An Unfair Value for Taxpayers
'Fair-Value' Budgeting Is a Dangerous Game to Play with Federal Loans and Guarantees
SOURCE: AP/Jacquelyn Martin
The House of Representatives this week passed a bill that would radically undo the budgeting process for government loans and guarantees. The proposed changes would add at least $55 billion per year in imaginary federal deficits while stifling critical government programs that create jobs, promote economic security for middle-class families, and pave the way for a more competitive future for our nation.
The Budget and Accounting Transparency Act of 2012 mandates the use of so-called fair-value budget reporting for all federal credit programs—budget parlance for an accounting trick that uses the private sector’s cost of funds instead of the government’s to make credit programs appear more expensive than they truly are. The new rules would add a premium to each program’s cost estimate based on the rates private lenders would charge to issue the same loan or guarantee.
At the heart of this bill is a debate over what types of risk the government should price and score in the budget. There’s no disagreement that the budget should accurately reflect “credit risk,” the likelihood a loan issued or guaranteed by the government will not be paid back. Indeed, current federal government budget rules already take that estimate into account. The question before Congress this week is whether the budget should add an additional cost to account for the rate a risk-averse private investor would charge for the perceived variability in those estimates, what is sometimes called “market risk.”
This brief lays out the context of this ongoing debate, summarizes and critiques the argument for fair-value reporting, and discusses the bill’s real-world impact on essential government programs, such as student aid and support to the housing market.
“Fair-value” actually means added costs
Despite its strategic misnomer, fair-value reporting is anything but “fair.” So for purposes of this issue brief, we’ll refer to it by a more appropriate name: “added-cost” reporting. Added-cost reporting is a bad idea for the following reasons:
- Instead of improving the accuracy of cost estimates for credit programs, it actually makes them less accurate by biasing apparent costs upward.
- It accounts for “phantom” costs that never actually materialize. This distorts the government’s true fiscal position, which is precisely what the budget is supposed to reflect.
- It causes serious harm to critical credit programs and adds tens of billions of dollars to the federal reported deficit while doing nothing to actually reduce the debt, minimize wasteful spending, or reduce taxpayer exposure to loss.
- It attempts to solve a problem that doesn’t really exist. The current budget rules have been effective, and the cost estimates reasonably accurate, over the past two decades.
- It gives opponents of particular credit programs a back-door way to scale back the government’s footprint in certain industries, under the guise of “responsible” budgeting.
With this understanding of the objectives of added-cost reporting in hand, let’s look at how the federal budget currently accounts for loans and loan guarantees.
A primer on federal credit budgeting
Prior to the early 1990s, costs for federal loans and loan guarantees were accounted for on a “cash basis,” tracking the amount of cash flowing into or out of the Treasury over the course of a year. This failed to reflect the long-term cost of credit activities, creating an inappropriate and misleading bias for loan guarantees that didn’t require up-front outlays. As a result, Congress and other federal policymakers lacked the information necessary to make informed budgeting decisions.
The Federal Credit Reform Act established a standardized system to capture the net value of a loan’s cash flows over the life of the loan. Since 1992, the government has estimated the lifetime cost for each new book of loans for each credit program. That estimate, also known as the “credit subsidy cost,” is then recorded in the federal budget and updated on an annual basis.
The first step in estimating the credit subsidy cost is to project the government’s expected cash inflows and outflows from the transaction. Projected cash flows include the disbursement of principal (for direct loans) or obligations (for loan guarantees), expected repayments, and any fees the government collects in the process. The government’s expected cash receipts depend on the likelihood of default, expected recoveries on defaulted loans, the borrower’s planned repayment schedule, expected prepayments, and the fee schedule.
Through that process, the Federal Credit Reform Act rules account for estimated “credit risk,” the chance a borrower will not be able to pay a loan back in full with interest. Of course, there is always a chance a loan will perform better or worse than expected; current budget rules reflect the most likely, or “base-case,” scenario. The budget baseline is updated based on the actual performance of the loans over time.
Under the law, projected future cash flows are discounted to reflect the so-called net present value of the direct loan or guarantee, which compares the value of a dollar today to the value of that same dollar in the future. Costs are discounted based on the interest rate on a Treasury bond with a comparable maturity as the loan or guarantee. In simple terms, this adjusts for the price the government has to pay to borrow the money it is lending out or using to back the guarantee.
In recent years, some have pushed for a new method for discounting subsidy costs to present value, claiming that current rules for discounting cash flows undervalue the “uncertainty” of certain credit programs. They argue that while future costs are relatively easy to estimate for some programs (say, short-term utility loans to thousands of rural house- holds), those costs are harder to estimate for other programs (say, working capital loans to a small number of startup companies). The Federal Credit Reform Act model accounts for the different probabilities of default in the budget, but does not treat the difference in uncertainty around these estimates as an additional cost—under the assumption that the federal government is in a unique position to absorb both levels of uncertainty.
Critics of the law argue that programs with high variability in cost estimates pose a higher market risk to the taxpayers: The less certain you are about the outcome, the more potential for losses above those estimated. (Set aside for the moment that there is equal potential for un-estimated gains.) They say that any “risk-averse” private investor would insist on being paid a premium whenever they invest in a financial instrument whose result is uncertain—a premium that is above and beyond the present value of expected defaults—so the federal government should do the same.
To solve this perceived problem, critics propose inflating the cost of credit programs to account for the price private firms would charge for the same loan or guarantee. They call this fair-value reporting but it is (as we demonstrated above) actually added-cost reporting.
The Federal Credit Reform Act reporting standards have proved effective and reasonably accurate over the past two decades, and almost all credit programs continue to use it today. Moreover, inaccuracies in existing Federal Credit Reform Act estimates come from imperfectly estimating the demand for loans and the actual default rates, not from failing to add a private-market premium.
But that’s only the beginning of the philosophical problems with added-cost reporting. Let’s examine each in turn.
Added-cost budgeting doesn’t make logical sense
The basic argument for added-cost reporting is that private firms are “risk-averse,” so the federal government should be too. But this ignores the simple fact that the federal government is not a private firm, nor is it simply an amalgamation of several million risk-averse taxpayers.
This argument taps into the core reason for federal credit programs. There are certain risks that private financial institutions are unwilling or unable to take, despite significant benefits to the public. In some cases, the government is in a unique position to assume those risks and spread them across a wide credit portfolio, all in an effort to achieve certain public goals. Since the government is not a profit-seeking firm, it should not be risk-averse; it should be “risk-neutral.”
It’s important to clarify what exactly we’re talking about here. The question is not whether government is particularly good or bad at estimating the likelihood of default, or even whether policymakers should better account for uncertainty when making policy decisions. It’s a question of whether that uncertainty ought to be explicitly scored in the federal budget.
The simple answer is that it should not. Scoring credit programs based on a discount rate with embedded market risk, which we’ll define as the level of variability in cost estimates, would add billions in phantom costs to the federal books, according to the bipartisan Center on Budget and Policy Priorities. Added-cost reporting requires that the budget “reflect amounts that the Treasury would never actually pay anyone,” totaling amounts that “would not be dollars that the government spends,” according to CBPP’s budget experts.
In other words, instead of helping the budget more accurately reflect the government’s fiscal position, added-cost reporting makes the budget less accurate.
Also, we shouldn’t kid ourselves that federal credit programs are the only government activities that involve financial risks. Several spending and revenue estimates are based on uncertain projections of economic activity, such as capital gains and other tax revenues, Social Security, unemployment insurance, and food stamp benefits. If anything, adding a risk premium would make credit programs appear more costly to government compared to equivalent grants or tax expenditures. The CBPP report puts it best:
If a risk-aversion adjustment were added to credit programs, it should be added to all such other costs as well. Not doing so would disadvantage credit programs relative to other forms of government assistance and distort the budget as a tool for allocating public resources.
But the fundamental point is deeper: It makes no sense to add a risk-aversion adjustment to the budget accounting of any federal government program.
To be sure, uncertainty is often an important consideration for sound policymaking. Lawmakers should consider the level of confidence in cost estimates when deciding authorization and funding levels for government programs. But the federal budget is not meant to assess the likelihood of positive and negative outcomes of a program. The budget is supposed to reflect the government’s fiscal position based on the best possible estimate of financial inflows and outflows, and nothing more.
There are separate questions of how the government can improve the accuracy of these cost estimates, or whether policymakers should be given more information on the variability of individual estimates when making policy decisions. But added-cost reporting accomplishes neither of these goals.
Finally, a sudden shift to added-cost reporting can cause serious harm to certain credit programs while doing nothing to reduce taxpayer exposure to loss. As the CBPP report points out, many of these phantom costs will require some sort of offset. In some cases, this means other programs will have to be cut to limit the net impact on the federal deficit. In others, specifically when programs are required by law to operate at no cost to government, credit programs will have to be scaled back significantly to account for these imaginary new costs, leaving behind otherwise creditworthy borrowers.
In either case, the American people suffer, both as taxpayers and recipients of these programs. And that’s no coincidence. There’s reason to believe this is the true motivation behind the conservative push for added-cost reporting—it is a back-door way to reduce the government’s footprint in certain industries.
Regardless of how their cost estimates are calculated and scored, federal loans and guarantees will still be grounded on the same basic assumptions and market forecasts. Biasing the estimates upward will not change the economic reality in which the government operates these programs. It will, however, overstate the costs government is likely to incur, which in turn will encourage misguided opposition and drive legislation to constrain their growth. Here’s how that might work.
Added-cost budgeting would cripple critical government programs
While there are currently hundreds of federal direct-loan and loan-guarantee programs, most government-assisted credit is provided through a small number of programs. Today the two biggest permanent federal credit programs budgeted under Federal Credit Reform Act reporting standards are the Federal Housing Administration’s single- family mortgage insurance program and a wide portfolio of student loan programs. Together these programs account for about 60 percent of all outstanding credit backed by the federal government.
An across-the-board transition to added-cost reporting, as proposed in the Budget and Accounting Transparency Act, would have severe financial implications on both programs, leading to a significant contraction in government support to both markets. Down the line, this would arbitrarily and unnecessarily raise the costs to borrowers served by these programs.
Let’s start with the Federal Housing Administration. Under the law, the nonpartisan Congressional Budget Office, or CBO, estimates that FHA’s single-family mortgage insurance program would produce budgetary savings of $4.4 billion in fiscal year 2012, mostly from fees collected from mortgage lenders. When calculated on an added-cost basis, the program would have a cost of $3.5 billion in 2012.
In other words, adding a market-rate premium would transform FHA’s flagship insurance program from a money-maker to a drain on the federal deficit, without altering the economic reality in which the program operates. (see Figure 1)
This raises a much bigger problem than pumping up the federal deficit. By law, FHA insurance programs must operate at no cost to government, so the agency would have to cover these new “losses” by increasing fees or tightening underwriting standards.
And that’s not easy. FHA’s insurance premiums are already the highest they have ever been in the agency’s 77-year history. FHA’s most recent fee increase in April 2011 increased the “economic value” of the agency’s 2011 book of business by less than $1.4 billion, according to agency estimates. So it would take significantly larger fee increases (or severe tightening of underwriting standards) to balance the agency’s books under added-cost reporting. Either action would severely scale back the government’s critical support to the struggling housing market, effectively kicking the legs out from under our economic recovery.
Added-cost reporting has a similar effect on the federal student loan portfolio. CBO in 2009 estimated that the federal student loan portfolio would save taxpayers nearly $46 billion between 2010 and 2020, based on Federal Credit Reform Act reporting standards. These savings are mostly from interest payments or fees charged to students for government guarantees.
When calculated through added-cost standards, the portfolio was estimated to cost taxpayers $157 billion over the same period. That’s a budgetary difference of more than $200 billion. (see Figure 2)
Unlike FHA’s insurance program, the federal student loan portfolio can run a cost to government. So assuming no change to lending activity, a change to added-cost reporting
would add an estimated $200 billion to the federal deficit over the next decade, just attributable of the student loan portfolio.
To be sure, added-cost reporting would have similar effects on other government programs that are essential to our economic growth and competitiveness, including small-business loans, clean energy loan guarantees, infrastructure loans, and loans for international development projects. Indeed, if enacted the Budget and Accounting Transparency Act would add a total of $55 billion in phantom costs to the deficit in the first year alone, according to CBO’s estimates. (see Figure 3)
It is beyond the scope of this issue brief to examine all of these programs to demonstrate the phantom costs of added-cost reporting, but let’s take a look at one key policy arena where enactment of the Budget and Accounting Transparency Act would torpedo our housing markets.
The added-cost budgetary treatment of Fannie Mae and Freddie Mac
In the years leading up to the financial crisis, federal support to the government-sponsored mortgage giants Fannie Mae and Freddie Mac was kept off the government’s balance sheet. Since the guarantees to Fannie and Freddie were “implicit” (meaning the government had no legal obligation to guarantee the debt of either institution) Fannie and Freddie were not covered by the Federal Credit Reform Act. That was profoundly wrong.
Ever since the government placed the two mortgage giants in conservatorship in 2008, both the Office of Management and Budget and CBO have accounted for the cost of conservatorship in the federal budget, but in different ways. The Budget and Accounting Transparency Act would drastically change this budgetary treatment by scoring all future Fannie and Freddie guarantees as traditional loan guarantees using added-cost reporting. And that’s chilling news for the struggling U.S. housing market.
As of September 2010, CBO estimated that, using added-cost report- ing, new guarantees made by Fannie and Freddie would cost taxpayers about $53 billion between 2011 and 2020. As a point of comparison, CBO estimated that under Federal Credit Reform Act standards these guarantees would generate net savings for the federal government of $44 billion over the same period. (see Figure 4)
To cover these phantom costs, the Federal Housing Finance Agency, Fannie and Freddie’s regulator and government conservator, would likely have to direct them both to increase their fees. And any fee increase would likely be substantial.
As a reference point, last year’s payroll tax cut extension included a mandate of a 10-basis-point increase to the fee charged on Fannie- and Freddie-backed loans (meaning an increase of 10 cents for every $100 dollars guaranteed), to be calculated using added-cost reporting. CBO estimated that increase would generate about $36 billion in revenues between 2012 and 2021. So even after accounting for that new revenue, Fannie and Freddie would have to bump up fees by significantly more just to offset the phantom costs of using added-cost reporting. By comparison, the average guarantee fee charged by Fannie and Freddie was just 26 basis points in 2010, according to the Federal Housing Finance Agency.
It’s important to understand the big-picture implications here. Each time Congress mandates an increase in guarantee fees—either explicitly or implicitly through changing the budget rules—fewer American families can afford a Fannie- or Freddie-backed loan. Through this stealthy effort to scale back government support, Congress is essentially pull- ing the rug out from under our still-struggling housing market.
Over the past 75 years, a government guarantee on certain residential mortgages has helped promote long-term stability in the housing market. It was also critical for creating and popularizing the affordable 30-year fixed-rate mortgage, now a pillar of the industry. Prematurely transitioning to a purely private market—the effective outcome of requiring Fannie and Freddie to account for their activities according to added-cost report- ing—could price millions of creditworthy homebuyers out of the market and trigger frequent boom-bust cycles, with devastating effects on the broader economy.
A responsible path forward
The concepts laid out in the Federal Credit Reform Act have correctly reflected the government’s fiscal position for nearly two decades. It would be unwise for Congress to try fixing a model that isn’t broken. To the extent estimates of program costs have been inaccurate for individual programs, the federal government should devise new methods of estimating defaults, repayments, prepayments, and recoveries. But it is not appropriate to change the scorekeeping rules to add substantial premiums, even if a program’s current estimates are perfect.
That said, more can and should be done to improve the way policymakers weigh the costs and benefits of federal credit programs. Lawmakers should have access to all the information necessary to make informed policy decisions, which includes some estimate of uncertainty regarding cost estimates. One possible solution would be to report a metric for each book of business as part of the annual budget’s Federal Credit Supplement.
Regardless of how they’re reported, though, these confidence measurements should not be priced in the federal budget. Such adjustments add arbitrary costs for certain types of risk while ignoring others, biasing the budget against federal credit programs.
The House of Representatives is entering precarious territory with the Budget and Accounting Transparency Act. At a time of strict fiscal discipline, the last thing our government needs to do is start conjuring up imaginary costs to tack onto the federal deficit.
John Griffith is a Research Associate with the Economic Policy team at the Center for American Progress.
To speak with our experts on this topic, please contact:
Print: Liz Bartolomeo (poverty, health care)
202.481.8151 or firstname.lastname@example.org
Print: Tom Caiazza (foreign policy, energy and environment, LGBT issues, gun-violence prevention)
202.481.7141 or email@example.com
Print: Allison Preiss (economy, education)
202.478.6331 or firstname.lastname@example.org
Print: Tanya Arditi (immigration, Progress 2050, race issues, demographics, criminal justice)
202.741.6258 or email@example.com
Print: Chelsea Kiene (women's issues, Talk Poverty, faith)
202.478.5328 or firstname.lastname@example.org
Print: Elise Shulman (oceans)
202.796.9705 or email@example.com
Print: Katie Murphy (Legal Progress)
202.495.3682 or firstname.lastname@example.org
Spanish-language and ethnic media: Jennifer Molina
202.796.9706 or email@example.com
TV: Rachel Rosen
202.483.2675 or firstname.lastname@example.org
Radio: Chelsea Kiene
202.478.5328 or email@example.com