Flawed Math on Student Loans

Aggregate student loan debt surpassed credit card debt in size for the first time in 2010. Since then, the gap has continued to grow and now exceeds $200 billion. Student loans (at over $1 trillion) are now the second-largest category of consumer lending, second only to home mortgage lending.

While the total pool has been growing, the share owned by the Federal government has grown even faster. In 2000, the government’s student loan was valued by the CBO at $149 billion; now, it exceeds $1 trillion. More than 90% of new student loans are being initiated by the government.

The rapid growth in the government’s portfolio can be traced to several policy changes: 

  • The government chose to largely remove banks from the lending process, following the financial crisis.  The commonly stated objective was to save of $60 billion in fees over ten years (though that number has been questioned by the CBO).
  • A decision was also made to expand the type of lending done without screening criteria.  Typically, student lending done by banks had involved application of some basic credit criteria, even if the government would ultimately own the loans.  Now, for the majority of loans, that is no longer the case.

What is the quality of this massive loan book, and what are the implications for both students and taxpayers?  The New York Times, on March 22 of this year, noted that “many” of these loans “appear to be troubled.” 

Unfortunately, it’s difficult to know exactly how bad the problem is.  The Federal government’s own lending is exempt from the stringent loan forecasting, accounting, and reporting requirements that apply to lending by financial institutions.   The March 22nd Times article noted that the Fed has had to resort to purchasing student loan data from credit bureaus in an attempt to get some metrics on this portfolio.  The Education Department does not provide even basic vintage delinquency data to the agencies that oversee the financial system.   This is ironic, given that Federal reporting requirements for banks have grown massively since 2008, and reporting of more than 100 data elements is now required at the individual loan level on a monthly basis.

Another unknown is equally troubling.  It’s really not clear whether the expertise to manage this kind of portfolio exists within the Education Department.  Consumer lending is primarily driven by technology and analytics.  These tools work best when deployed by staff with deep expertise in management of risk assets.  Has the Federal government had the time (or budget) to invest in the massive loan tracking and management systems that underlie the operations of consumer banks?  Constant updating of data (both from the students themselves and external sources) should be taking place, leading to frequent loan-level modeling of default risk.  This type of modeling could drive targeted rollout of both pre-delinquency and early delinquency programs.  Such programs could potentially aid borrowers before it’s too late.  

Proactive management of this $1 trillion portfolio is crucial, both for the borrowers and the lenders (the taxpayers).  It’s important that the implications of this coming wave of defaults for future Federal budgets be clearly understood.  How much of a loss do we expect to take?

It is particularly hard to know the answer to the last question, as the CBO is required to use an unusual method of accounting for these loans.  They have released quite a few reports noting that fair value accounting would yield a much less favorable assessment of the Federal loan book.  A recent report documented a negative swing of over $200 billion.  Crucially, these estimates are being made without the benefit of true credit quality data, which should underlie forecasting on all risk assets.  The real “hole” may be much worse.  And the problem is just kicking in, as the no-payment grace period is just now expiring on many of the loans made in recent years.

Good intentions followed by poor execution can bring damaging results. 

Some years back, the idea of providing a way for most Americans to own homes sounded very appealing.  But the result was a situation in which many lower incomes households became excessively leveraged and terribly illiquid.  Many were badly harmed when housing prices started to drop, and suffered further pain when the job market tanked.   In its execution, these home ownership programs seemed to end up hurting some of the very folks that they had been intended to help.

More recently, the idea of extending a loan to anyone who qualified for college also sounded appealing.  However, not all courses of study will provide sufficient additional income potential to ensure payback.  Both the economic value of the asset (the degree itself) and the available credit data on the borrower should have been considered when lending was expanded.  Finally, and more controversially: interest rates on Federal student loans should have been varied in relation to the risk of the loan.  This is Risk Management 101.  By not doing this, the government is essentially admitting upfront that they plan to subsidize loans to riskier borrowers, or in areas of study that do not typically bring large returns.  Our national policy on this point should have been debated openly.  College tuition grants should be done explicitly, in accordance with a comprehensive policy framework, not through the back door (and not in a way that demeans students by first having them default on obligations).

It’s worth noting that household debt can be discharged in bankruptcy court.  However, Federally-backed student debt cannot.  Many students who were given loans initiated without appropriate risk assessment face a situation in which repayment will be difficult, and legally available opportunities to reset their obligations will be few.  It seems likely that the rules governing discharge will have to be changed.  As noted, loan forgiveness programs will doubtless be greatly expanded, and a sizable portion of this $1 trillion loan book will likely be written off.   The impact to the lives of the graduates in question will be severe, as discharges and forgiveness programs must be reported to the bureaus.  Credit scores will drop hugely as a result.  The graduates in question will therefore find it harder to get jobs, credit cards, car loans, and even apartments.  Credit scores are routinely checked in relation to many transactions these days, including potential offers of employment.

The longer we wait to face this growing problem, the more future graduates (and taxpayers) we will put at risk.  The bell is ringing.  It’s time to get the math right on student loans.

For a pdf version:

This entry was posted in Regulatory Updates by Kate Flanagan. Bookmark the permalink.

About Kate Flanagan

Kate Flanagan has spent 25 years in consumer banking, most of it with Citibank. She has extensive experience in the credit card business and in consumer lending generally, with particular expertise in risk analytics and technology. Her consumer credit experience spans 40 national markets. In her 23 years at Citibank, she moved between functions and regions to implement rigorous, analytics-based business methods across all stages of the consumer credit cycle (from originations to collections). Since 2010, she has been doing consulting in risk analytics. Kate holds a BA in Math from Brown University and an MBA in Finance from Columbia University.