The $300 Billion Question


© Shutterstock

September 29, 2014 23:58 EDT

How should Congress budget for federal lending programs?

Lending programs create special challenges for federal budgeting. So special, in fact, that the Congressional Budget Office (CBO) estimates their budget effects two different ways. According to official budget rules, taxpayers will earn more than $200 billion over the next decade from new student loans, mortgage guarantees and the Export-Import Bank. According to an alternative that CBO favors, US taxpayers will lose more than $100 billion.

Those competing estimates pose a $300 billion question: Which budgeting approach is best?

As I document in a new report and policy brief, the answer is neither. Each approach tells only part of the story. Congress would be better served by a new approach that fairly reflects all the fiscal effects of lending.

Compared With What?

If lending programs perform as CBO expects, they will bring in new money that the US government can use to reduce the deficit, increase spending or cut taxes. In that sense, taxpayers may come out more than $200 billion ahead.

But these programs do not fully compensate taxpayers for their financial risk. If the government took the same risk by making loans and guarantees at fair market rates — perhaps by investing in publicly traded bonds — taxpayers would make much more. Taxpayers are subsidizing the students, homeowners and companies that borrow through these programs. In that sense, taxpayers come out more than $100 billion behind.

The same issue can arise in personal life. Suppose your aunt asks for a $10,000 loan to start a business. You’ve got exactly that much in a government bond fund earning 2.5%, and she offers to pay 5%. She’s got a good head for business, so the risk of default is very low; realistically you expect a 4% annual return.

The loan sounds like a winner, right? Her 4% beats the bond fund’s 2.5%, if you can handle the risk. But there’s one other thing: Your brother-in-law, equally good at business, would like a similar loan, and he’s willing to pay 6%, with an expected net of 5%.

Now the loan to your aunt sounds like a loser. Your brother-in-law’s 5% beats her 4%. You might still prefer to lend to her, but you would come out behind in financial terms.

The competing CBO estimates reflect this dichotomy. One approach compares the financial returns of lending with doing nothing (the $200 billion gain in CBO’s case, 4% versus 2.5% in yours). The other compares the returns with taking similar risks and being fully compensated (the $100 billion loss in CBO’s case, 4% versus 5% in yours).

Suppose your aunt asks for a $10,000 loan to start a business. You’ve got exactly that much in a government bond fund earning 2.5%, and she offers to pay 5%. She’s got a good head for business, so the risk of default is very low; realistically you expect a 4% annual return.

Both comparisons provide useful information. If you want to predict the government’s future fiscal condition, you should compare the financial returns of lending with doing nothing. If you want to measure the subsidies given to borrowers, you should compare returns with the fair market alternative.

When you discuss your aunt’s proposal with your spouse, you would be wise to mention not only the potential financial gain (“4% is better than 2.5%”) but the subsidy to your aunt (“4% is less than the 5% your brother would pay”). Only then can you have an open discussion of your family’s financial priorities.

Today’s Approaches

The same information is necessary for an open discussion of federal budgeting. But official budget rules, created by the Federal Credit Reform Act of 1990 (FCRA), require CBO to use just the first approach in its budget analyses. Official estimates thus measure the fiscal effects of lending, not the subsidies provided to borrowers. CBO rightly believes, however, that policy deliberations are incomplete without measuring the subsidies, which CBO calculates separately using an approach known as fair value.

Policy analysts have vigorously debated the pros and cons of the FCRA and fair value for years. Neither side has scored a decisive win for a simple reason: both approaches are incomplete. Fair value measures subsidies well, but tells us nothing about fiscal effects; this is its missing-money problem. The FCRA measures lifetime fiscal effects well, but tells us nothing about subsidies.

By recording expected fiscal gains the moment a loan is made, moreover, the FCRA makes lending appear to be a magic money machine. Lending may pay off over time, but the gains do not happen the moment the loan’s ink is dry. Like any lender, the government must be patient to earn those returns. It must hold the loan, perhaps for many years, and bear the associated financial risk.

A Better Approach

For those reasons, I believe we should replace both approaches with a more accurate budgeting method, which I call expected returns. As the report and brief describe, the expected-returns approach forecasts the fiscal effects of a loan by projecting the government’s expected returns year by year, rather than collapsing them into a single value at the time the loan is made, as both the FCRA and fair value do.

Expected returns accurately track the fiscal effects of lending over time, thus avoiding both fair value’s missing-money problem and the FCRA’s magic-money-machine problem. It also provides a natural framework for reporting the fiscal effects of lending and the subsidies to borrowers. Expected returns would give policymakers and the public a more accurate assessment of federal lending than either of the approaches we use now.

*[This article was originally featured on Donald Marron’s blog.]

The views expressed in this article are the author’s own and do not necessarily reflect Fair Observer’s editorial policy.

Ufuk Zivana /


Only Fair Observer members can comment. Please login to comment.

1 comment

  1. Alan

    September 30, 2014

    In evaluating the value of a lending asset, Fair Value Accounting principles require, first, that actual, historical market data on sales of similar assets be used as a benchmark for determining the fair value of the assets if such data exists. For federal student loans, such data exists, since federal loans have been securitized and traded for years, and except for a brief period following the financial crisis in 2008, the market has been active, healthy, and relatively stable. Therefore, any competent fair value accountant tasked with evaluating federal student loan assets would, of course, use this data as the primary benchmark, and look no further. It is, in fact, the best data one could hope for by FVA standards. This is where the CBO analysis goes horribly wrong. Rather than using historical federal loan sales data, CBO chooses to ignore it, and instead chooses to use private student loans as their relevant benchmark. These loans have much higher interest rates, are far riskier, come with none of the guarantees or collection powers attached to federal loans, and the private student loan market is far smaller (less than a fifth the size of the federal market), more volatile, and far more difficult to find reliable historical sales data on generally. Even assuming there was no federal loan sales data available, to use private student loans as a second choice would be a dicey proposition, at best, What is worse, however: The private loan market has always been highly dependent upon the federal loan system in many ways. One example: many if not most borrowers turn to private loans only as a last resort because they could not get federal loans. The lenders know this well, and raise their interest rates accordingly. No credible fair value analyst, given all this, would choose to use private student loans for the purposes of valuing federal loans. But this is what CBO does. As such, the large difference in interest rates between private and federal loans is used to calculate a "loss" for the federal lending portfolio. This explains in large part how a profit estimate of $180 billion becomes a loss of $100 billion under this particular FVA methodology. The best analogy I can think of for this type of accounting is a Big Oil company pointing to a guy selling biodiesel out of his garage for $8/gallon (compared to the Big Oil company's price of $4), and using this to justify writing off $4 for every gallon the corporation sells. Is Big Oil actually losing money? If this guy's brother came up with his own batch that he could sell for $12/gallon, would Big Oil then be able to claim an $8 loss for every gallon it sold? Of course not! The “costs” are not only completely fictitious, there is ZERO chance that they will ever occur based on all available data. Also: years of presidential budget data show that for defaulted FFELP loans, the government’s recovery rate is 122%, a rate far higher than any loan, public or private could ever claim (this is directly attributable to unprecedented collection powers, and the removal of bankruptcy and other protections). For Direct loans, this rate has been about 110%, but I estimate that with FFELP ended, this rate is about 115% . Even discounting generously for collection costs (using cost data for generalized bank loans which typically involve seizure of property, significant court costs, etc.), and factoring in the government’s cost of money leaves a hefty profit on defaults, and in fact shows that the government makes more money on defaults than loans which remain in good stead (a defining hallmark of a predatory lending system). This analysis should be repeated using the correct benchmark, and also the historical data available for the "cost" of defaults. This important issue, the larger policy debate it informs, and the large number of affected citizens warrant at least this much from the CBO.

Leave a comment

Support Fair Observer

We rely on your support for our independence, diversity and quality.

For more than 10 years, Fair Observer has been free, fair and independent. No billionaire owns us, no advertisers control us. We are a reader-supported nonprofit. Unlike many other publications, we keep our content free for readers regardless of where they live or whether they can afford to pay. We have no paywalls and no ads.

In the post-truth era of fake news, echo chambers and filter bubbles, we publish a plurality of perspectives from around the world. Anyone can publish with us, but everyone goes through a rigorous editorial process. So, you get fact-checked, well-reasoned content instead of noise.

We publish 2,500+ voices from 90+ countries. We also conduct education and training programs on subjects ranging from digital media and journalism to writing and critical thinking. This doesn’t come cheap. Servers, editors, trainers and web developers cost money.
Please consider supporting us on a regular basis as a recurring donor or a sustaining member.

Will you support FO’s journalism?

We rely on your support for our independence, diversity and quality.

Donation Cycle

Donation Amount

The IRS recognizes Fair Observer as a section 501(c)(3) registered public charity (EIN: 46-4070943), enabling you to claim a tax deduction.

Make Sense of the World

Unique Insights from 2,500+ Contributors in 90+ Countries

Support Fair Observer

Support Fair Observer by becoming a sustaining member

Become a Member