FROM ISSUE NUMBER 21 ~ FALL 2014 GO TO TABLE OF CONTENTS

TEXT SIZE A A A

The Case for Fair-Value Accounting

JASON DELISLE and JASON RICHWINE

In each of the last two years, Congress has wrestled with the politics and policy of student loans. On the surface, the terms of the debate seem straightforward: Lowering the interest rate on federal loans would ease the burden on students, but it would also cost taxpayers money. It's a standard, unremarkable political trade-off.

Close followers of the student-loan debate, however, may have noticed something peculiar. The Congressional Budget Office simultaneously provided lawmakers with two different estimates of what the existing loan program would cost over the next decade. One stated that the program would cost $95 billion, while the other claimed a profit for the government of $184 billion. Even more puzzling, the estimate showing a profit was the official figure, though the CBO argued that the $95 billion loss was a more comprehensive estimate of the costs that taxpayers would bear.

How could this be? The short answer is that the government's official method for estimating cost is incomplete. It fails to incorporate the cost of the market risk associated with expecting future loan repayments. So-called "fair-value accounting," an accounting method favored by the vast majority of finance economists as well as the CBO itself, factors in the cost of market risk. The difference transforms the official student-loan "profit" into a loss, for a budgetary swing of $279 billion over ten years. That figure demonstrates why the stakes are so high in the debate about fair-value accounting.

This seemingly dry accounting issue illustrates an important way in which the costs of government programs are systematically understated. In nearly every government credit or insurance program — from student loans to public pensions to green-energy subsidies — the government makes risky investments without recording the full price of the risk. The government, at both the federal and state levels, exploits this accounting flaw to generate dishonest financial schemes — and some would even be illegal if they occurred in the private sector. Political activists and politicians only add to the confusion by taking inconsistent stances, supporting fair-value accounting when politically convenient but opposing it otherwise. Accounting should be a matter of objective accuracy about the nation's fiscal situation, but it has become a political tool exploited for partisan ends.

ACCRUAL ACCOUNTING

Accounting, at its simplest, is a matter of keeping track of how much money comes in and how much goes out. Until 1990, the federal government handled all its finances exactly that way — on what is called a "cash" basis. Cash accounting is easy to understand, and there can be little argument over how to do it. But cash accounting is inadequate for credit programs, which by definition involve future revenue streams. A commitment to spend money later or the promise of future revenue changes the government's financial position today, but "money in, money out" accounting would not immediately reflect that change.

Cash accounting was distorting how Congress made spending decisions, enticing politicians to favor credit programs with back-loaded costs. For example, "money in, money out" shows that in the year a direct loan is made, the loan costs the government the full amount disbursed, regardless of how of much of it is likely to be repaid in the future. A loan guarantee, by contrast, appears to generate a profit when the government makes the commitment, because the recipient of the guarantee pays an immediate fee and no cash leaves the treasury unless and until there is a default at some later date. Over a long enough period of time, the cost of the direct loan would be the same as the loan guarantee, but because of the short-term effect on the budget in a cash-accounting system, politicians prefer loan guarantees.

At the urging of economists and budget experts, Congress enacted the Federal Credit Reform Act (FCRA) in 1990, which changed the budgetary treatment of credit programs by assessing their costs on an "accrual" basis. Under accrual accounting, all projected losses and gains from a new investment are bundled together into a single cost applied to the current year.

A simplified example illustrates how accrual accounting works. Imagine that the federal government disburses a $300 student loan this year, and the receiving student pays back the loan in $100 installments over three years. Assume for a moment that there is no default risk, interest charge, or time-loss discounting — we'll build up to those issues as we go along. Under accrual accounting, this simple loan would be entered on the current year's fiscal ledgers as -$300 + $100 + $100 + $100 = $0. Lending the student $300 has no effect on the budget.

Not all students will pay back their loans in full, however, so the government projects how much it will actually receive in loan repayments. If the government expects only $0.95 of every loaned dollar to be repaid, then the new calculation would be -$300 + $95 + $95 + $95 = -$15. With the possibility of a default accounted for, lending $300 to the student will be entered as a $15 cost on the government's accounting sheet.

Taking the example a step further, the government charges the student interest at, say, 6% annually. Including interest payments along with the expected defaults, the government's projected revenue stream becomes -$300 + $106.62 + $106.62 + $106.62 = $19.86. Now the loan has a positive value to the government of $19.86.

Up to this point, there is no difference between the official government accounting method and fair-value accounting. Both systems are accrual methods that use the same estimates of how much money the government should expect to be repaid, taking into account default risk and interest rates. There may be better ways to estimate those expected repayments, but the fair-value critique of the status quo accepts the estimates as given.

ACCOUNTING FOR MARKET RISK

The sole difference between the official and fair-value methods is the rate at which the identical projected cash flows are converted (or "discounted") to a present value. In other words, the discount rates are different. (Technically, fair-value accounting should also include administrative costs, which the government's current method under FCRA does not. However, the CBO excludes administrative costs from both its official and fair-value calculations, so we won't consider the issue here.)

A discount rate reduces the value of future revenue based on the time-value of money. It's always necessary to discount future income streams because a dollar received later is worth less than a dollar received now. To decide how much less for the official budget estimate, the government references the yield on Treasury bonds. Assume, for example, that the yield on a one-year bond is 3%. Since the government could invest a dollar today in U.S. Treasuries and have $1.03 next year, the government reasons that $1.03 next year is really like having $1 today — and thus that receiving $1 in a year is like having only about $0.97 today.

Because money now is worth more than money later, the government applies that 3% annual discount rate to the expected payments. (In reality, the discount rate used is whatever the current Treasury yield is — it fluctuates and is right now much lower than 3% for a one-year bond. For simplicity, we will assume a 3% yield.) Taking into account the time-value of money using a 3% rate, the income stream in our simplified student-loan example is now -$300 + $103.52 + $100.50 + $97.57 = $1.59.

By using the Treasury yield as the discount rate, the official government accounting under FCRA effectively requires that cost estimates treat the projected cash flow from a student loan as if it were just as safe and reliable as the payout on a U.S. Treasury bond. The U.S. Treasury yield is the interest rate that market participants demand as compensation for bearing no market risk; that is, irrespective of the state of the economy, the security will pay off exactly as expected. Yet the projected payments from a student loan — or almost any loan for that matter — are quite different from payments flowing from a U.S. Treasury bond. They are subject to risk.

In finance, uncertain amounts are worth less than certain amounts, owing to the risk aversion of most individuals. Imagine a game in which a coin is flipped one time to determine a player's winnings. "Heads" results in a $100 payoff, whereas "tails" gives the player nothing. Although the potential payoff for this game is $100 and the average payoff is $50, few people would elect to play the game if they were offered the option of just taking a certain $50, or even $45. People tend to prefer certainty even when the certain option pays less than the average value of a risky alternative. The same principle explains why people invest in bonds rather than stocks much of the time — bonds are safer, even though they have a lower rate of return.

Some risks can be mitigated by diversification — spreading the risk across multiple "bets" so as to reduce the risk to the whole without affecting the expected value. The standard financial advice to build a "diverse investment portfolio" rests on this reasoning. But unlike coin flips, student-loan payments (and payments for most types of loans) are subject to a special kind of uncertainty known as "market risk": the risk of a general downturn in the economy. Because a downturn would affect the financial markets as a whole, market risk cannot be diversified away by holding other assets.

Fair-value accounting uses a higher discount rate than the official method. It incorporates not just the time-value of money (as the current system does) but also the cost of the uncertainty surrounding future loan payments. Students might pay back the expected principal and interest, but taxpayers bear the risk that they will not. If the United States enters another recession, students will not make their loan payments as expected, and the government will have less money just when it can ill afford to be short of funds. By using a higher discount rate than the risk-free one, fair-value accounting captures the cost of that risk; FCRA accounting, by using the risk-free Treasury rate, does not.

In our example above, the government's 3% discount rate (accounting for only the time-value of money) produced a $1.59 profit for the government. But let's say the cost of market risk adds two more percentage points to the discount rate. The income stream with a 5% discount rate would be -$300 + $101.54 + $96.71 + $92.10 = -$9.64. Just like that, the student-loan "profit" becomes a loss.

It is important to note that the government, under its current accounting method, is not expecting a full repayment on its credit programs — the expectation in the example above was a 95% repayment rate. The government's error is to treat the expected repayment rate as risk-free when in fact there is considerable uncertainty surrounding it. The repayments may be lower than expected, or they may end up higher, but the uncertainty adds a cost. Put another way, a private investor would never purchase the right to collect student-loan payments for a price that was the same as the expected payments. The investor would demand a lower price to take on such a risky asset.

In the view of the CBO, "adopting a fair-value approach would provide a more comprehensive way to measure the costs of federal credit programs and would permit more level comparisons between those costs and the costs of other forms of federal assistance." But the CBO is compelled by Congress to follow FCRA procedures in all of its official budget estimates. The CBO does produce fair-value estimates on occasion — that's how there came to be two separate cost figures during the student-loan debate — but the FCRA method is always used as the official estimate, and only an act of Congress can change that.

UNDERSTATED COSTS AND PERVERSE INCENTIVES

The push for fair-value accounting has been most closely associated with student loans, and for good reason: The difference in costs between the two accounting measures is larger for the student-loan program than for any other credit program. But fair-value accounting would fix accounting problems in a wide range of federal credit programs. The CBO estimates that, after accounting for taxpayer risk exposure, the federal budget is actually about $56 billion larger per year than the official figures suggest, due to the accounting rules for credit programs.

The effects of these flaws are widespread. Consider the Federal Housing Administration's single-family mortgage-insurance program, which provides default guarantees to lenders making home mortgages to first-time and lower-income home buyers. These home buyers secure subsidized mortgages, which are loans with terms better than any private lender would offer without the government guarantee. Because FCRA rules exclude a market-risk premium, the program appears to both subsidize homeowners and generate profits for the government, "earning" a $60 billion free lunch for the government over ten years. But once a market-risk premium is added to these tallies, the loan guarantees show a $3 billion annual cost.

The Troubled Asset Relief Program is another high-profile example of the dangers in our current rules. It authorized the government to buy assets so "toxic" or risky that no private-sector entity would go near them as the recession took hold. Under normal conditions, a private institution would charge an extremely high interest rate to cover the riskiness of buying these assets, if it wanted to assume the risk at all. The government, however, purchased preferred stock from (i.e., made loans to) teetering financial institutions at about one-third the market interest rates. Under FCRA rules, the government would have claimed an immediate profit of 11% in late 2008 and early 2009 when it brought these risky assets onto its books.

Fortunately, due to a Congressional mandate, TARP is one of the few programs in which budgetary impact is officially calculated using the fair-value method. A 2009 analysis by TARP's Congressional Oversight Panel (chaired by Elizabeth Warren, now a senator from Massachusetts) found that when market risk was factored in, the government actually provided a 22% subsidy to the struggling financial institutions. In dollar terms, this works out to be a taxpayer subsidy of $43 billion.

If it is valid to ignore the market risk that government credit programs impose on taxpayers, then the government could buy out almost any private-sector loan at its current market price and claim an immediate profit. That profit would be equal to the market-risk premium that the private sector had priced into the loan's interest rate but that the government would earn for "free." In other words, the government would buy the loan at the market price and simply put it on the books at a higher value. That's budgetary alchemy, but it's required by law under FCRA.

Legislators have exploited the shortcomings of the FCRA method in perverse ways. For example, in the depths of the recession, Ohio senator Sherrod Brown proposed that the federal government buy up private student loans, convert them to federal loans, and then reduce the interest rates that borrowers pay. Lenders holding the loans would be paid face value for them — that is, the government would pay the lenders the full outstanding balance on the loans. Borrowers would receive new, better terms and repay the remainder of their loans to the Department of Education. The CBO was required under FCRA to show that this transaction would result in an immediate $9.2 billion profit to the government.

Bear in mind that this was a debt swap in which borrowers would pay less interest to the government than they would pay to private lenders. But, miraculously, $9.2 billion in new cash for the government would appear out of thin air as soon as the transaction was made. This money could then promptly be spent on more government programs. Accounting for market risk shows that the transaction would actually have cost about $700 million.

FCRA's flaws have given members of Congress a ready-made justification for even more spending on credit programs. Last year, none other than Senator Elizabeth Warren called for lowering the interest rate on federal student loans because FCRA was showing a profit on those loans. "This is just plain wrong," she said in a floor speech. "The government is making obscene profits on these loans — profits we can and should cut back on to help our kids who are struggling to pay for college." Of course, in reality the government was subsidizing students to the tune of $95 billion over ten years, and Senator Warren was actually proposing to increase the subsidy even further. Nevertheless, the government's student-loan "profit" pervades popular commentary on the subject, with journalists and political activists often resting entire policy arguments on the myth.

Another way lawmakers have gamed flaws in FCRA is through what is known oxymoronically as a "self-pay" loan program. In 2005, Congress created a loan-guarantee program to back private financing for green-energy projects. The program would supposedly pay for itself by charging fees to loan recipients in exchange for the guarantees. As a result, Congress didn't need to budget a single penny to guarantee billions of dollars in loans.

It may seem strange that a private business would eagerly pay the government a fee in exchange for a subsidy of equal value, but that is business-as-usual for FCRA accounting. FCRA rules ensure that the fee on a self-pay loan guarantee will be underpriced relative to what it is worth to the company developing green-energy technology. The fee will be less than what private lenders would charge for the same loan, by a value equal to the market risk that FCRA ignores. For one type of loan under the program — loans for the construction of nuclear power plants — the CBO reports that the fees are below fair value by 8% to 16% of the amount borrowed. The nuclear-power industry, however, brags that the loans it receives "are executed at truly no cost to the government."

Collectively, these schemes would leave taxpayers on the hook for billions of dollars in the event of another economic downturn, but the budget does not account for this risk exposure. Moving to a fair-value accounting system would incorporate the price of risk, more accurately representing the cost of the subsidies that government credit programs give their beneficiaries.

PUBLIC PENSIONS

While the above examples focus on federal programs, state and local governments are responsible for the single biggest accounting offender in the United States — the public pension system. Disputes over pension accounting are regularly featured in news stories about state budgets, but the connection between pensions and student loans has been underappreciated. The accounting principles at stake are, in fact, identical.

Public pensions are designed to be fully funded, meaning that plan administrators set aside sufficient money each year to eventually pay the future pension benefits that active workers earn in that year. Administrators put these annual pension contributions into an investment fund. The combination of the annual contribution and the interest earned on that contribution is then supposed to pay for the future benefits that have been promised to current public workers.

About three decades ago, pension administrators began gradually chasing higher returns by taking on more risk in their investment funds. As pensions sold bonds and bought stocks, the expected return on their investments increased, as did the risk, reflecting what is called the "equity-risk premium." From a fair-value perspective, that risk-return tradeoff is a wash, with no impact on projected liabilities. After all, millions of knowledgeable investors around the world forgo the greater expected return of stocks in exchange for the safety of bonds. It would be nonsensical for the government to simply exchange bonds for stocks and declare a net asset gain — that would imply that one dollar in stocks is worth more than one dollar in bonds. Nevertheless, states claim exactly that: They believe they can reduce their required pension contributions without cutting benefits, just so long as they put more of the money in stocks.

That extra risk comes at a cost — pension funds might achieve the expected return, but benefits are a guaranteed obligation of the government, and taxpayers are responsible for any shortfalls. Nevertheless, when calculating pension liabilities, administrators discount future payments at the expected rate of return on those risky portfolios, typically around 8%, and do not incorporate market risk. State pension administrators are assuming, in effect, that expected returns on pension investments are certain, just as the federal government assumes that expected student-loan repayments are certain. By failing to apply fair-value accounting to their pensions, state and local governments are effectively concealing trillions of dollars in pension liabilities.

Government accounting for pensions is a mirror image of its accounting for student loans. Whereas state and local pensions treat their liabilities (the future pension payments they must make) as uncertain even though they are guaranteed, the federal government treats its student-loan assets (the expected loan repayments) as guaranteed even though they are uncertain. In both cases, costs to the government are made to appear lower than they really are.

As discussed earlier, official FCRA accounting enables the federal government to claim a "profit" simply by purchasing a private-sector loan. In the pension world, the analogous transaction is the "pension-obligation bond," which allows states to conjure money through an interest-rate arbitrage scheme. In essence, a state sells a government bond that pays, say, a guaranteed 5% interest rate and then places the proceeds from the bond sale into the pension fund. The trick is that the pension fund is assumed to return 8%, so the state nets 3% per year in "free" money. The fallacy, of course, is that the pension fund's 8% expected return carries risk — which is why investors are willing to buy the (safer) pension-obligation bonds in the first place.

It is illegal for private pensions and corporations to hide the cost of risk as public pensions do. In a recent speech, Commissioner Daniel Gallagher of the Securities and Exchange Commission made that point explicit: "[F]or years, state and local governments have used lax governmental accounting standards to hide the yawning chasm in their balance sheets....In the private sector, the SEC would quickly bring fraud charges against any corporate issuer and its officers for playing such numbers games."

A clear double standard exists here. When private businesses offer a pension, federal law requires them to make regular contributions to their pension funds, accurately measure their liabilities, and pay insurance premiums to the Pension Benefit Guarantee Corporation. Public pensions do none of those things, in large part because they do not have fair-value accounting to rein them in.

CRITIQUING FAIR VALUE

Despite strong support from the CBO and most finance economists, the idea of incorporating fair-value accounting into government budget practices faces opposition from the Obama administration and left-of-center think tanks (although they have not always opposed it — a point we will return to later). Some of the criticisms have a superficial appeal, and it is worth examining each in turn.

The first and most obvious criticism concedes that while market risk is real and imposes a kind of cost on taxpayers, it does not necessarily impose a budgetary cost, and thus should not be included in budgetary estimates. Since the price of risk is something that the government may never have to pay, as this argument goes, it has no place in accounting ledgers that are supposed to reflect actual revenues and expenditures. Budgeting with fair-value accounting would sow confusion because the budgetary cost (including market risk) would be higher than the projected cost.

This argument seems reasonable at first glance, but it actually misses the point. Consider the estimated cost of TARP mentioned earlier. The 2009 report of TARP's oversight panel estimated a cost to taxpayers of $43 billion for the subsidies given to the bailed-out banks, but by the end of 2013 the treasury had recouped more than it had paid out. So was fair value wrong about TARP? No. Russian roulette is not a risk-free game just because the gun's first chamber turned out to be empty. The $43 billion was the cost of entering into the obligation, based on the risk understood before the government knew how TARP would turn out. The eventual net gain is the hindsight value, known only after the risk has become irrelevant.

Fair value is not a prediction of anything. It simply reflects the market price of risk. In fact, the primary reason for fair-value accounting is that no one can predict the future with certainty. If another financial crisis occurred today and Congress had to debate whether TARP should make another round of loans, lawmakers would again need to account for risk to reflect the immediate cost of the program.

The government's risky investments do not always pan out as TARP did, and public pensions offer a perfect example. As noted earlier, pensions began pursuing higher rates of return by buying stocks rather than safer bonds. This move may have seemed like a wise choice in the 1990s and mid-2000s when the stock market was growing at a good clip. But then the 2008 stock-market crash occurred, and the value of pension investment funds plummeted, forcing spending cuts and tax hikes at a time when the public could least afford them.

Fair-value accounting would not outlaw risk-taking, of course, but it would require the government to be honest about the cost of an uncertain future. While unexpected events may never affect the government's bank accounts directly, they certainly do affect the government's financial position in the present. That is why the cost of market risk must be included in the budget.

The second objection offered by opponents of fair-value accounting is that the government has special advantages over the private sector. They say the government is less risk-averse than private actors because it can spread risk out across millions of taxpayers and across generations, unlike individuals who often need money at a particular moment. The Center on Budget and Policy Priorities, for example, has argued that low treasury borrowing rates in the midst of a weak economy reflect the government's ability to tolerate more risk: If the government can borrow inexpensively even during recessions when money is scarce, as this reasoning goes, then the ups and downs of the marketplace should not be worrisome.

But market risk cannot be diversified away, and spreading it across millions of taxpayers does not reduce its magnitude. As for government bonds, the market values them as risk-free because the government can raise taxes when it needs money. The government's power to tax assures investors that they will always receive the promised interest payments on government bonds. So the government does not make risk disappear; it merely passes it on to the taxpayers. As the CBO has noted, "When the government finances a risky loan or loan guarantee by selling a Treasury security, it is effectively shifting risk to members of the public." If government receipts from risky assets come in lower than expected, it is taxpayers — a collection of risk-averse individuals — who must make up the difference.

Third, defenders of the current arrangement often point out that the government, unlike a private investor, is not in the business of generating a profit off of these credit programs. This point, even if true, is irrelevant. Just because the government is not explicitly pursuing profits does not mean that it is immune to market risk. If non-profit entities could ignore market risk in how they price assets, then credit unions, public pension funds, charitable endowments, local governments, and countless other non-profit investors would have gobbled up all of the world's credit, considering it underpriced due to a special premium for "for-profit risk bearing." No such premium exists.

Fourth, both the CBPP and the Obama administration argue that fair-value accounting would lead to inconsistent treatment of risk. Programs like Medicare and Social Security have future costs that are uncertain, they point out, so why not risk-adjust those as well? But there is no inconsistency here. The risk that entitlement programs may be more costly than expected is analogous to the uncertain costs of future loans disbursed by the government. Fair-value accounting applies to neither of these cases. A disbursed loan is a contractual agreement that represents an immediate change to the government's assets and liabilities, so fair-value accounting applies. By contrast, Congress is free to change entitlement policy before entitlement benefits are disbursed, meaning taxpayers are not yet subject to any market risk. As the CBO points outs, it is actually the absence of fair-value accounting that creates budgetary inconsistency, with cash transactions priced at market levels and credit programs priced below.

Finally, the Obama administration worries about the cost of the complicated analyses necessary to generate fair-value discount rates. Unless one is already convinced that fair-value accounting is a bad idea, however, it is difficult to take this objection seriously. The administration's argument is essentially that it's expensive to calculate the price of market risk, so the price is assumed to be zero. This position is akin to assuming there is no risk of a meltdown at a nuclear plant because figuring out the exact risk is too much work. As with all budget estimation, fair-value precision will vary depending on the resources invested. How much to invest is a question for the political system to answer, but assuming an obviously wrong number — a market-risk cost of zero — is not a sound approach.

FAIR VALUE'S FAIR-WEATHER FRIENDS

At the federal level, the CBO has been working to show lawmakers the flaws in the FCRA by producing a number of alternative estimates using fair-value methods. The CBO has also argued that public-pension liabilities should reflect fair value. Its position is in line with finance experts opposed to the exclusion of market risk from government cost estimates, including scholars with the Federal Reserve, the Financial Economists Roundtable, and the Simpson-Bowles fiscal commission.

The CBO is an impartial observer, though, and in the political arena conservatives have led the charge lately for fair-value accounting. Congressmen Paul Ryan and Scott Garret have sponsored legislation to amend the FCRA so that estimates reflect market risk. Left-leaning lawmakers, think tanks, and advocacy groups oppose the change, as does the Obama administration, which makes fair-value accounting seem like a "conservative" position. But that has not always been true. Some of today's fair-value heroes were the villains of the same story less than a decade ago, and vice versa.

The Social Security "privatization" debate during the late 1990s and 2000s is a standout example of how political activists have selectively applied fair-value accounting depending on their policy preferences. The key issue in the debate was whether Social Security's budgetary health could improve if payroll taxes were invested in stocks or other financial assets. In some cases, the proposal was for the government itself to invest the funds and boost the return that the program could earn. In later debates, the focus was on allowing individuals to invest their own share of contributions to the program in private accounts. Either way, the justification was the same: Investments in financial assets, particularly stocks, offered much higher returns than what the Social Security trust fund (notionally investing in U.S. Treasury securities) would earn.

One popular study by the Advisory Council on Social Security from 1997, right at the beginning of the dot-com bubble, assumed that government bonds would return 2.3% annually, while stocks would return 7%. No risk adjustment was included in the projections. As economists John Geanakoplos, Olivia Mitchell, and Stephen Zeldes pointed out, that would mean every $100 of Social Security funding shifted from bonds to stocks would instantly (and implausibly) transform into an investment valued at $385. The apparent gain from this transaction comes from the same flawed FCRA rules that make government loans appear profitable when they are not. Stocks pay a higher return than bonds because they require investors to take on more risk. Supporters of investing Social Security funds in the stock market disregarded that risk, just as supporters of the current accounting method for student loans and pension portfolios disregard the risk inherent in them.

As noted earlier, the government cannot become wealthier by simply exchanging bonds for stocks. Conservatives rightly criticize public pensions for indulging that absurdity, but this was exactly the free lunch pushed by Social Security privatization advocates on the right. Here is Stephen Moore arguing for private accounts on behalf of the Cato Institute in 2000:

If a middle-age worker — man or woman — younger than 30 were permitted to invest his or her payroll tax dollars in a safe [sic], diversified mutual fund account, and that account earned 6 percent per year, which is below the historical average for the financial markets, the worker would have more than $1 million in a retirement nest egg at the age of 65. That nest egg would pay a monthly annuity payment, not 2, not 3, but 4 times higher than the benefit promised from Social Security.

And yet, writing in the Wall Street Journal a decade later, Moore suggested that the Illinois public pension system's expected return was "pure fantasy," and he promoted the fair-value estimate of pension liabilities over the official state estimate. One could just as easily characterize Social Security privatization's fiscal benefits as "pure fantasy" on identical grounds. The Heritage Foundation, where Moore now works, has experienced the same cognitive dissonance, hyping the higher returns on private accounts while endorsing fair-value accounting for federal loan programs.

There are certainly some legitimate reasons to endorse private accounts — we are supporters ourselves — but selling them as a budgetary panacea is deeply misguided. Ignoring market risk in private retirement accounts is just as imprudent as ignoring it in public pensions and in student-loan repayments. Regardless of the specific policy question at hand, the accounting arguments here are effectively interchangeable.

Liberal groups are just as guilty as conservatives on this front, if in reverse. During the Social Security debate, it was Democrats and left-of-center think-tank experts who argued for respecting the cost of market risk. One of the most effective arguments came from Peter Orszag, who would later become director of the Office of Management and Budget under President Obama. In his 2002 book with Peter Diamond, Saving Social Security, Orszag called the comparison between stock and bond returns "fundamentally misleading" because it "ignores the greater riskiness of stocks than of bonds." Orszag's reasoning never carried over to the Obama administration, however; after he left his job with OMB, the agency began to defend explicitly the continued disregard of market risk in all federal loan programs.

Other Obama administration economists agree with Orszag but have had no effect on the administration and other liberal organizations' opposition to fair-value accounting. Jason Furman, the current chairman of the president's Council of Economic Advisers, wrote a paper in 2005 for the liberal Center on Budget and Policy Priorities that was devoted almost entirely to debunking the higher-return claim made by Social Security privatization supporters. In the paper, Furman appealed to the CBO and its argument for risk-adjusting costs using fair-value methods.

Today, however, the CBPP flatly disagrees with the CBO and Republicans who are advocating for a shift to fair-value methods. The inconsistency is obvious: The CBPP endorses fair-value in the context of Social Security privatization but rejects it for federal credit programs. Logically, of course, the principles on which fair-value accounting rests cannot be valid only some of the time.

The self-pay loan-guarantee program discussed earlier provides a more recent example of Democrats having supported fair-value accounting before they were against it. Former congressman Dennis Kucinich, an opponent of nuclear power, used fair-value calculations to make the case against loan guarantees for nuclear power plants. He personally requested that the CBO calculate the full cost of the government subsidies in the self-pay loan program by factoring in market risk. When the CBO released the results in 2011 showing that the self-pay loan guarantee provides recipients with large government subsidies but appears free on the government's books, Kucinich and his Democratic colleague Elijah Cummings wrote a letter to the energy secretary: "In light of CBO's findings, the continued use of this current model [under FCRA] to establish credit subsidy fees for nuclear reactor construction projects could expose taxpayers to the risk of billion dollar bailouts."

Nevertheless, Congressmen Kucinich and Cummings are not supporters of fair-value accounting in general. They voted against budget-reform legislation that would have amended the FCRA to incorporate market risk. And neither lawmaker has voiced concern about the distortions FCRA creates for other, much larger federal credit programs such as student loans, nor have they expressed concern that state and local pensions are dramatically underfunded because of their exposure to market risk. Their dalliance with fair-value accounting appears to have been a one-time event intended to expose the higher costs of one federal program they oppose.

HONEST SELF-GOVERNMENT

Government accounting has been used as a partisan tool by conservatives and liberals alike. Political players have essentially latched on to whichever budget estimates support the stories they want to tell, without regard for consistency in accounting principles. That inconsistency exacerbates a fundamental problem with the way the government budgets for credit and insurance programs. Risk has a cost, and it has a cost regardless of whether it is borne by an individual investor or by taxpayers collectively. When the government disregards the market risk associated with its credit programs, it understates costs and misleads the public.

While fair-value accounting may be an arcane issue, it addresses something central to democratic governance: the need for transparency and honesty about the real cost of government programs. When the government places a value on a risky asset that is greater than its market value, it is effectively buying a widget for $100 and immediately declaring its value to be $110.

If Congress were to mandate fair-value accounting across the board, it would take a big step toward a more honest and transparent budget process. When crafting legislation and examining existing programs, Congress would face squarely the risks it assumes and the subsidies it provides through credit programs. Legislators may ultimately decide that these risks and subsidies are in the public's interest, but under the current budgeting rules they do not see the potential consequences of their policies in the official analyses, to the detriment of honest self-government. This should change, and adopting fair-value accounting would be a big step in the right direction.

Jason Delisle is director of the Federal Education Budget Project at New America, a think tank in Washington, D.C.

Jason Richwine is a public-policy analyst in Washington, D.C.