The Panic of 2007 started in shadows

When the Panic of 2007 broke out in August of that year with the collapse of the housing market, I was in a unique position to observe the events. For 25 years, my academic career — in the Federal Reserve System, at the Wharton School of the University of Pennsylvania and at the Yale School of Management — focused on banking, financial crises and banking panics. My 1983 Ph.D. dissertation was on the subject of banking panics. One paper from my thesis was the first (and, as far as I know, the only) econometric study of panics to this day.

When I wrote that thesis, I never dreamed that I would live through a banking panic. The experience of the current one would be surreal if it were not so tragic. Certainly, the events are confusing, and the reality of what happened is hard to accept. But what exactly did happen? How could it happen? Answering these questions is important because the narrative of what happened provides a framework for new regulations, laws and policies, ones that are relevant and effective. The lenders of last resort, the central banks, in the future will need a record of what happened in the Panic of 2007. A. Piatt Andrew argued a century ago about the Panic of 1907: “The unique dimensions of the recent panic among the experiences of the present generation render important the preservation for future study of all records concerning its phenomena.”

SYSTEMIC EVENT

Professional economists tend to focus on the post-World War II era and on the stock market. But the earlier history offers important clues about how to think about what a “systemic event” is and what happened in the current crisis. Indeed, another way to understand what happened is to ask, “Why was it a systemic event?” Like many terms, this one has lost any precise meaning and has come to signify “bad financial events.” It is important to recover some precision about this because its meaning is so closely linked to lender-of-last-resort policies, that is, what the central bank does in a financial crisis.

The modern financial system is complex, but still it is surprising that it has been so difficult to figure out what happened. One reason may well be that the events themselves were largely invisible to all but the participants in certain financial markets. In the Panic of 2007, most people had never heard of the markets that were involved, didn’t know how they worked or what their purposes were. Such terms as subprime mortgage, asset-backed-commercial paper conduit, structured investment vehicle, credit derivative, securitization or repo market were meaningless. These markets were obscure and esoteric for most, including economists. In the earlier panic episodes, not only could everyone see the runs on banks, most people likely participated, rushing to their bank to withdraw their money out of fear the bank would not survive the coming recession.

PAST PANICS

These runs would occur at all banks, usually starting in New York and spread from there. Everyone knew that the panic had happened and consequences would follow — firms would fail, and there would be difficulties making transactions.

The visibility of earlier panics did not make the events themselves explicable, but it’s clear to everyone what happened. In the Panic of 2007, the “bank run” was invisible to almost everyone because it was a run by banks and firms on other banks. These interbank markets were invisible to the public, journalists and politicians. Without observing the bank run, what became visible were only the effects of the run, and in many cases, the effects were mistaken for the cause. Without the details of what happened, new policies may end up addressing effects rather than causes.

If we think about a 19th-century bank run, like the one on the Seaman’s Savings Bank in 1857, we can get a sense of the problem. When everyone demands to withdraw cash from their banks, it is not possible for the banking system to meet these demands. The money has been lent out, and banks do not hold significant amounts of cash (because it does not earn a return). The banking system becomes insolvent because it cannot meet the contractual demands of the depositors; that is, banks are simply unable to pay back all the cash that depositors want. Because the banks have lent the money out, there is no easy way to get it back. The banks cannot sell their loans — the assets of the banking system are simply too large for anyone to buy. This is what makes a banking panic a systemic event. One bank could possibly sell its loans and pay off its depositors. But when all banks have to sell loans, there are no other banks to buy them.

The events of 2007 are essentially a repeat of a 19th-century bank run — only in 2007, some firms ran on other firms. What has become known as the “shadow banking system” is, in fact, genuine banking and it turns out, vulnerable to the same kind of bank runs the U.S. has seen in the past. Where do firms and institutional investors save their money when they do not want to make long-term investments? In other words, what is the equivalent of a checking account for firms? There are no insured-deposit accounts large enough for these depositors. But they have large amounts of money that they would like to deposit safely while having easy access to it, like a checking account.

Over the last 25 years, a number of forces led to a banking solution. The solution is banking, but it does not happen in the familiar form of a depository institution. Firms “deposit” in the sale and repurchase (“repo”) market, a short-term institional market.

REPOS EXPLAINED

Here’s how it works: Imagine that a large institutional investor wants to save $500 million short-term. The investor wants to earn some interest, wants the money to be safe (no risk) and wants to have easy access to the money. One thing this investor could do is buy U.S. Treasury bonds. But there are many demands for Treasurys. Not only do foreign governments and foreign investors want to invest in Treasurys, but there are many domestic demands for them, as well. The demands for this type of (information-insensitive) bond are enormous. Treasurys are used as collateral for derivatives positions and in clearing systems. There is a shortage of such collateral. So our institutional investor may well engage in the following transaction: the $500 million is “deposited” overnight with a bank (investment bank or commercial bank, foreign or domestic). The institutional investor will receive bonds (not necessarily government bonds) with a market value of $500 million; in other words, he receives collateral. In the panic, the collateral most likely will be securitization-related bonds, which represent claims on the portfolios of loans held by special legal entities that hold only that portfolio. The institutional investor will earn interest on the deposit. The bonds have to be given back when the institutional investor withdraws his money by not renewing (not “rolling”) the transaction. Note that the firm receiving the deposit of $500 million has just financed the bonds that were given as collateral.

This transaction has some notable features. It resembles checking in that it is short-term. It often lasts overnight, is backed by the collateral, and the bond received as collateral can be “spent,” that is, it can be used as collateral in some other transaction that the institutional investor may undertake. And that party can pass it on. This process of reusing the collateral repeatedly is called “rehypothecation.” In short, repo is banking. You can see why the Federal Reserve system counted these transactions as “money” when it computed a measure of money called M3, now discontinued.

The problem that will arise stems in part from the demands for collateral and how the private sector responds to this, by producing and supplying collateral. Simply put, there is a shortage of information-insensitive collateral that can be used in repo. For various reasons, the financing of bank loans began to move out of the regulated bank sector and into capital markets. Many important forces have led to the evolution of the banking system, but in this case, the private sector began to produce bonds that could be used as collateral in repo.

RIPE FOR FAILURE

You can see the possibility of a panic; it could occur if the depositors in the repo market decide not to renew their deposits and withdraw instead. Once a panic occurs, things get complicated fast. Transactions (or “liquidity”) are best accomplished with information-insensitive securities, such as demand deposits or repo with collateral.

These markets are defined by the fact that individuals do not perform due diligence on the credit risk, precisely because they have confidence in the value of the securities, and because they are sure the other side does not know more than they do about the security’s value. This is called “confidence in the system.” No one needs to know the details of the securities, precisely because they don’t matter.

It is a bit like electricity. When you wake up in the morning, you put your lights on. When you leave for work, you turn them off. Return from work, turn them on; go to sleep, turn them off. You don’t need to know anything about electricity for this system to work. In fact, the idea is that you shouldn’t have to know; you don’t need to be an electrician. But if it happens that there is a blackout in which the whole electrical grid breaks down (this came close to happening in August 2003), then there is a problem. No one saw what happened to the grid, and many people do not actually know what electricity is. For the average person, the concept of electricity is incredibly complicated.

Of course, the solution is not for everyone to become an electrician: rather, it is to restore the credibility of the system so that no one has to think about electricity.

Gary B. Gorton is the Frederick Class of 1954 Professor of Management and Finance at the Yale School of Management and research associate at the National Bureau of Economic Research.

For archived columns, go to InvestmentNews.com/advisersbookshelf.

Related Topics: Adviser's Bookshelf