June 17, 2015
FCRA vs. Fair Value Accounting: A Comparison and Recommendation
*The views expressed here are my own and not those of the American Action Forum. I thank Gordon Gray, Chad Miller, Meghan Milloy, and Andy Winkler for their assistance in preparing for this testimony.
Chairman Coats, Ranking Member Maloney, and Members of the Committee, thank you for the opportunity to speak with you today. In this testimony, I’d like to make three basic points that support incorporating Fair Value Accounting into the federal budget process:
1. The Federal Credit Reform Act (FCRA) was needed to place the cash flows for economically equivalent activities on a level budgetary playing field, but has proven to underestimate the ultimate taxpayer cost of credit activities.
2. Fair Value Accounting (FVA) provides a more comprehensive measure of the cost of federal credit programs by better accounting for risk.
3. FVA already has been used successfully as the standard for budgetary treatment for TARP and for federal assistance to Fannie Mae and Freddie Mac.
Let me address each in turn.
Creating a Level Budgetary Playing Field
The Federal Credit Reform Act (FCRA) was passed in 1990 to ensure that the federal budget more accurately reflected the taxpayer costs for both federal direct loans and federally guaranteed loans, and did so in a comparable fashion. These two activities have the same economic purpose: provide credit using the taxpayer as a guarantee of repayment. But in the absence of FCRA they appear quite differently on the federal budget.
On the one hand are direct loans: loans made directly from the federal government to the borrower in which the federal government makes one large payment for the total amount of the loan and then, over the course of the next several years, receives incremental payments from the borrower for the principal amount and the interest of that loan. If the borrower fails to repay, the taxpayer picks up the shortfall. Before FCRA, accounting for a direct loan showed the full amount of the loan as an outlay in the year that it was disbursed, recorded principal and interest as receipts in the year received, and had no explicit recognition of the probability of failure to repay in full.
On the other hand are federally guaranteed loans: loans made to the borrower by a private lender in which repayment is guaranteed by the full faith and credit of the federal government. Repayment of those loans is made directly to the lender, but the government receives various fees at the time of application and/or throughout the repayment period. However, if the borrower defaults on the loan, the federal government is liable for the amount of the percentage of the loan that it guaranteed. Unlike direct loans, before FCRA, federally-guaranteed loans showed a budgetary “profit” at the time of approval since the government received the guarantee fee yet owed no money unless the borrower ended up defaulting down the road.
After FCRA, however, direct and guaranteed loans were accounted for in the same way: a single lump sum of the estimate of the net present value of receipts and outlays (excluding administrative costs and any incidental effects on government receipts or outlays). In this way, any taxpayer subsidy is identified in the budget (the present value of receipts is smaller than the present value of outlays) and is directly comparable between loans and guarantees.
FCRA Underestimates Taxpayer Costs
FCRA mandated the computation of subsidy rates in credit programs. In a CBO letter to then Ranking Member Judd Gregg of the Senate Budget Committee the process is described: “FCRA facilitates the comparison of the budgetary effects of direct loans and loan guarantees by converting the net outlays for each program into a single lump-sum estimate of net costs (that is, the discounted present value of all cash flows). Those cash flows are discounted using the government’s costs of borrowing – that is, the interest rates it pays on Treasury securities of comparable maturities. The resulting subsidy estimate is recorded in the federal budget in the year of a loan’s disbursement. Subsidies computed under FCRA do not include the government’s costs for administering the loans; those administrative costs are recorded separately, on a cash basis.”
Unfortunately, the FCRA approach has systematically underestimated costs. As pointed out in a recent study conducted by the American Action Forum, the budget estimates required under FCRA are fundamentally uncertain due to market changes and economic fluctuations. In those circumstances, one would expect that some estimates would turn out to be too high, while others would be too low. Instead, a quick look at some existing federal direct or guaranteed loan programs shows both a wide variance in revised estimates of program costs (compared to their original estimates) and a bias toward revising the costs upward.
Table 1: Lifetime Re-estimates for Select Federal Credit Programs
(Direct loan programs are in standard font; guaranteed loan programs are in italics, savings relative to original estimates are noted with parentheses)
|Department and Program||Lifetime Re-estimates||Lifetime Disbursements|
|Department of Agriculture|
|Agricultural Credit Insurance Fund||($1 billion)||$25 billion|
|Rural Housing Service||$278,000||$37 billion|
|Rural Utilities Service||$264 million||$82 billion|
|Agricultural Credit Insurance Fund||($2.2 billion)||$100 billion|
|Rural Housing Service||$2,522,000||$129 billion|
|Rural Utilities Service||($2.4 million)||$388 million|
|Department of Commerce||($68 million)||$814 million|
|Department of Education|
|Federal Direct Student Loans||$10,963,000||$762 billion|
|Department of Housing and Urban Development|
|Mutual Mortgage Insurance Program||$77,717,000||$2.773 trillion|
Under FCRA estimates, the federal direct student loan program has cost the Treasury (and taxpayers) almost $11 billion since its inception. While not inconsequential, this is but a fraction of the nearly $78 billion re-estimate on federal housing guaranteed loan programs.
While FCRA has improved the federal government’s ability to account for credit programs compared to the cash flow accounting in use for the majority of federal programs, it is far from perfect. Examining only a handful of the largest federal credit programs shows that since implementing FCRA accounting methods, the federal government has had to come up with roughly $90 billion to offset underestimated costs of credit programs. Those are costs that are ultimately born by the taxpayer.
The bias stems from omitting a source of risk: market risk. An alternative would base the budget cost of federal credit programs on an estimate of the market value of the federal government’s obligations, an estimate that accounts for market risk and includes administrative costs associated with originating, servicing, and collecting the loan.
Fair Value Accounting is A More Comprehensive Measure of the Cost of Federal Credit Programs
This approach is commonly referred to as Fair Value Accounting. Per a 2014 CBO study on fair value estimates of the cost of federal credit programs, “[m]arket risk is the component of financial risk that remains even after investors have diversified their portfolios as much as possible; it arises from shifts in macroeconomic conditions, such as productivity and employment, and from changes in expectations about future macroeconomic conditions. The government is exposed to market risk when the economy is weak because borrowers default on their debt obligations more frequently and recoveries from borrowers are rare. When the government extends credit, the associated market risk of those obligations is effectively passed along to taxpayers, who, as investors, would view that risk as having a cost. Therefore, the fair-value approach offers a more comprehensive estimate of federal costs.”
Such a significant reform to budget procedures should not be undertaken lightly. However, my views are informed by the fact that, during my tenure as Director, the Congressional Budget Office undertook a number of studies of the implications of accounting fully for economic risks in the budgetary treatment of financial commitments such as the credit programs being discussed today. In example after example (pension guarantees; deposit insurance; flood insurance; student loans; and assistance for Chrysler and America West Airlines) it becomes clear that an incomplete assessment of risks leads to misleading budget presentations and may engender poor policy decisions. Fair value accounting would be a significant step toward improving this informational deficit.
My views are echoed by a wide array of budget experts. In addition to a March 2010 CBO report which discussed the use of FVA for federal student loan programs and stated that budget rules do “not include the costs to taxpayers that stem from certain risks involved in lending,” the Pew-Peterson Commission on Budget Reform proposed “fair-value accounting” for credit programs and the President’s National Commission on Fiscal Responsibility and Reform advocated for reform of budget concepts that would more accurately reflect costs. It is clear from many studies, reports, and varying points of view, that FVA is a better, more comprehensive, and more accurate measure of the cost of federal credit programs.
Fair Value Accounting Already has Been Used Successfully in TARP and Fannie/Freddie
In closing, I would like to draw your attention to two recent instances in which FVA has been widely used and accepted. First, in a 2009 study, commissioned by chair of the independent Congressional Oversight Panel Elizabeth Warren, FVA was successfully used to show that the total market value of TARP assets was $176 billion – $78 billion lower than what the government had paid. In dismissing the methods used by former Treasury Secretary Henry M. Paulson and advocating for the FVA study on the matter, Warren said, “Despite the assurances of then-Secretary Paulson, who said that the transactions were at par – that is for every $100 injected into the banks the taxpayer received stocks and warrants from the banks worth about $100 – the valuation study concludes that Treasury paid substantially more for the assets it purchased under the TARP than their then-current market value.” In doing so, Warren further dismissed the administration’s official FCRA estimate which stated that the assets were, at the time of purchase, valued at $12.6 billion more than taxpayers paid – an estimate that failed to include the cost of market risk.
Finally, in a 2011 report, CBO stated that the real cost of the federal government guaranteeing Fannie Mae and Freddie Mac is $317 billion, more than double the $130 billion claimed by the Obama Administration using cash payment accounting methods. Specifically, the report explained that CBO considered the mortgages guaranteed by Fannie and Freddie as new guarantee obligations of the federal government, and “[f]or those guarantees, CBO’s projections of budget outlays equal the estimated federal subsidies inherent in the commitments at the time they are made.” This was a sharp contrast to OMB’s treatment of Fannie and Freddie as nongovernmental entities and its limited reporting of only the net cash payments provided by Treasury to Fannie and Freddie. Since then, the CBO estimates have been used as the standard, and OMB’s FCRA estimates have been widely disregarded.
Thank you. I look forward to answering your questions.
 Electronic versions of The 2015 Federal Credit Supplement in both PDF and Excel formats are available at http://www.whitehouse.gov/omb/budget/Supplemental. Tables 7 and 8 provide information on the lifetime reestimates of direct loan and loan guarantee cohorts 1992 through 2013 by agency, bureau, program, and risk category. The tables show the following for each cohort: the original subsidy rate at the point of obligation or commitment; the current reestimated subsidy rate; the change in subsidy rates due to interest and technical updates; dollar reestimates for the current year; net lifetime reestimate data; and lifetime disbursements through September 30, 2013. Some cohorts displayed in the 2015 Supplement have closed—there are no outstanding direct or guaranteed loans in the cohort and no further activity is expected. These cohorts are identified with a footnote and will not be displayed in future Federal Credit Supplement volumes.