Click here for a printer-friendly version of this article
Fiscal conservatives, budget experts, and free market-supporters lost a debate 20 years ago over how the federal government should report the costs and risks of its loan programs. The outcome of that debate continues to wreak havoc on the federal budget today, as the rules that were adopted make it appear as if the government can earn a profit by lending at below-market rates to everyone from homeowners to big banks and college students. The new Congress set to arrive in Washington in January should fix this problem as one of its first acts of fiscal rectitude. They should update the budget rules so that loan program costs reflect “fair-market” values like those that would be used by the private sector.
Three previous articles appearing on these pages have explained that the 1990 Credit Reform Act – still in place today – improved the way Congress and federal agencies assess the costs of loan programs but that the law included an unfortunate provision that also dramatically understates the risks and costs of many of these programs. Specifically, this flaw in the Credit Reform Act makes government-backed loan programs, such as FHA mortgage insurance and government-backed student loans, appear profitable for the government even though they provide below-market interest rates and other subsides to borrowers that would be unprofitable for even the largest, most efficient private lenders to make.
A review of the early debates that led to the adoption of the budget rules for loan programs reveals that there were in fact many policymakers who favored using market prices to assess the costs of government-backed loans. These proponents were rebuffed when drafters of the 1990 Credit Reform Act chose to adopt rules that effectively exclude private market values for the loan programs.
As explained in our earlier commentary, the budget rules for loan programs make it appear as if the government’s costs are always lower than the private sector’s for making identical loans even when the government has no inherent efficiency advantage. This is because the rules operate on the faulty assumption that taxpayers do not demand the same compensation for bearing risk that private lenders do for making the same loans. This faulty assumption stems from a part of the Credit Reform Act that measures the costs of government-backed loans using the government’s borrowing rate (the discount rate applied to the cash flows from the loans). Treasury rates are “risk-free” interest rates since there’s little chance the government will default, and using them to price loans suggests that loans will perform exactly as the government expects. Of course loans rarely perform exactly as expected. So using the Treasury rate to value them actually excludes unexpected risks from any cost calculation.
This fundamental problem with using the government’s borrowing cost to measure loan program costs instead of market-based values was highlighted early on in the debates that preceded enactment of the Credit Reform Act. In fact, that history shows how the decision to use Treasury rates was not the obvious choice for policymakers when the Credit Reform Act was adopted, that the arguments in favor of Treasury rates weren’t compelling, and that there was a strong case made for using private market values instead.
Prior to the adoption of the Credit Reform Act in 1990, what Congress spent on government loan programs was almost impossible to discern and couldn’t be compared with what was spent on other programs. This was because loan programs were measured by volume and by yearly cash flows of disbursals and repayments, which aren’t a reflection of costs. After a debate that lasted for much of the 1980s, policymakers reached a consensus that the costs of government-backed loans should be measured by the “subsidy” that they provided to borrowers. There was also consensus that the subsidy should reflect the lifetime costs of the loans all at once, in the year that the loans were originated, so that the costs of the program could be assessed when the obligations became binding. This is essentially a form of accrual accounting. Despite that early consensus, a deep divide emerged among the reformers over exactly how to measure the subsidy.
The Congressional Budget Office and the Office of Management and Budget, the Reagan Administration, and the George H.W. Bush Administration all wanted government subsidies for loan programs valued on par with private market prices. They generally defined the subsidy as the difference between the interest rate a borrower paid on a fully private loan compared to what that borrower would have paid on a government-backed loan. To measure the subsidy or difference, the Reagan Administration proposed selling all direct loans in the private market or buying reinsurance for all loan guarantees and then recording any negative difference as a cost. For direct loans, the subsidy would be the difference between the government’s expected cash flow versus what the price private lenders paid to buy the loan. For guaranteed loans, the cost of the loan would have been whatever the cost was to the government to fully reinsure it in the private market.
Later iterations proposed by the George H.W. Bush Administration in 1989 and 1990 dropped the idea of major loan sales, proposing instead to have the Treasury Department estimate the private market value of the loans by discounting their expected cash flow with market-based rates. This change was likely in response to critics who argued that comprehensive loan sales would be too difficult to implement given that markets for the loans did not exist.
The GAO and lawmakers on the Senate Budget Committee opposed the market-based plans and argued that private market values were not appropriate for valuing government programs. Their proposals would still have measured loans in terms of subsidies and accrual accounting practices, but they favored using the U.S. Treasury borrowing rate as a discount rate.
In 1987 congressional testimony and a subsequent 1989 report, the GAO argued that market-based discount rates shouldn’t be adopted because they measure the “benefits to the borrower” that a government loan provides. The agency believed that the appropriate measure was instead a loan’s “cost to the government” which, according to the agency, is captured by using Treasury interest rates as discount rates in the budget rules. The GAO never offered a convincing explanation as to how these two measurements – benefits to the borrower and costs to the government -- could be different, nor did the agency refute that the difference might be an accounting illusion. Instead the agency argued the following:
For determining the present value of federal loans… we use the interest rate paid by Treasury on its borrowings for securities with comparable maturities to discount the expected loan income stream. This reflects the idea that, because the government has debts, the present value to the government of an expected income stream relates to interest-cost avoidance. If the government sells the loan assets, it receives cash which enables it to pay off existing debt and avoid future interest payments. Alternatively, if the government makes additional loans, depending on whether it is in a deficit or surplus situation, it either incurs additional debt or forgoes revenues that could be used to avoid future interest costs.
While the GAO’s argument certainly sounded logical on its face, it does not address the practical implications of using the government’s cost of borrowing to value loan programs. That is, the agency does not answer why taxpayers would demand less compensation for making a risky loan than they would in a purely private market.
The Congressional Budget Office, on the other hand, understood the pitfalls of using risk-free Treasury rates to value loan programs. In 1987, CBO Acting Director Ed Gramlich pointed out to lawmakers that the government bears the same risks as any investor or lender, so it cannot magically earn higher returns than the private sector when making an identical loan. In fact, he was quite clear in congressional testimony that the government’s risk-free borrowing rate was irrelevant to valuing the costs and risks of government loan programs.
The higher cost of borrowed money for private firms [compared to the government] is also irrelevant to the pricing of financial assets… it is important to recognize that the use of capital has a cost not matter who is using it. Finally, government, like all other entities, must use a risk-adjusted interest rate to discount risky income streams. In sum, I am unaware of any evidence that market prices for federally-originated financial instruments would be consistently biased because private [borrowing] costs are higher than the government’s cost.
Later in a 1989 report, the CBO explained that using the government’s borrowing rates to measure the costs of loan programs would turn conventional finance on its head. The report clearly outlines how discount rates used by the private market distinguish between higher and lower risk loans while using a Treasury rate would treat them the same way.
Use of a Treasury borrowing rate to discount future cash flows on a credit contract treats those payments as though they were certain to be received. If the government were to use its own risk-free cost of borrowing to discount uncertain future cash flows, it would be the only financial institution to do so. An aversion to risk causes others, including federally insured banks and thrifts, to discount risky income at a higher than risk-free rate.
Others recommend that the government select a discount rate equal to the rate that the government would receive by investing in other equally risky assets. They maintain that the use of such a rate is necessary to capture the opportunity costs of credit or the value of alternatives forgone. The use of rates that account for risk is also necessary to distinguish the cost of assets with equal expected income but different degrees of risk.
Despite the fact that the more cogent and logical arguments were put forth by the proponents of market-based cost estimates, they were steamrolled (or dismissed) when the Credit Reform Act was slipped into a 1,000-page omnibus budget bill signed into law in 1990. The drafters of the law decided that only interest rates on U.S. Treasury debt could be used to value the subsidy on the government’s direct loans and guarantees.
Today, two decades of experience with the Credit Reform Act has revealed that using Treasury rates to value the government’s loans programs – with some $2.8 trillion in loans outstanding – is every bit as flawed as early critics believed. Using modern financial tools and data that reflect market prices (rather than selling government loans as the Reagan and Bush Administrations proposed), analyst have documented how these programs impose much greater costs on taxpayers and expose the government to far greater risks than what official figures state. In short, those who warned against using Treasury rates to value government loan programs have been proven right over the past two decades.
Unless the Credit Reform Act is amended, the artificial cost advantage that the federal government has in its loan programs remains an open invitation for Congress and the President to expand the scope of the government without acknowledging the risks and costs that taxpayers will ultimately have to bear. That should give the new Congress cause for concern. Updating or changing the Credit Reform Act to reflect fair-value accounting should be a key part of their agenda when they arrive in Washington in January, and it would right a wrong that has mislead policymakers for two decades.
Jason Delisle is the Director of the Federal Education Budget Project at the New America Foundation.