Is Bubble? Is Not Bubble? (Part 2 of 2, with debt charts)


In Part 1, I distilled student debt bubble skeptics’ points (Annie Lowrey, the Economist, and Baum and McPherson) This post will conclude and add some research on student debt.

There are two problems I have with the skeptics’ thinking that haven’t come up yet.

1). If student debt isn’t a bubble, then what’s the solution?

The Economist says nothing. Baum and McPherson, who think student debt surpassing credit card debt is irrelevant, whistle Dixie past the problem, so they’re of no help. Lowrey’s attitude, though, is simply disturbing.

For better or worse, students cannot discharge college loans through bankruptcy.

Anyone with common sense should conclude that the undue hardship exception is a terrible idea on both economic and moral grounds. The only people who could counsel maintaining a lopsided system that allows banks to take on zero risk are banks and universities. It is indefensible.

My solution for student debt: cancel all government guarantees, return all consumer bankruptcy protections, and any banks that fail get nationalized by the government, which then wipes out their bad assets, breaks them up, resells them to the public, insures the depositors, and convicts anyone who committed fraud.

But wait, that’s the same solution I’d give to anyone asking me about the solution to the housing bubble, or the Japanese real estate bubble, or the Great Depression. If it looks like the housing bubble, and the solution’s the same, then why is student debt not a bubble? The remaining questions are how big is it, and what’s its effect on the economy?

2). Student debt is inhibiting economic recovery.

Anyone who knows anything about the American economy will tell you that after the housing bubble burst, households began paying down outstanding debt. The problem is that when everyone does this simultaneously, you get two things you don’t want: (1) nonexistent growth, for growth = spending = income, and (2) disinflation because in principle every loan issued is inflationary while every loan paid off is deflationary. These two factors both contribute to high unemployment. In short: a depression.

Take a look at this graph of three components of household debt: mortgage debt, revolving debt, and nonrevolving debt (which includes student loans).

You can see the severity of the mortgage bubble and the ongoing market correction. You can also see that nonrevolving debt is still growing while the others aren’t.

When we discuss debt, we should expect it to always be at a nominal record high since output is also usually at a nominal record high, so to see where we’re at, here’re the debts as a ratio of GDP:

(Sources: GDP (BEA), Nonrevolving, Revolving, and Total Consumer Debt (Federal Reserve, G.19 Release), Mortgage Debt (Federal Reserve, 1.54 Release (2009-2011, archives 2004-2008 (the February updates show the data from five years earlier)); note that the last data point is the 1st quarter 2011 (annualized), so the graphs are slightly distorted but project 2011’s trend)

Notice how in 2007-2008 all debts start deleveraging while nonrevolving debt does not? While we don’t know the exact amount of student debt, we do know who owns it, courtesy of FRED.

Your first thought might be, “Holy shit! Look at how much nonrevolving debt Uncle Sam picked up since 2009!” Ah, but according to the Fed, much of that’s actually attributable to a change in accounting rules:

The shift of consumer credit from pools of securitized assets to other categories is largely due to financial institutions’ implementation of the FAS 166/167 accounting rules.

By “other categories,” the Fed means the federal government. More subtly, the government category included Sallie Mae until it was privatized in 2004, hence the bulge in “Finance Companies” back then. Here’s the same data in a stacked chart:

The new accounting rules went into effect by January 2010, but even then, the federal government is the only owner of nonrevolving debt whose holdings are growing. Most of the $150 billion in the last year and a quarter are likely student debt (unless the government is buying up home equity lines of credit (HELOC) as part of the Fannie/Freddie takeover, which I doubt). Please gaze.

The Federal Reserve released its quarterly G.19 Release a few weeks ago. By the end of March 2011, nonrevolving debt grew at a 5.6% annualized rate, which is bad enough, but the government’s share grew at an annualized—I kid you not—58.8%, from $316.4 billion to $355.2 billion (or 19.7% to 21.8% of total nonrevolving debt; 2.16% to 2.41% debt-to-GDP). That’s in just one quarter!! Revolving debt continued dropping, except in March. Meanwhile, the bad news is that GDP only grew 1.8% first quarter.

Government holdings of what are likely student loans aren’t taking over the financial system, but their rapid expansion against a depressed economy is very bad. I emphasize that the remaining student debt is held privately, though the quantities are jumbled with the other kinds of debt they own (HELOCs, vacation loans, auto loans, etc.). We only know that their aggregate holdings are decreasing, for now.

I can’t imagine any instances when debt growing faster than output is good, aside from some kind of significant long-term investment that doesn’t pay off in the short run, which we know isn’t the case here. The bubble skeptics don’t see these problems. I will watch what they say as growth stagnates and student debt and resistance increase.


  1. Matt, Chris Martenson has highlighted this exponential growth in debt. Anyone with an IQ above room temperate – and a sense of honesty – knows that real wages have stagnated or slightly declined for the average worker, over the last 30 years.

    Policymakers and politicians keep telling us that “higher education” is the key. It is implied that the benefits will go to the students. In reality, the administrators at colleges and universities are making money hand over fist.

  2. It is not due to a change in accounting rules. It is due to federal student loan reform. If you look at the G19 notes from 2004-2006, the Fed was only counting two categories of federal student loans: outstanding direct loan principal and outstanding Sallie Mae on-balance-sheet (non-securitized) principal. It has nothing to do with when Sallie Mae formally phased out its GSE function. In fact, the Fed only began counting it AFTER Sallie Mae privatized, not before.

    [You’ll have to provide a link to the “G19 notes” you refer to. I see no such thing on the Fed’s historical releases.]

    The increase in recent quarters is because of the shift to Direct Loan. After July 1, 2010, all new federal student loans are Direct Loans. No new guaranteed student loans. Even before that, during the 7/1/09/6/30/10 award year, colleges were already panicking about the instability of guaranteed lending, despite the successful liquidity program (bailout program) which actually enabled FFELP volume to grow from 07/08 to 08/09. Many colleges were shifting to Direct Lending even before the new law took effect; FFELP volume, in an unprecedented event, actually declined from 08/09 to 09/10.

    Bottom line: the G-19 student loan category during 2004-2009 contained a fraction of outstanding federal student loans because the vast majority of outstandings were in FFELP, not DL, and G-19’s only sliver of FFELP was Sallie Mae on-balance-sheet outstandings. New loan volume has increased very little. It has simply shifted from one bucket (FFELP) to another (DL).

    [You will have to link to this “G-19 student loan category” you refer to.]

    The huge missing gap of FFELP loans held by lenders (banks, state gov’t agencies, nonprofits and other non-bank lenders) other than Sallie Mae, not to mention the defaulted loans held by guaranty agencies and DoEd, is probably why the Fed changed the category title from student loans to “federal.” (And why doesn’t it include other types of federal loans then — VA, USDA, ExIm, HUD, etc?)

    Keep in mind that the Fed also doesn’t understand that OPB (outstanding principal balance) is not the total outstanding balance. The Fed refuses to include OIB (outstanding interest balance) in its figures, despite the fact that the borrower legally owes this amount. (While occasionally a retroactive deferment or forbearance will erase the accrued interest [OIB], this is generally a temporary benefit, because interest will begin to accrue again.) Accrued interest (OIB) is generally very small on the non-defaulted portfolio but is quite significant on the smaller, defaulted portfolio.

    [Interesting point, but cite sources]

    1. These Federal Reserve G-19 notes reflect the sequence of events in the changes of FR’s definitions:

      Click to access attachment2.pdf

      This is FR’s attempt at presenting stock values (snapshots of outstandings at a particular point in time). To show that this is just part of the total outstanding DL + FFEL universe you will have to look at the actuals in the annual President’s budget, or you can look at the audited financials and piece together the financial notes which have the outstanding balances from the different parts of the two programs, i.e., balances with lenders, balances with guaranty agencies, balances with ED, non-defaulted balances, defaulted balances.

      Because of FR’s lack of understanding of what outstanding balance means, there is no “citation” from FR saying that OIBs are missing, i.e., it is likely that FR does not believe/does not understand that something is missing.

      There are many sources for flow values (new volume of student and parent FFEL & DL), for example,

      [Interesting. Thank you.]

  3. There’s no bubble. [You’re invited to explain why tuition is increasing above inflation and why student debt is increasing faster than GDP while all other types of consumer debt are not.] The solution you propose would take us back to the 1950s, when college was largely a class-based, social-sifting mechanism. Admissions at places like Harvard College and Harvard Law School were well over 50%. As Dershowitz stated, in a rare moment of humility a couple of decades ago, the folks who were teaching at Harvard in the 1990s would not even get admitted to Harvard in the 1990s as students. The admissions standards had increased meteorically. That is not to say that everyone who was admitted to colleges, even elite ones, was low-caliber “in the old days.” It was just that a strong message was sent across society that postsecondary education was only for certain classes of people. That is still the way it is in most of the world.

    Before October 2008, it could be said that at least 20 colleges in the US could survive a semester without the availability of federal student aid (federal student loans, Pell Grants, federal campus-based, etc.). That number has surely dropped since then, as even the colleges with the largest endowments have faced cash-flow crises, as they are generally restricted from spending more than 5% of their endowment in any given year.

    Thus, with your reforms, we would return to a class-based postsecondary system, one that includes only those with significant amounts of financial and social capital. The exception, as in the 1950s would be the availability of a small amount of merit scholarships for ambitious white males of lesser means. This would of course be anathema to today’s USA, in which the vast majority of financial aid is based on financial need, not merit. The only change from the “old days” might be that, in the absence of the fed’l gov’t from student lending, students with very high FICO scores and/or co-signers would still be able to borrow on the private market to pay for part of their postsecondary education.

    [You’re assuming college must work as it did in the 1950s. With the advent of things like distance learning, Harvard could easily have online courses with very large enrollments and far lower tuition. The vast sums shoveled to the universities are preventing innovation, not encouraging it. Additionally, (a) college isn’t a good fit for everyone, (b) its unquantified democratizing effects that you appear to favor must be balanced by the quantifiable needs of the labor market, and (c) even if college were closed only to the wealthy due to expense (which I again emphasize would be untrue), that doesn’t mean people who couldn’t go wouldn’t have upward mobility or would be socioeconomic untermenschen; they would be perfectly capable of leading prosperous, productive lives just like many of my ancestors did.]

    While the 2005 bankruptcy “reform” made bankruptcy tougher for everyone, it should be amended so that alternative education loans are placed back on the same bankruptcy procedure as other types of consumer credit. The special treatment of one type of private sector credit does not make sense.

    Howwever, putting them back for federal student loans is a non-starter. Bankruptcy is not a consumer protection. It is a debtor relief valve which is not available for government-related obligations, such as tax debts, child support, social security overpayments, and defaulted student loan debts. Most Americans would almost surely agree in a poll that giving people the authorization to write off these types of debts completely freely would take the country in the wrong direction. Particularly in cases where the borrower cannot prove there were any genuine rip-offs involved.

    [The undue hardship exception was only made perpetual for federal loans in 1998. There’s no principle that prevents the law from being changed back.]

    What is far worse is that implementing 1975-style “free and open bankruptcy” into the federal student loan programs would not have any of the positive effects cited by advocates for the policy. It would not put any “skin in the game” from either colleges or federal agencies. There is no way that Americans will accept a federal student loan program where the student needs a FICO of 720 to qualify to attend college. Nor would taxpayers supply the costly infrastructure to perform that type of individualized underwriting annually for millions of students. And will Americans really want a bureaucracy that reviews which type of postsecondary program (cosmotology, medical assistance, engineering, business administration, sociology, etc.) is “appropriate” for each federal loan applicant and Pell Grant applicant, before advancing the funds?

    [There’s no reason the federal government should be in the student loan business any more than it should be issuing credit cards. Moreover, you’re assuming there are no good alternatives to financing college other than federal student loans. If higher education is so critical, we could fully fund it as European countries do. Or, we could just adopt a human capital contract model for financing education.

    Undue hardship is so hard to prove, because any (or almost any) federal bankruptcy judge can see that you can consolidate into direct loan, get on the income-contingent-repayment plan and then pay $0 per month if your income is very low. Capitalization is limited to 10% lifetime, and then the whole debt is forgiven if the balance still exists after 20 years. How is that “undue hardship”? And this isn’t about “banks vs. people.” The banks are bye-bye. All new federal loans after June 2010 are direct loans. That faux-populist argument is passe. While the direct loan servicers are former-FFEL companies, that was always the case.

    [IBR and ICR are good for whatever number of debtors are eligible, but they still leave taxpayers holding the bag for whatever costs universities feel like charging.]

    One way to pay for a vast, new bankruptcy discharge bureaucracy — and to avoid abuse — might be to allow the bankruptcy discharge on federal loans but then automatically seize 20 cents of every dollar earned for the rest of that person’s life. Provisions would be put in place to prevent underground economy income
    and to prevent living as an expatriate.

    [20 cents on the dollar? Preventing people from leaving the country? This is absurd, and the latter is certainly unconstitutional. You’re assuming that everyone will rush to the courthouse just because they can. I believe Americans are more honest than that and actually want to pay down their debts, except the jobs aren’t there and the incomes aren’t either to justify the degrees’ costs in many cases. The government should be more concerned with creating jobs for poor student debtors rather than finding novel ways to punish them.]

    Another option to put some real skin in the game would be to require annual matching funds from schools that want continued participation in Pell and federal student loan. There is long precedent for this in the federal campus-based programs. This would show that the school and/or state actually had enough faith in its programs to invest some of its own funds in the process.

    [Or they could self-finance out of human capital contracts.]

  4. Tuition has been increasing faster than inflation for a long time — at least since the Vietnam War, when, for most of the time when the draft was still in place, young men could obtain a draft exemption (“deferral”) by enrolling in postsecondary education. Then, when everyone saw the impact on military enrollment numbers, they restricted deferrals to grad school — which, by the way, kicked off the boom in law school enrollments. Before the threat of being sent to Vietnam was around, law school was not something that most Americans were willing to consider. Later on, school exemptions from the draft were eliminated entirely. Nixon had a goal of repealing conscription and moving to an all-volunteer military — he saw ending the draft as an effective way to undermine the anti-Vietnam war movement, since he believed affluent youths would stop protesting the war once their own probability of having to fight in it was gone. By the time they stopped drafting people, American ground actions in Southeast Asia had terminated anyway.

    Why are other types of consumer debt not increasing faster than GNP? Despite TARP and a number of other efforts, banks and non-bank lenders essentially have refused to lend, despite the assurances they made when receiving so-called “bail out” funding. Even where there are generous guarantees available to lenders, such as single-family-home mortgage lending, lenders have greatly tightened underwriting standards. Just as standards were too loose for a couple years there, standards have been too tight since 2007. It is an over-correction which will take time to reach equilibrium. Anecdotally, the same type of “raising of credit standards” has taken place with credit cards, car loans and personal loans. Supposedly, thousands of consumers have received notices that their credit card accounts have been cancelled.

    On the demand side, consumers are tapped out, after a long period of growth in spending. A high rate of employment and under-employment also has greatly reduced Americans’ taste for credit. On the supply side, the international liquidity crisis which took hold during the winter of 2007-08 has not permanently resolved. At least during the past few decades, Americans have never saved enough so that we could fund our own credit needs internally. Consumer credit was driven by investors from Europe and Asia who viewed American consumer spending as a safe, dependable long-term growth market. While it was not as safe as these investors thought it was in 2006, it is certainly not as risky as they have believed since 2008. Someone needs to reassure them. It is not that difficult. If you go through all the securitizations you will see that the media’s meme of omnipresent “toxic assets” is largely a myth. For mortgages specifically, even if the macroeconomy were to become incredibly strong tomorrow, growth in new borrowing may be limited for many, many years by the overhang from market oversaturation during the mid-2000s.

    Part of the growth in federal student loan volume is directly related to the doldrums elsewhere. Growth in use of Pell Grants and federal student loans has always been a leading indicator of recession. When times are bad, people go back to school to re-tool. In addition, those already enrolled tend to stay in school more than normally. Unlike during the dot-com boom, for example, most college students are not tempted to drop out by some great job that is available.

    The other part of the growth in federal student loan volume relative to other consumer debt products is the lack of a liquidity problem. FFELP actually got a bail-out which began months and months before Americans first heard about things like TARP. While we will never know whether the threats were real, FFELP lenders raised the specter in late winter 2008 of students unable to enroll in fall 2008 due to lack of loan capital to make loans to the students. An emergency liquidity program was quickly enacted. No one has ever reported that any student was unable to enroll (or remain in school) due to inability to obtain a FFELP loan in 2008-10. In fact, the bail-out program was so successful that FFELP volume actually increased during 2008-09 over 2007-08. Almost 100% of FFELP consolidation came to a grinding halt, as none of the liquidity programs included consolidations. The schools simply wouldn’t allow policy makers to take their eyes off the ball — the school’s habitual focus on current and future students, rather than former students. While FFELP volume declined in 2009-10 as 100% direct lending seemed an inevitability to many colleges, it was still quite high, bolstered by the emergency liquidity program. And today, with all colleges in direct lending, liquidity is not only readily available, but no budgetary appropriation is required, as the program makes money on a standard accrual basis.

    1. “You’re assuming college must work as it did in the 1950s. With the advent of things like distance learning . . . ”
      All evidence to-date indicates distance learning is more expensive than classroom learning, not less expensive. Part of that is due to the tendency of traditional universities to use distance education as “full-price” cash cow to cross-subsidize scholarly research and other operations. Part of it is due to the types of other schools which tend to utilize distance education quite heavily — for-profit schools. There is no need to throw the baby out with the bathwater. The fix here with the for-profits is very simple — reinstate the 50% rule, which was repealed in 2006.
      The thing possibly contributing most to the outrageous cost of distance learning is “innovation” itself. There are no standards, no common platform, no shared code. As long as every vendor has the incentive to develop its own, shiny new platform, utterly incompatible with everything else, cost will be an issue. Eventually, industry and the postsecondary community will coalesce around one or two “best idea” platforms, and costs will probably decline.

      There’s no reason the federal government should be in the student loan business any more than it should be issuing credit cards.”
      Not true. If you believe there should be loans, then there unfortunately has to be a federal role. It sounds like you don’t believe in loans, though. In addition, these are not pure credit products. It is a hybrid social program/loan program. Before the fed’l guaranteed student loan program, a variety of approaches failed to provide access to loans. And this was back when Americans saved a lot more and the credit markets were somewhat healthier than today. There were a couple of successful exceptions — states which were able to run programs without federal subsidies. Some wealthy colleges were able to fund their own internal loan programs as well.

      “If higher education is so critical, we could fully fund it as European countries do.”
      As costly as those systems are, ours would be exponentially costlier. To be similar, we would have to move to a tracked system, where the system decides very early in life whether someone is “college material.” The current American postsec system offers nearly-endless second chances. There are additional distributional problems with the European model, such as the social classes who work and pay taxes that fund the free tuition but generally would never be allowed to participate. And I am sure you would be interested in the German and French systems where law school is not only free but there are mandatory paid internships in (1) prosecution; (2) judicial; and (3) private practice. Maybe it is just me, but it turns my stomach that a janitor would be paying high taxes in these countries so that a fifth generation lawyer could receive such a heavily-subsidized education.

      Or, we could just adopt a human capital contract model for financing education.
      Sounds worth trying.

      “IBR and ICR are good for whatever number of debtors are eligible, but they still leave taxpayers holding the bag for whatever costs universities feel like charging.”
      Because the federal loan program makes money, there is no bag. One programmatic “correction” which needs to be enacted is that the debt discharge needs to be made tax-free. Although it will be a while before a borrower reaches that point, the fact that the ultimate loan forgiveness under these repayment plans is still in the tax code as taxable is probably discouraging borrowers from choosing these repayment plans.

      “whatever number of debtors are eligible”
      Good point. These plans really only work when everyone has to participate. Then the subsidy for the needy borrower can be made more generous. The American system offers a stingy borrower subsidy, particularly in ICR, and particularly for borrowers with small households. For example, in Australia, it is very simple. Everyone is in an income-based repayment plan, and the monthly payments are collected automatically through the national tax agency. In the USA, we largely base subsidies (Pell Grants, subsidized Stafford loan eligibility) on how your parents did in life. Does that make sense? In several other countries, the subsidy is based on how the student does in life (after leaving school). For some Americans, the idea that the more successful borrowers effectively cross-subsidize the less-successful borrowers sounds “socialist.” Some other capitalist nations seem to have no problem with this, though.

      We only reached the full “undue hardship” standard gradually in stages for federal loans. Before 1998 there were several years when the standard was 7 years of active repayment or undue hardship. Before that there were several years when the standard was 5 years of active repayment or undue hardship. Before that the “years” option was pure years after entering repayment, and time in deferment and forbearance could count; it didn’t need to be actual attempts at active repayment. In 1998 the change was one of several “savers” to help pay for continuation of lender subsidies for another five years. To loosen the standard and reverse the tightening of the federal student loan bankruptcy standard would require coming up with some money elsewhere in the program.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s