Testimony on Estimates of the Cost of the Credit Programs of the Export-Import Bank

Report
June 25, 2014

Testimony by Douglas W. Elmendorf, CBO Director, before the Committee on Financial Services, U.S. House of Representatives

This testimony describes CBO’s recent estimates of the budgetary costs of the Export-Import Bank’s (Ex-Im Bank’s) credit programs, which were provided in Fair-Value Estimates of the Cost of Selected Federal Credit Programs for 2015 to 2024 (May 2014), and discusses the two different approaches that CBO uses to estimate the costs of credit programs:

  • One approach reflects the procedures currently used in the federal budget as prescribed by the Federal Credit Reform Act of 1990 (FCRA), and
  • Another approach, known as fair value, shows estimated costs that reflect the market value of the federal government’s obligations.

For fiscal years 2015 to 2024, CBO found that Ex-Im Bank’s six largest credit programs would generate budgetary savings of about $14 billion under FCRA accounting but cost about $2 billion on a fair-value basis. Both the FCRA and fair-value estimates are based on Ex-Im Bank’s projections of cash flows for those credit programs as reported in the Federal Credit Supplement to the Administration’s 2015 budget. Thus, both estimates reflect the program terms and outcomes—including the amount of lending, fees, and borrowers’ rates of repayment and default—that are expected to prevail under the current structure of the programs and the President’s budget request for those programs in 2015.

The difference between the two estimates lies in the treatment of the cost of market risk, which is one component of financial risk. Much of the risk of financial investments can be avoided by diversifying a portfolio; market risk is the component of risk that remains even after a portfolio has been diversified as much as possible. It arises because most investments tend to perform relatively poorly when the economy is weak and relatively well when the economy is strong. People value income from investments more when the economy is weak and incomes are relatively low, and so assign a higher cost to losses that occur during economic downturns. The higher cost of losses in bad times (as well as lower cost in good times) is captured in the cost of market risk.

The government is exposed to market risk through its credit programs because, when the economy is weak, borrowers default on their debt obligations more frequently and recoveries from defaulting borrowers are smaller. That market risk is effectively passed along to taxpayers and beneficiaries of government programs because they bear the consequences of the government’s financial losses. Moreover, that risk is costly to those taxpayers and beneficiaries because they also tend to value resources more highly when the economy is weak.

Under the FCRA approach to accounting for federal credit programs, Treasury borrowing rates are used to discount expected future cash flows—that is, to translate future cash flows into current dollars. That approach essentially treats future cash flows subject to market risk as having the same value as Treasury securities that promise the same average payments with no risk. This means that the market risk of federal credit assistance is treated implicitly as having no cost to the government, which has important consequences:

  • The costs of federal credit programs reported in the budget are generally lower than the costs to private financial institutions for providing credit on the same terms—at least in part because private institutions require compensation for market risk;
  • The budgetary costs of federal credit programs are generally lower than those of grants for similar purposes that involve equivalent economic costs; and
  • Purchases of loans and loan guarantees at market prices appear to make money for the government and, conversely, sales of loans and loan guarantees at market prices appear to result in losses.

To incorporate the cost of market risk, the fair-value approach generally entails using the discount rates on expected future cash flows that private financial institutions would use. That approach effectively uses market prices to measure the cost to the public of the lower returns on federal loans and loan guarantees when the economy is weak and incomes are relatively low. In CBO’s view, therefore, fair-value estimates provide a more comprehensive measure of the costs of federal credit programs, and CBO has provided fair-value estimates for many programs to help lawmakers more fully understand the trade-offs between certain policies.

Some analysts have expressed concern, however, about potential drawbacks of using the fair-value approach in federal budgeting, including the following:

  • Fair-value estimates include costs that will not be paid directly by the federal government if actual cash flows turn out to match expected cash flows, and including those costs makes comparisons with estimated costs for some noncredit programs more difficult;
  • Fair-value estimates are somewhat more volatile than FCRA estimates over time because of changes in the cost of market risk;
  • Producing fair-value estimates is more complex than producing FCRA estimates; and
  • Communicating the basis for fair-value estimates to policymakers and the public is more difficult than communicating the basis for FCRA estimates.

Proponents of the fair-value approach respond to those concerns by arguing the following: Decisions about spending the public’s money should take into account how the public assesses financial risks as expressed through market prices; by taking those prices into account, fair-value estimates provide unbiased estimates of the expected cost of loans and loan guarantees when that credit assistance is offered; and, other concerns can be mitigated by using established accounting practices.

This testimony provides further background on the cost of market risk and on the FCRA and fair-value approaches, drawn primarily from CBO’s earlier report Fair-Value Accounting for Federal Credit Programs (March 2012).