An overwhelming consensus has been developing that the Federal Family Education Loan (FFEL) Program should be ended, and all federal student loans should be made through the Direct Loan Program, where students borrow directly from the federal government. The main justification for this is that direct lending costs less than FFEL, where students borrow money from private financial organizations, typically with some federal subsidy. Proponents cite new figures indicating that the government stands to make a lot of money on the switch, money that will be used to fund more Pell grants.
As someone who likes the idea of bigger Pell grants, and thinks that the FFEL subsidies are a waste of taxpayer money, I can certainly appreciate the goals of these proponents.
But a closer look at their argument leaves me quite worried. Advocates are pointing to a figure of $94 billion that could be saved. This number is derived from the subsidy rate calculations of the loan programs produced by the Congressional Budget Office (CBO) through a process known as “scoring.” Proponents are using the CBO figures to argue that shutting down FFEL will save massive amounts of money -- a strange argument given that a former head of the CBO has explicitly warned against drawing such conclusions.
The warning appeared in "Budget-Scoring Barriers to Efficient Student Loan Policy ," a paper prepared for and presumably financed by groups of lenders. The author was Douglas Holtz-Eakin, the former head of the CBO. The paper describes the reasons why the CBO subsidy rate calculations are not sufficient for making policy decisions. Such decisions should be based on a cost-benefit analysis. Though no fault of its own, the CBO has a good handle on neither of the required totals.
To begin with, the CBO does not even look at the benefits of the programs. Advocates are assuming that both programs have the exact same benefits, which is highly questionable, given what Holtz-Eakin terms “a plethora of anecdotal evidence that private sector lenders offer a portfolio of un-priced borrower benefits (fee waivers, rate reductions, etc.), credit counseling, expedited delivery, superior information technology, college access in initiatives and other enhancements and programs not offered by the [DL], but not easily quantified.” It is also possible that DL provides greater benefits. The point is that we should not assume both programs have the same benefits.
Moreover, the CBO does not look at all the relevant costs. As Holtz-Eakin summarizes, the CBO figures do “not capture the economic cost of the loan programs. This is not a secret. The [CBO] itself has acknowledged the fact” in a 2005 report  stating that “the subsidy calculations … are not designed to fully capture the economic costs to the government … nor do they capture all of the effects of the programs on federal spending and revenues.”
The CBO provides some of the most authoritative, objective and accurate estimates on a wide range of budgetary issues. Thus, if you are going to take issue with their numbers, you had better have a good reason. While I am not qualified to offer a detailed critique, as the former head of the organization, Holtz-Eakin is, and he’s offered a number of reasons to doubt the usefulness of the estimates.
As he explains, these programs are required to be scored according to the Federal Credit Reform Act (FCRA). But the fact of the matter is that DL and FFEL “do not receive equal treatment under federal budget scoring rules.” While the switch to FCRA removed a bias in favor of FFEL, it instituted one in favor of DL. A few of these differences in treatment that lead to bias are explored below.
To begin with, there are risks that are not accounted for by budgetary scoring. The two big ones are interest rate risk (the uncertainty about what rate the government can borrow at) and market risk. Market risk is a broad category that accounts for uncertainty due to fluctuations of the economy. For instance, will the recession push up default rates? While these are real risks with real costs, those costs are “not captured by federal scorekeepers.” This puts FFEL at a disadvantage since they face the cost of insuring or hedging against these risks, while for DL, these risks and costs are simply ignored.
Then there is the risk of programmatic failure (DL had to shut down in 1997, and without FFEL to fall back on, students would have incurred substantial hardship) and indirect taxes (FFEL lenders pay significant corporate income taxes), neither of which is reflected in the budgetary scoring.
The differences that have been getting the most attention are administrative and guarantee costs. These costs were generally not included in the scoring, but some estimates by the OMB and CBO indicate that these costs for FFEL are higher than previously thought. What doesn’t seem to get mentioned is that while taking these costs into account is appropriate, this is a relatively minor source of difference in program cost.
Most importantly, according to an earlier CBO report, the key way in which the programs are treated differently is that for the DL program, “principal and interest payments are discounted at a different, and generally lower, rate than the borrower pays. The result is a net budgetary gain to the federal government that does not exist in the FFEL program.” This gain reflects the fact that the government expects to borrow the money for DL at low rates (0.76 percent in 2010) and charge students 6.8 percent.
This substantial gain would be reported for any program that borrows at the Treasury rate, and lends at a higher one. But that doesn’t mean it’s a good idea. To understand why, note that the exact same logic -- that the government can borrow more cheaply than it lends -- could be used to argue that the government should take over all lending in any market.
Consider an analogy to mortgage lending. Just as with FFEL, there are private lenders that have received subsidies from the government (we’ve already provided Fannie Mae and Freddie Mac $200 billion, and are on the hook for losses on their $5.2 trillion combined portfolio). By the logic of the pro-DL advocates, this subsidization is much more expensive than if the government provided the mortgages in the first place, so why not have the government take over all mortgage lending? I don’t know of anyone who thinks the government should be the only provider of mortgages, but there seem to be quite a few who think such a policy is a good idea for student loans.
In spite of these concerns about the relevance of the CBO figures in comparing the costs of these programs, advocates of switching to DL continue to rely on them. If this is how policy is to be made, then perhaps we haven’t quite put faith based initiatives behind us after all. In the words of former CBO director Holtz-Eakin, “When the budgeted cost of a federal program fails to reflect its actual economic cost, policy decisions regarding that program are likely to be skewed. The federal student loan programs provide a case study.”
Andrew Gillen is the research director of the Center for College Affordability and Productivity.