Harvard Business Review
Friday June 8, 2012
by Edward Conard
Some politicians claim that unsystematic loan losses, like the $2 billion loss JP Morgan recently suffered, significantly increase the risk of another financial crisis. They do not. A 30% drop in real estate prices, which systematically threatened all lenders with upwards of a trillion dollars of losses, caused the financial crisis. Subprime mortgage losses turned out to be much smaller than expected —$300 billion, according to the Federal Crisis Inquiry Commission—and non-bank lenders suffered most of those losses (notwithstanding mark-to-market losses from credit downgrades). Nevertheless, institutional investors withdrew $1.5 trillion of short-term funding from intermediaries—five times more than loan losses—despite $15 trillion of explicit government guarantees made during the crisis. Had the government guarantees been smaller, the withdrawals would have been much larger. Two billion dollars of losses may be eye opening, even for JP Morgan with equity worth $170 billion, but it’s orders of magnitude smaller than risks large enough to spark a financial crisis.
The financial crisis stems from the fact that the U.S. economy finds itself on the horns of a dilemma. Price-insensitive, risk-averse savers hoard rather than invest their savings and insist on the right to withdraw their funds “on-demand.” If these savings sit idle, growth slows and unemployment rises. Banks are the vehicle through which the economy puts short-term savings to work. Because it takes time to repay loans, the economy runs the risk that short-term savers may panic and demand withdrawals en masse—when real estate prices fall 30%, for example. When that happens, lenders must sell assets—namely loans—to fund withdrawals. In a panic, when sellers vastly outnumber buyers, asset values sink to fire sale prices. At low enough prices, banks cannot sell enough assets to fund withdrawals and still remain solvent. Depositors race to withdraw funds before insolvency, which amplifies withdrawals.
Banks, therefore, face two types of risk: loan losses (or credit default risk), which are relatively limited, and withdrawal risk (or liquidity risk), which is nearly unlimited. Overcollateralization protects lenders from loan losses—traditional 20% homeowner down payments protect lenders from losses in the case of conventional mortgages; comparably sized subordinate tranches of mortgage securitizations protect AAA-rated debt in the case of subprime no-money-down mortgages. Real estate prices, however, fell 30% during the crisis. We can demand that banks hold loans that can withstand a 30% drop in real estate prices and still maintain their AAA-ratings, but this requires banks to make loans with no less than 50% overcollateralization.
To dampen the amplified risk of withdrawals without the need for excessive collateral, the government makes explicit guarantees of liquidity in the case of retail deposits and implicit guarantees in the case of institutional deposits—implicit guarantees that were made explicit in the financial crisis. But the world woke up in 2009 and recognized that implicit government guarantees of short-term institutional deposits did not and, now, likely will not hold deposits in place as it assumed they had since 1929. Borrowers and lenders have dialed back lending to compensate for the now-recognized risk of damage from withdrawals. Savings sit idle and unused as a result. The economy has contracted, growth has slowed, and unemployment remains high.
Angry voters conflate the risk of loan losses with the risk of withdrawals. They insist lawmakers hold banks responsible for both. They fail to see that institutional depositors can withdraw from a fragmented banking system as easily as from a consolidated one, and that it is near impossible for the private sector to hold enough equity in reserve to make guarantees credible enough to stem withdrawals. The United States has only about $20 trillion of publicly traded equity. And equity is a constrained resource, especially in a recession. Reductions in risk underwriting elsewhere offset gains from diverting equity to underwrite the risk of withdrawals. The government, however, was able to guarantee and fund withdrawals without either idling equity or suffering losses despite a 1929-like bank run. In fact, the government expects to turn a profit.
A better solution than threatening banks and the economy with the risk of damage from withdrawals, and suffering slower growth and higher unemployment as a consequence, is to strengthen the government’s ability to guarantee withdrawals in a panic rather than politicizing its ability to act. Were the government to do this—chiefly, by making explicit rather than implicit guarantees—it could take steps to reduce the heightened risk of moral hazard (where risk takers are encouraged to take excessive risks by capturing gains while avoiding some of the losses). This could be done by increasing capital adequacy requirements to hold banks more accountable for the risk they can manage—default risk—and by restricting guarantees to systematic risks that affect all banks. The government could also charge banks for the guarantees they are, in fact, making. It could improve the accuracy of that pricing by taking steps to maximize visibility into the risks banks are taking and by selling a portion of each bank’s insurance to the public. Entities too illiquid for price discovery could borrow short-term funds from those that are liquid enough.
Instead, the Fed cuts the short-term interest rate to near zero and dares anyone to use hair-triggered, short-term debt to fund longer-term loans. As if nobody learned their lesson! No surprise: that no longer increases growth. Until the government takes steps to manage the risk of damage from withdrawals, the recovery will remain anemic.