury have not been recognized to date in budgetary totals. I should also mention FDIC., Our audit last year of the Corporation's 1988 and 1987 financial statements4 showed a weakened fund, mainly reflecting banking problems in' the 'energy, agriculture, and real estate sectors of the economy. A record level of bank failures resulted in the Corporation experiencing in 1988 its first net loss since, inception--$4.2 billion. Its fund balance dropped to about $14 billion,'producing the lowest ever ratio of the balance to insured deposits. We are currently conducting a new financial audit of FDIC and are concerned by the threat to the fund of some money center and other large banks. Failure of several,of the larger banks could bankrupt the fund, just as FSLIC was bankrupted a couple of years ago. There is also the Pension Benefit Guaranty Corporation (PBGC) and its approximately $800 billion in outstanding commitments. I am particularly worried about the government's 4Fin&cial Audit: Federal Deposit Insurance Corporation's 1988 and 1981 Financial Statements (GAO/AFMD-89-63, April 28 I 1989 1 l 8 exposure in those instances where there are large, financially troubled companies with underfunded plans. While the government col.lects insurance premiums on most of its insurance programs', it is clear that the level of premiums is insufficient to cover costs in many cases. Already, losses are occurring, and they could grow significantly if the economy' undergoes a downturn. This brings us to the GSEs. Very little of the approximately $776 billion in GSE debt and outstanding mortgage- backed securities is expressly guaranteed by the government. Nonetheless, the financial markets treat most of these GSE instruments as "agency debt" under a perceived implicit federal guarantee. Accordingly, the GSEs are able to market their issuances at favorable interest rates. That this assumed federal backing can become an explicit federal commitment was demonstrated in 1987 when federal legislation wa,s enacted providing billions in payments and explicit loan guarantees to the failing Farm Credit System, a GSE. Such a result is not a foregone conclusion. It 2s worth noting that another GSE, the Federal National Mortgage Association, ran into a serious interest rate squeeze in the early 1980s but managed to work out of it without federal P assistance. Nevertheless, the farm credit crisis and impact on 9 the federal budget was one of the reasons that the 1989 Financial Institutions Reform, Recovery, and Enforcement Act (FIRREA) required GAO and the Treasury to study the safety and soundness of GSEs. Our first report under this FIRREA requirement will,be issued shortly. Given these cred,it, insurance, and GSE problems, what should the government do? My remaining remarks will mainly address' proposals for improving budgetary controls over the government's direct loans and loan guarantees. IMPROVED CONTROL OVER DIRECT LOANS AND LOAN GUARANTEES The history of budgeting for federal credit programs over the past decade has been one of gradually changming the timing and nature of congressional action on credit programs. The trend has been for the Congress to get involved at ea,rlier stages in the life cycle of loans and guarantees, and to increasingly consider the cost implications of new loans and guarantees. The current proposals of GAO and others represent the latest stage in this evolution. Prior to this decade, most direct loans were made through revolving funds not requiring annual appropriations, while loan guarantees required appropriations only when defaults occurred. u As a result, credit programs typically came under budget scrutiny 10 only when appropriations were needed to recapitalize loan accounts or to cover defaults experienced on previously extended guarantees. Congressional budgetary actions were after-the-fact. This began to change when OMB included a “credit b&et” in the President’s budget submission for fiscal year 1981. The credit budget was a compilation of the various loan and loan guarantee accounts in the budget, accompanied by recommend,ed appropriation act limits on the new loans and guarantees of those accounts. In subsequent years, increasing numbers of such accounts have been placed unde.r such appropriation act limitations, with the result that by fiscal year 1989, about 38 percent of new direct loan obligations and 59 percent of new guarantee commitments were made under such statutory limitations. In 1985, the Congress added budget resolution controls to these growing appropriations controls. The idea was to’ provide in the annual resolutions overall targets for credit programsl and allocations to the various functions of the budget (agriculture, housing, etc.). Prior to this, the levels were left to the actions of several appropriations bills passed at differ.ent times. Yet even with these changes, the credit decisions of.the Congress today are still not based on estimates of final costs. Appripriations controls about to be enacted for fiscal year 1991 11 will indirectly govern future costs by limiting new levels of activity. These limits will not, however, directly control credit costs because the amounts in the appropriations bills authorize gross levels of credit activity, not the ultimate costs to the government on those programs. One result is that the current cash-based budget continues to misrepresent the costs of credit activities. ‘The costs of ‘a new direct loan program are overstated in the initial years of the cash disbursements and understated in the later years of cash repayments. Similarly, loan guarantees appear to be cost- free and are excluded from the budget’s cash-flow totals until default payments are made. Under such a system, it is practically impossible to make valid comparisons of grant, loan, and guarantee alternatives for extending aid. 12 I , ” This weakness in current controls has s timulated several proposals for budgetary reform from leaders of the Budget Committees ,5 OMB,6 CB0,7 and GAO.8 Each of the proposals endorses the notion of estimating the total credit subsidy costs for proposed direc t loans and loan, guarantees, and appropriating funds for the subsidy costs before the loans and guarantees are made. Credit subsidy costs would be recorded and controlled at the time of the decis ion to extend credit assis tance, thus' providing the information necessary to permit comparisons with other programs during budget review. 5See Chairman Panetta's H.R. 3929 introduced in the 2nd session of the 10ls t Congress; and H.J. Res. 324, a joint resolution passed by the Senate on July 31, ,1987. 6See Budget of the United States G overnment, .Fiscal Year 1989 (Part 6b) an,d Budget of the United States G overnment, Fiscal Year 1990 (Part 6). 7See CBO's Credit Reform: Comparable Budget Costs for Cash and Credit (December 1989). 8Budget Issues: Budgetary Treatment of Pederal'.Credit Programs (GAO/AFMD-89-42, April 10, 1989). 13 l ’ . In m ost of the proposals, credit subsidy accounts would be established in agencies having credit programs, to which projected credit subsidy costs would be appropriated. In addition, m ost would establish credit financing accounts in Treasury or in the originating agencies to provide the nonsubsidy portion of the funding for direct loans and to m ake ,default paym ents for loan guarantees. When a loan is m ade, the originating agency would p$y the appropriated subsidy amount to a financing account where it would be com bined with the balance of the loan amount and disbursed to the recipient. We would add a word of caution about the placem ent of the financing accounts-- in Treasury or the agencies? we would be concerned about any centralized approach if it is accom panied by debt m anagem e'nt procedures that lessen agencies accountability and responsibility for their programs. However, as I said at the outset, the areas of agreem ent outweigh any differences among these proposals and provide a sound basis for m oving ahead withm needed reform . Never,theless, there are som e differences worth m entioning among the proposals. One such difference relates to the m ethod'used to calculate credit subsidy costs. Our proposal, and that of the Budget Com m ittees, could be term ed a "cost-to-the-governm ent" approach. For direct loans, it would m easure the net present value of the difference between the costs to the Treasury of m aking the loans, 14 . and the expected receipts flowing back to the Treasury from loan repayments. Treasury costs would essentially be Treasury's interest costs for borrowing funds to finance the loans. The calculation of expected receipts from loan repayments would mainly take into ,consideration interest earnings and default rates. OMB, on the other hand, prefers a different method that could be termed a "market-valuation" approach. In the past, Cl30 has also seen some merit inthis approach. This method would calculate the economic benefit borrowers receive as a result of obtaining federal loans at more favorable terms than would be available to them from the private sector. The subsidy costs would be the present value of the additional payments that a federal borrower would be required to pay for a similar loan from the private sector. Although a market valuation approach may be useful for some purposes, we prefer'the cost-to-the-government method for governmental budgeting purposes. It would be more consistent with current budgeting practices, and with;the cost valuation practices followed for most other federal programs. When the Congress funds the School Lunch program, for example, it provides amounts to cover the ,government's costs, not amounts equal to the benefits to recipients, which may be quite u different. "15 Also, we think that it would be ve'ry difficult to calculate with reasonable confidence a market valuation for loans given to marginally creditworthy recipients, many of whom will default on their loan repayments. For many such persons, there is no real credit market OK set of applicable interest rates.' Such persons come to the government for loans because they cannot get commercial financing. All of the proposals would affect agencies' credit-related budget authority and outlays by requiring appropriations for credit subsidy outlays. However, they wou,ld not all affect the unified budget deficit in the same way. The GAO, OMB, and Senate Budget Committee proposals would not alter the government's reported, bottom-line deficit. They would be deficit-neutral in their effects. Consider, for example, a $50,000 loan. Under current practices, a $50,000 loan disbursement would count, as a $50,000 budget outlay in the year of disbursement, and add to the deficit that amount (assuming no offsetting repayments that year). Under the GAO, OMB, and Senate Budget 'Committee approaches, there would still be a $50.,000 outlay and impact on' the deficit. The subsidy portion, say $5,000, would be highlighted in budget reporting and funded in appropriations. On the other hand, the CBO proposal and the bill introduced by Chairman Panetta of the House Budget Committee (H.R. 3929), ly would ultimately move the nonsubsidy portion of the direct loans 16 "below the line," or off-budget, on the grounds that such outlays do not change the government's financial condition. The argument is that this portion of the loans represents money that the government will recover over time through loan repayments. CBO ,and H.R. 3929 would redefine these expenditures as a "medns of financing the deficit." In our $50,000 loan example, CBO,and H.R. 3929 would count as a budget outlay only the $5,000 subsidy amount: GAO would distinguish between the subsidy and nonsubsidy amounts in a different way. I am referring to our proposal to restructure the unified budget into general, trust,, and enterprise funds, with an operating and capital component for each. In that proposhl, we would classify the credit subsidy in the operating part of th'e budget and the nonsubsidy amount in the capital part. 'Both, however, would remain within the unified budget. We believe,it is important to maintain a link between the government's overall budget deficit and its borrowing needs. We urge that Congress, OMB, and CBO look carefully at our proposed budget restructuring, which we believe would permit a better understanding of the government's financial performance and financing requirements. In the meantime, however, it is clear to us that credit reform would be'an important improvement that can be accomplished within the present budget structure. 17 I want to emphasi'ze that the differences among the proposals should not be seen as major impediments to reform. I am confident that the common features outweigh the differences and that ways can be found to reach agreement on an approach that would be a significant improvement over current credit budgeting practices. I would like, now, to turn to some related matters. RELATED ISSUES H.R. 3929 would treat the government's deposit insurance.as a type of loan guaranty and require appropriations for the estimated costs not covered by insurance fees and premiums. We think that such a proposal deserves careful consideration. Indeed, we would probably extend the concept to include all forms of insurance programs where it has been determined that there are costs that cannot be covered by existing or future fee'and premium receipts. This has proved to be the case with FSLIC, for example. The basic issue is similar to that of loan guarantees. There needs to be a timely recognit,ion of, and appropriations for, the expected costs that cannot be covered by fees and premiums. On insurance programs, this would essentially be the u costs that are projected for payment from Treasury. 18 . Of course, the government's current rescue of the thrift deposit insurance system comes to mind. This is a clear case where a combination of deregulation, industry mistakes, and economic downturns created an insurance need not covered by the premiums and fees. The government will have to pay hundreds of billions of dollars to clear up the problem. How much of this Treasury price tag, reflecting costs and liabilities already incurred, has been included in budgetary totals? Only about $19 billion has been reflected in the budget authority totals. This is an amazing budgetary shortfall. It is past time for the government to get its books right on this savings and loan problem. I would urge all parties concerned to do the following as we approach anothesr Gramm- Rudman-Hollings (G-R-H) deficit snapshot date later this year: make full and accurate estimates of the costs in a timely fashion and get Treasury's share figured into the budget's totals. Also, on GSEs, we are currently studying certa,in budget treatment issues as well as the' safety and soundness of GSEs. Some are'inappropriately being used to move federal activities off-budget, and we will report to the Congress soon on our recommendations to minimize this kind of problem in the future. 19 In conclusion, I would urge the Congress to move forward at this time on cr,edit budgeting reform. I would also suggest an examination of the concepts for possible application to the insurance areas of the budget. This completes my statement, Mr. Chairman. I would be glad to respond to the questions you or Members of the Task Force may have. 20
Need for Improved Budgetary Controls Over Federal Credit Programs
Published by the Government Accountability Office on 1990-04-11.
Below is a raw (and likely hideous) rendition of the original report. (PDF)