A Risky 'Systemic' Watchdog

By Sebastian Mallaby
Monday, March 2, 2009

Barney Frank, the thoughtful chairman of the House Financial Services Committee, wants to create a new "systemic risk regulator." This general concept has been endorsed by some extremely distinguished economists. Nevertheless, the Frank proposal is dubious.

Frank and his allies begin with the accurate insight that existing regulation is inadequate. We have a set of overseers who evaluate financial institutions one by one, but "systemic risk" is created by the interactions between institutions. A bank or hedge fund can take what looks like a reasonable bet on its own terms, but it still may blow up if others are making the same bet.

Consider a bank that takes $1 billion of its capital and $4 billion of debt and buys a $5 billion portfolio of Latin American bonds. It figures that even if the bonds go down, they will take at least a month to fall by a fifth, allowing plenty of time to get out before the $1 billion cushion is vaporized. Based on the history of the bond market, this may be a perfectly smart bet. But if a hundred other banks make the same trade, the calculation is thrown off. If something goes wrong, all these banks will run for the exit at once. The market will crash in a day, not a month. Some will get stuck inside the building.

Moreover, these sorts of "crowded trades" are not merely harder to exit. The odds of needing to exit are higher because the crowding is itself a source of instability. Suppose that one bank loses money on an unrelated oil trade and has to dump its Latin holdings to free up capital. Its selling will drive down the price of Latin bonds, which will prompt some of the other banks in the Latin trade to sell, which may set off a disastrous chain reaction. Leveraged financial institutions -- which include commercial banks, investment banks, insurers and hedge funds -- transmit shocks to one another.

In short, a regulatory system that evaluates financial institutions individually misses the big picture. So why not support proposals for a systemic risk regulator?

The answer starts with the sad truth that crowded trades are difficult for the government to identify. If it somehow mustered the resources to employ a large, talented staff, a U.S. systemic risk regulator might theoretically collect complex risk information from Wall Street, then put it all together in an attempt to identify potential trouble. But it would be missing large pieces of the puzzle because trades can be crowded by foreign players as surely as by U.S. ones. Even if foreign regulators joined in the effort, no international dragnet could be expected to capture trades placed by an obscure Thai trading company, a Russian oil outfit or an Arab sovereign wealth fund. It is hard to identify crowds when half of the people in them are invisible.

Perhaps an imperfect effort to analyze systemic risk would be better than no effort. But that's not so clear when you consider what the new regulator would do with its findings. Having decided, for example, that the Latin bond trade is crowded, the regulator would be expected to tell the big banks to cut their exposure. The banks, which believe the Latin trade is profitable, would object that the government is spoiling the party. If the regulator is issuing its order on the basis of imperfect information -- and if the banks themselves arguably have better information courtesy of their global networks -- it's not clear the government will win this argument. It's not even clear that it ought to.

Finally, what if, by some miracle, the regulator persuaded the banks as a group that they should scale back? How would it then distribute the burden of prudence? Given that the trade is profitable when 10 banks make it but dangerous when a hundred pile in, each bank will strive to keep as large a position in the trade as possible. The more the regulator talked some banks out of the trade, the more it would create an incentive for others to dive in. Herding these financial cats may be beyond the capabilities of even the best government agency.

All of which suggests that a systemic risk regulator would deliver less than expected. There is a real question as to whether the costs of this effort are justified by the benefits. And the potential costs are especially high if, as Frank has suggested, the new systemic risk regulator is lodged within the Federal Reserve. A central bank's credibility is precious and fragile. Saddling the Fed with mission impossible could damage the best institution that we have for fighting financial crises.


© 2009 The Washington Post Company