Fault Lines: How Hidden Fractures Still Threaten the World Economy

by Raghuram G. Rajan

  Public exposure can reduce tail risk taking in its early stages, for tail risk is of significant value to management at that stage only if the public is unaware of the extent of the company’s exposure to risk. Bond holders will typically not be happy to learn that the hundreds of millions in dollars of profits made in the past quarter came from taking on trillions of dollars’ worth of exposure to particular rare events.

  The Repricing of Financial Claims and Incentives

  Public exposure can work by making the financial claimants on a financial firm check excess risk taking. Of course, equity holders, who get much of the benefit, may actually favor tail risk taking if the financial firm’s equity cushion is thin relative to the size of the firm: they have little to lose. So a larger capital cushion is necessary to deter risk taking. Below, I suggest ways that such capital can be raised. Also, as I argue above, many debt holders—not just insured depositors but even holders of long-term unsecured debt—thought they would be bailed out by the government, so the interest rate they demanded did not adjust upward as the financial firm took more risk. But debt holders could be an effective constraint on risky behavior if they bore the downside risk, for their future anticipated losses would be reflected in the higher interest rate they demanded from the financial firm. If interest rates move up substantially on a large portion of the financial firm’s debt when it takes more risk, thereby reducing profits, both management and the board may be deterred from risky behavior. At the very least, such an increase in rates paid will be a strong signal to the public and to regulators that the financial firm is taking substantially more risk. Therefore, regulatory reform should focus on ensuring that an important segment of a financial firm’s debt holders know they should expect painful losses if the financial firm takes too much risk.

  But neither equity holders nor debt holders will worry much once risk taking becomes so systemic that the market comes to rely on government intervention to prop up markets when the risk hits. In that situation, management and stockholders or bondholders unite to see risk taking as a value-maximizing activity. And if the firm is deemed too systemically important to be allowed by the authorities to fail, the financial firm’s investors are unlikely to ever bear the full cost of big losses. This is a situation tailor-made for encouraging tail risk taking, because the firm makes a lot of profits in good times—everyone purchases tail risk insurance from firms like AIG that will be propped up by the government—and AIG runs to the government for a bailout if the costly tail risk ever hits. Let us take these situations in turn.

  Government Intervention in Markets

  With the government so heavily involved in mortgage lending, both directly through the FHA and Ginnie Mae and indirectly through Fannie Mae and Freddie Mac, there was little chance that the market for housing finance, especially subprime housing finance, would be left unsupported even if a big shock hit. Similarly, the Fed and Treasury have supported virtually all the big banks that chose to take liquidity risk.

  Clearly, the way to reduce tail risk taking under these circumstances is to back off from government intervention, to the extent possible. There is no inherent reason why the government should have such a large presence in the market for housing finance, other than the fact that the middle class has grown used to implicit housing subsidies. Government conservatorship of Fannie and Freddie gives it an opportunity to create true private-sector housing finance by breaking up these monoliths and creating a handful of private entities, none of which have an implicit government guarantee. At the same time, both the FHA and Ginnie Mae should be shrunk steadily. Although some think the fickleness of the private sector—as evidenced by the current drying up of private subprime housing finance—is a reason for the government to maintain its presence, it is the unsustainable levels to which house prices have been pushed, in part because of that very government presence, that has caused private finance to dry up.

  Similarly, there really is no reason other than political pressure for the Fed to take us from bubble to bubble by cutting interest rates to near zero and flooding the market with liquidity. Ironically, the lesson from the Great Depression—that letting the banks go under is not a good idea—has been so well absorbed by the Fed that it is played for a patsy by the banks. One reform I discuss in chapter 9 is to create a stronger social safety net. That would take some pressure off the Fed to keep injecting stimulus so long as jobs are not growing.

  In addition, the exercise of monetary policy should be rethought to take into account its effects on risk taking by the private sector. Financial stability should become a more explicit part of the Fed’s mandate, given as much weight as employment and low inflation. The Fed ought to ask itself whether it is possible that nominal interest rates could be too low, even when the economy is in trouble. The nineteenth-century editor of the Economist, Walter Bagehot, was fond of saying, “John Bull can stand many things but he can’t stand 2 percent.” Similarly, John Doe cannot stand interest rates near zero, and when the Fed pushes short rates very low, especially when deflation is not a clear and present danger but just a possibility, savers move to holding riskier assets, pushing all manner of risk premiums down and prices up. A rock-bottom nominal short-term interest rate prompts risk taking and makes price bubbles more likely; it is unclear, however, that it is much more helpful in prompting corporate capital investment and job growth than a somewhat higher but still low nominal short-term interest rate.

  Moreover, if the Fed on occasion has to cut interest rates sharply, it should be prepared to raise interest rates higher than strictly warranted by economic activity once the emergency is over. Such interest-rate policies will reward those who maintain liquid balance sheets and prevent others from becoming overly illiquid in their activities.10 It will also offset the one-sided discriminatory intervention by the Fed that favors debtors at the expense of savers.

  Neither set of recommendations—disengaging from housing or following a more evenhanded monetary policy—will be easy to adopt. Powerful groups, including real estate developers and financiers, have an interest in seeing continued government involvement in real estate. Should we wait for the next deep crisis for the government to distance itself? Similarly, monetary economists and central bankers (often cut from the same cloth) will think it crazy that interest rates should be related to anything other than real activity—a point made abundantly clear by Ben Bernanke in his speech to the American Economic Association in Atlanta on January 3, 2010. As we have seen though, the financial sector can affect real activity with a vengeance. We have made the mistake of not paying attention to risk taking before; we should not make it again.

  Eliminating “Too Systemic to Fail”

  One reason why markets did not penalize financial firms that took on large amounts of tail risk was that they thought the firms were too systemically important to be allowed to fail by the government. Here again the government has substantiated these beliefs by bailing out nearly every large or interconnected bank, including, most controversially, large banks that had bought insurance from AIG. The government simply cannot afford to be seen as a soft touch by financial firms or the market. Hence one of the most important items on the reform agenda is to ensure that no private financial firm is deemed to have the protection of the government.

  Entities that are widely known to be too systemic to fail not only have warped incentives, but they also have a competitive advantage over entities that do not enjoy such implicit protection: they can take on costly tail risks secure in the knowledge that they can appropriate the resulting revenues in good times while passing on the risk to the government in bad times. Equally problematic, market investors, also knowing that they will be kept whole by the government, have no incentive to discourage them. Indeed, because the strategy offers tremendous potential profits and limited losses (to the equity holders, that is), equity investors will applaud it: the most risk-loving banks had the highest stock prices before the onset of the crisis. Because bond holders will not
demand a higher premium for bearing the risk of losses, banks can double up their bets with leverage. Finally, because these banks enjoy a lower cost of financing, they can grow even larger and more complex; and they have an incentive to do so to make themselves even more systemic.

  When a downturn hits, the problems associated with entities deemed too systemic to fail multiply. Resources are trapped in corporate structures that have repeatedly proved their incompetence, and further resources are sucked from the taxpayer as these institutions destroy value. Confident in the knowledge that the government will come to their rescue, these institutions can play a game of chicken with the authorities by refusing to take adequate precautions against failure, such as raising equity.

  Perhaps just as important are the political consequences of such rescues. It is hard for the authorities to refute allegations of crony capitalism. Aside from the stated intent of saving the economy, there is no discernible difference between a bailout motivated by the sense that institutions are systemically important and one motivated by the desire of those in authority to rescue their friends or their once and future employers. Even as conspiracy theorists have a field day, painting everyone remotely associated with the financial system into a web of corruption, the damage to the public’s faith in the system of private enterprise is enormous: it senses two sets of rules, one for the systemically important and another for the rest of us. And the conspiracy theorists do have a point: the leeway afforded to the authorities in choosing who is too systemic to fail allows tremendous scope for discretion, and hence corruption.

  I have avoided referring to institutions as too big to fail. This is because there are entities that are very large but have transparent, simple structures that allow them to be closed down easily—for example, a firm running a family of regulated mutual funds. By contrast, some relatively small entities—examples include the monoline bond insurers who guaranteed municipal bonds, and Bear Stearns—caused substantial stress to build up through the system. A number of factors other than size may cause an institution to be systemically important, including its centrality to a market, the extent to which other systemically important institutions are exposed to it, the extent to which inflicting losses on the liability holders will also inflict disproportionate losses on the assets, and the complexity of the institution’s interactions with the financial system, which may render the authorities uncertain about the systemic consequences of its failure and reluctant to take the risk of finding out.

  There are three main ways of dealing with these problems. First, try to prevent institutions from becoming systemically important. If they do become important, force them to have additional private-sector buffers to minimize the need for the government intervention. If, despite these buffers, they do become truly distressed, make it easier for the authorities to allow them to fail. I discuss each of these measures in turn.

  Keeping Institutions from Becoming Systemically Important

  A number of suggestions are circulating on how to keep institutions from becoming systemically important. The most common is to stop them from expanding beyond a certain size, an idea most closely associated with the former Fed chairman, Paul Volcker. While prima facie attractive, this proposal has weaknesses.

  Although large institutions are more likely to be systemic, size is neither a necessary nor a sufficient condition. Bear Stearns was not considered large by most calculations, though it was considered connected enough to need saving. On the other hand, the mutual-fund group Vanguard manages more than a trillion dollars in assets but would probably not qualify as systemic. Not all aspects of size are equally troubling.

  Moreover, even if we could be specific about the aspects of size that are troubling, banks would try to evade any restrictions on size by economizing on whatever measure served as a criterion, whether assets, capital, or profits. Crude size limits would likely lead banks to conceal a lot of financial activity from the regulator, only to have it come back to light (and to balance sheets) at the worst of times. There are many legal ways to mask asset size. Instead of holding assets on their balance sheet, banks can offer guarantees to assets placed in off-balance sheet vehicles, much like the conduits of the recent crisis. If, instead, capital were the measure, then we would be pushing banks to economize on it as much as possible; this is hardly a recipe for safety. And if it were profits, we would be inviting healthy banks to park profits elsewhere while rewarding sickly ones by allowing them to expand indefinitely.

  Also, being big has some virtues. Larger banks may be better at diversifying and attracting managerial talent (including risk managers). Although a poorly managed $2 trillion bank creates immense problems for the system, the problems could be even greater with one hundred banks of $20 billion in size, each of which has taken similar risks. What is important is not size per se but the concentration and correlation of risk in the system, as well as the extent of exposure relative to capital. Indeed, in the past regulators have intervened to bail out a system because entities were too numerous to fail—the forbearance displayed to the U.S. savings and loan industry in the early 1980s being one example.

  Instead of imposing a blanket size limit on institutions, regulators should use more subtle mechanisms, such as prohibiting mergers of large banks or encouraging the breakup of large banks that seem to have a propensity for getting into trouble. Entities such as the Federal Trade Commission already have such authority. Although there are always concerns about whether regulators will use these sorts of powers arbitrarily, they are no more difficult for legislators and courts to oversee than are powers based on anticompetitive considerations.

  Turning to proposals to limit activities by insured banks through some modern version of the Glass-Steagall Act of 1933, these again seem less attractive on further reflection. Obviously, some activities of a large bank add considerably to risk and opacity. To the extent possible, these should be clearly separated in legal entities that are not affected by the bank’s failure. For instance, banks should not attempt to use client assets that are pledged to them in their prime brokerage units (units that lend securities, offer loans, and undertake asset-management functions for clients, typically hedge funds) in further transactions.11 The commingling of client assets with the bank’s own funding activities reduces transparency, increases risk, and was an important reason why many investment banks experienced runs in the current crisis, as clients tried to withdraw their assets before they got entangled in the bank’s bankruptcy. Of course, such a separation would increase a bank’s cost of borrowing, but the benefits here might outweigh the costs.

  Proprietary trading—in which the bank uses its balance sheet, partly funded by government-insured deposits, to take speculative positions—is another activity that has come in for censure. The reason critics want to ban it is, in my view, wrong, but there is another reason to consider limiting it. Critics argue that proprietary trading is risky. It is hard to see this as an important cause of the crisis: banks did not get into trouble because of large losses made on trading positions. They failed because they held mortgage-backed securities to maturity, not because they traded them. Regional banks have failed by the dozen because of loans they made for commercial real estate, an activity that no one is talking of prohibiting. It is a fallacy to think that just because certain activities are prima facie riskier than others, keeping banks from those activities will make banks safer overall. In truth, if banks want to take risk, they can simply go further down the spectrum of risk in any of the activities permitted to them. For example, so long as they can lend, they can freely make unsecured, long-term, “covenant-lite” loans to heavily indebted firms. The focus should be on limiting their overall incentive to take risks and their propensity to join with other banks to take risks as a herd, rather than to ban a specific activity, unless the activity has no redeeming financial purpose.

  Nevertheless, there is another reason for considering ways to limit proprietary trading. Banks that are involved in many businesses o
btain an enormous amount of private information from them. This information should be used to help clients, not to trade against them.12 Banks effectively have an unfair informational advantage over the rest of us in trading: they can use inside information, despite the presence of firewalls within a bank that are meant to prevent sensitive client or market information from being shared with traders. This advantage should be reduced by limiting proprietary trading. I say limiting because some legitimate activities, including hedging and market making, could be hard to distinguish from proprietary trading. A crude overall limit on a bank’s trading for its own account, no matter what the purpose, is one possibility, but it suffers from the same problems as any crude limit has. Perhaps an initial crude limit, refined over time with experience, as was the case with capital requirements, may be the way to go.

  The best way to keep institutions from becoming systemically important might not be through crude prohibitions on size or activity but through the collecting and monitoring by regulators of information about interinstitution exposures as well as risk concentrations in the system. The regulator could ensure, through command and control, that the system is not overexposed to any single source of risk, institution, or class of institutions. Regulators themselves would need to be monitored; hence, as I suggested earlier, information on exposures should be released periodically and publicly, after the passage of an appropriate amount of time. Such measures would be less dramatic and punitive than size or activity limits but more easily enforced and probably more effective.

  Building Better Buffers