Sharing the Leverage

by Jay Weiser

Editor’s note: The following is a guest post by Jay Weiser, associate professor of law and real estate at Baruch College. In it, he addresses how policymakers should address consumer debt buildup – and, just as importantly, how they should not.

In their new book House of Debt, economists Atif Mian and Amir Sufi connect the vast increase in consumer debt with the Great Recession and slow-motion recovery. But rather than sing a requiem for a half-century of proxy Keynesianism, when borrowers leveraged up in pursuit of a government-defined American Dream, they shout hosannas for even more debt. They claim that compelling mortgage and student lenders to share equity risk with borrowers will reduce the severity of future financial crises by supporting consumer purchasing power. This is like saying that the overloaded Sewol ferry wouldn’t have capsized if the captain had added one more car in exactly the right place. Consumers and the financial system need less leverage and simpler loans.

Government guarantees and moral hazard

As the authors note, government incentives were a major cause of the consumer debt buildup. Without government backing, as guarantees or direct loans, lenders charge high rates for risky loans and expect high chargeoffs in an economic downturn – the reason why unsecured credit card debt, with its extortionate rates on unpaid balances, was not a systemic problem in the Great Recession. Express and implied mortgage and student loan guarantees undermined market discipline, encouraging risky borrowers to load up on artificially low-rate debt. 

Rather than eliminating government backing, whose benefit has gone mainly to the real estate and higher education industries rather than consumers, Mian and Sufi would add complexity by having lenders assume more of the risk of economic downturns. It’s unclear which lenders they’re talking about, since both student loan and residential mortgage debt have been effectively socialized.  Post-bust, the government backs 90 percent of residential mortgages and is the lender for 85 percent of student loans. If government guarantees remain in place for new loans, politicians will latch on to the supposed additional safety to pump even more loans to riskier borrowers.

Mian and Sufi claim at one point that consumer borrowers are not subject to moral hazard. But consumers have repeatedly leveraged to the max on subsidized loans and bet on good times later to pay them back. Bubble-era 100% loan-to-value home mortgages were premised on rising prices permitting cash-out refinancings. Relying on the exaggerated “college premium” for future earnings, student loans more than doubled from the start of the Great Recession in December 2007 to March 2014 ($1.26 trillion outstanding), notwithstanding a 14.7 percent three-year default rate for borrowers graduating in 2010. If borrowers know that their debt will be reduced during downturns, they will have even more incentive to borrow, bubbles will inflate further, and government losses in downturns will compound.

Doubling down on derivatives

If government guarantees disappear, Mian and Sufi’s risk-sharing requirement – reducing the amount owed when the unemployment rate rises (for student loans) or housing prices drop (for mortgages) drop will cause further distortions in the debt securities markets. Lenders already bear the risk of economic downturns, when defaults spike. If lenders are required to double their bet (by taking on the downturn risk currently borne by borrowers), they will cover the cost of this downturn insurance by raising interest rates. The authors confidently assert that lenders can price the risk based on past performance, but this reminds us that economics is the queen of the social sciences in the Ru Paul sense, offering vivid, but distorted, models. Lenders’ ability to predict the length and severity of downturns over the typical terms for student loans (ten years) and mortgages (30 years) is nil: as late as September 2008, the economists’ Consensus Forecast poll predicted that no country would be in recession by 2009. In the absence of the authors’ risk-sharing, borrowers with less ability to ride out a downturn wouldn’t borrow – reducing the default rate and financial instability when the inevitable downturn comes.

Perversely, the authors’ insurance scheme will force the massive use of derivatives not currently demanded by the market. Bond buyers are generally interested in a predictable fixed return. To provide this, the investment banks that securitize student loans and mortgages structure the securities into tranches (segments). Upper tranche buyers typically get relatively low interest rates that are comparable to Treasury or high-quality corporate bonds.  In exchange, they are the last to bear the risk of default on loans backing the bonds. 

Buyers of lower tranches (known as “B-pieces”) get higher interest rates (akin to junk bonds) in exchange for accepting equity-like risk: higher payments in a good economy but bearing the first losses when defaults rise in a downturn.  B-pieces are hard to value.  In past decades, they led to a series of investment bank blowups culminating in the Great Recession’s CDO conflagration. In response to Mian and Sufi’s structure, securitizers will dump the index risk into a speculative B-piece that will leave the incentives of upper tranche buyers untouched. 

Arbitraging indexes

The specific student loan and mortgage proposals have further problems. Mian and Sufi propose reducing student loan indebtedness when unemployment increases. There is little reason to subsidize groups such as recently graduated engineers, who command high starting salaries and have just 5 percent unemployment. There is even less reason to encourage borrowing by students who are unlikely to earn a college premium on graduation. Undergraduates have a six-year bachelors degree graduation rate of just 59 percent, and noncompleting students have a 59 percent incidence of student loan delinquency or default. Even among recent graduates who have completed their bachelors’ degrees, 56 percent are unemployed or work at jobs that do not require a college degree. Graduate student loans have exploded despite tiny wage premiums that make many degrees little more than cash cows for universities. Under the authors’ proposal, students will have even less incentive to limit their debt — and higher education institutions will have even less incentive to limit tuition. 

For home mortgages, Mian and Sufi propose reducing principal and payments when housing price indices drop, claiming that there are housing price indexes accurate down to the zip code level. But houses, even in a single zip code, are of multiple ages, sizes and quality. It took years to develop the data on repeat sales of the same houses that make the Case-Shiller index reliable — and Case-Shiller’s finest grained public indexes go down only to the metropolitan area level. Nor has economist Robert Shiller’s effort to promote housing index futures succeeded so far: as of July 3, the CME Group’s national S&P/Case-Shiller Home Price Index had all of 16 futures contracts outstanding.  

The more complex the reconciliation of the actual market price with the index value, the more litigation: commercial shared appreciation mortgages generated substantial litigation when they were popular in the 1970s and 1980s. Even without litigation over the amount of the writedown on sale, as of June 2013, it took an average 1,033 days to foreclose in New York and New Jersey, 907 days in Florida and 817 days in Illinois. 

With the loan principal reduced in a falling market and borrowers in control of the time of sale, borrowers will leverage up with even less concern for price than during the bubble, then strip the debt when the market falls. (Similar opportunism led to a prohibition on principal-stripping in Chapter 13 bankruptcy.) Nor will lenders be comforted by the authors’ proposal that, if prices go up, they get just 5 percent of the gain on sale, given that prices dropped by about a third nationally from peak to trough in the bust. (The authors offer student loan lenders no upside at all if the unemployment rate drops.)

The Soprano solution

Sufi’s University of Chicago colleague, Ronald Coase, won the 1991 Nobel Prize in Economics for demonstrating the importance of minimizing transaction costs. Coase recently died of natural causes at age 102, but given the authors’ treatment of his life’s work, it’s as if, Tony Soprano-style, they had smothered him with a pillow. Complex and opaque debt structures enable politicians and crony capitalists to disguise high leverage while spinning the predictable blowups as black swan events.

If a consumer debt hangover is hindering the economy, as Mian and Sufi plausibly argue, then the government should be encouraging writedowns in exchange for the elimination of future guarantees and other hidden debt subsidies. Borrowers and lenders, not taxpayers, should bear the risk.