Two Points on Reducing Student Loan Interest Rates, Featuring: Recoveries, Algebra, Arguments.

Sarah Jaffe has an article up about where or not Congress is going to let the interest rate on student loans double this year.   They are currently at 3.4%, and they are set to go back to 6.8%.  Recommended.  President Obama called for extending this one year in his State of the Union adress.

There’s two points I want to bring up in response.  The second point is a list of additional reasons to keep interest rates low and lower them more.  But before that we need to talk a little bit about recovery rates and why they are so important to understanding how these loans work.

I: Recovery

Aaron Bady, reviewing David Graeber’s book Debt, warns us that when “you quantify a debt with financial precision…you take what was a human relationship of mutual imbrication” and turn it into a “kind of moral dominance, and thereby subject the indebted party to the mechanisms of financial debt collection instead of the precepts of human morality.”  I’d like to go further and abstract financial debt collection into the precision of algebra.

Let’s say you make a loan.  There’s a probability p you get paid back and a probability (1-p) that you don’t get paid back.  If you get paid back you get the loan L, and if you don’t you get the recovery R – think collateral or whatever they can pay up.  The normal way to write out the value of this loan is to take the chance you’ll get paid and multiply it by the value of the loan, the state when things work out, and add it to the probability you don’t get paid but you get the recovery value, the state when things don’t work out.

Writing out this version of the equation emphasizes that the probability of being paid back is what drives the value.  But with a little bit of algebraic manipulation, you can see another way of saying the same statement is that the recovery also drives the value:

Since you get the recovery value in both states (with 100% probability, paid or not paid) a high recovery makes the party in question less concerned whether or not the loan defaults.  This is why unsecured debt like credit cards, where there is low recovery, has a higher interest rate than a home loan – because the “recovery” is taking the home, which has value.

Note that at values approaching 100% recovery (which is the value of the loan itself) we are indifferent to whether we get paid back; at recoveries over 100% we want to force the person to default.  I’ve never seen a recovery rate (and I’ve seen several) over 100% before.  But if you look at the recovery rate pre-collection costs for a student loan, it’s over 100% in situations where there are defaults.  Department of Education FY 2013 Budget, page 31:

The cash recovery rate before contract collection costs (CCC) is over 100%.  This is because there’s a huge collection cost added to the loan, which also collects interest.  The collection costs go to private collection agencies, not to the government – but the government still has a recovery of over 90% on a defaulted loan.

Algebraically we showed above that recovery is a major driver of valuation – and even when you include the losses involve for the government in having to collect a 90% rate can make them indifferent on whether or not they get paid.  Hence little incentive to work out lows, make sure students are aware of their various options, etc.  This 90%+ is the government sees is the result of having student loans being treated to the harshest regime of debt law we have.  And from the student’s point-of-view, going into default adds a significant amount of burden over the lifetime of the loan.

[I understand better the proposals floating out there to keep student loans at 5% of lifetime income.  With that kind of repayment scheme, these 100%+ recoveries (from the point of view of the student) and perpetual indebtedness don’t happen.]

II:  Arguments

So what are the arguments to keep student loan interest rates low?

In a disinflationary period with balance-sheets frictions, with real interest rates negative 10 years out, with the Federal Reserve President urging Congress to find ways to use the law to refinance current underwater mortgages into cheaper loans (“the low mortgage rates that we’ve achieved have not been as effective as we had hoped”), the fact that we are charging interest at all on student loans is dumb.  I’ve proposed mass refinancing of government loans into very cheap rates, and I haven’t heard very convincing counterarguments.  Jaffe’s article has several convincing arguments that add to the case.

Let’s go further.  From the Department of Education student loan overview (R-10):  “For Direct Loans, the overall weighted average subsidy rate was estimated to be -13.91 percent in FY 2011; that is, the overall program on average was projected to earn about 13.91 percent on each dollar of loans made, thereby providing savings to the Federal Government.”  After taking in the portfolio as a whole, including the bad economy through 2011, the direct loan program is a big money-maker for the government.  Lowering the rates will just mean fewer profits, not more losses.

Sarah Jaffe mentions that Jason Delisle at New American Foundation argues that we should let the rates go up in a white paper.  His basic argument is that taxpayers need to be compensated for their risks, that “market risk” might lead to correlated defaults hitting taxpayers harder in bad times, and that if we “discount” (think adjust for risk and time if you are unfamiliar with the term) using the rates the private sector provides student loans at, a technique called “fair value”, instead of the cheap rates the government can borrow at, the program loses money.

Two responses.  The “market risk” as described in the paper is technically just correlated defaults, something the financial engineering community has been modeling for quite some time.  The risk issues are completely covered between interest rate risks (prepayable fixed-rate loans have negative convexity) and the credit risks (if one person defaults it’s likely many more will as well), and since the portfolio as a whole makes a profit credit and interest rate risk is well compensated on the first approximation.

The second involves a fascinating view of neoliberalism.  There are always fights over the appropriate discount rate to use, especially for policy (see the global warming debate for this).  Here I am hearing that we need to use the private market’s estimates of the price.  But note that the private market only exists because of huge government subsidies, removing bankruptcy protections for private loans, among many other things.

So, keeping our neoliberal hats on – the only function of higher education is human capital development, and the only problem is that students can’t collateralize their future earnings from that increased human capital at the beginning of school.  Debt markets are very dependent on recoveries – see why credit cards have a higher interest rate than homes.  Since students can’t collateralize their knowledge, a special type of loan needs to be created.  The government is a uniquely situated agent to create this – its time frame is larger, it can compel repayments thereby handling the agency problems better, its size allows for the lowest rates, it’s lack of profit motive allow for easier payment terms than the “sweatbox” model of private credit lending, etc.  The private market, in contrast, is not.  The private student loan market is incomplete, prone to collapse and very dependent on government goodwill to adjust bankruptcy laws for them.  This private market charges a higher rate.  The neoliberal’s response is to see the private market and go “that higher rate must be the correct one,” even though it exists because of government manipulations.

I’m not seeing the downside on keeping rates lower.  What am I missing?