Shadow Banking

• “Bankruptcy’s Financial Crisis Accelerator: The Derivatives Players’ Priorities in Chapter 11,” by Mark Roe. March 2010. SSRN #1567075.
• “Regulating the Shadow Banking System,” by Gary Gorton and Andre Metrick. September 2010. SSRN #1676947.

In last summer’s “Working Papers,” I discussed several papers that examine the shadow banking system. To briefly recap, in the shadow banking system, excess corporate and investor cash is “deposited” in (what used be called) investment banks. Because those banks were outside the regulated deposit-insurance commercial banking system, the investment banks supplied collateral to the depositors to “guarantee” the deposits. If the investment bank failed to return the cash at the end of the specified period of time — often just one day — the depositor took possession of the collateral, which had a market value equal to or more than the cash plus the owed interest.

Originally, collateral in the shadow banking system was limited to Treasury securities, Fannie Mae and Freddie Mac debt, and federally insured certificates of deposit. But demand for collateral was exceeding the supply and securitized loans filled the gap. In September 2008, investors lost faith in the quality of the securitized loans, withdrew their money from the shadow banking system, and instead invested in short-term government treasuries. This withdrawal of funds was the 21st century equivalent of a bank run from the Depression era, except that the scared parties were large investors rather than the ordinary individual depositors of the Depression.

A new paper by Gary Gorton and Andrew Metrick recounts this story in more detail and describes their proposal to bring shadow banking back within the regulated system. They argue that the Dodd-Frank financial reform bill that became law earlier this year did little to reform the shadow banking system. They propose to create a new chartered entity that they refer to as a Narrow Funding Bank whose sole purpose would be to purchase securitized loans and issue short- and long-term debt to fund itself.

Gorton agrees with Mark Roe (whose work was among the papers I reviewed in the summer issue) that the spectacular rise in the use of short-term repurchase (“repo”) agreements collateralized by securitized loans was the result of special advantages given to such agreements in bankruptcy reforms in 1978 and 2005. Those advantages put repurchase agreements ahead of all other liabilities in bankruptcy proceedings. Thus, “deposits” in the shadow banking system were bankruptcy-remote, which lowered risk for investors and lowered the cost of capital. The bankruptcy of an investment bank would not tie up investors’ cash because they would take possession of the collateral (rather than give it back to the bank to be part of the pool of assets divided among all investors through bankruptcy) and liquidate it outside of the bankruptcy process.

The downside of the bankruptcy provisions, however, is that if investors ever lost confidence in the collateral used to “guarantee” their deposits, they would flee the shadow banking system just like ordinary retail depositors did during the Depression before deposit insurance. And that is exactly what happened in the last quarter of 2008. The increased default rates in subprime mortgages caused investors to lose confidence not only in repurchase agreements that were collateralized with subprime mortgage securities, but all repurchase agreements that had any type of securitized loans as collateral.

In response, Roe argues that all creditors should be treated identically in bankruptcy. He thus proposes eliminating the special bankruptcy provisions for repurchase agreements. This would increase market monitoring by the suppliers of deposits on investment and decrease the use of short-term funds to back longer-term investment. Gorton, on the other hand, suggests eliminating the bankruptcy provisions only for repurchase agreements that take place outside the new Narrow Funding Banks.

What makes these new regulated narrow banks any safer than the shadow banking system they replace? Gorton says that bank runs can be prevented only through deposit insurance or “safe collateral.” He proposes that regulations and regulators will define safe collateral and adequate capital requirements for this new class of narrow securitized loan banks:

If we fear that regulators are not up to the task, then we must pay them more and pay them better. We do not see any pure private-sector solutions to ensure the safety of the banking system, so the role of regulators will remain essential. To the extent that this role is found to be impossible, then we are either destined to have more crises or forced to live with a greatly constrained financial system.

But until securitized housing loans proved more risky than expected, everyone thought that such loans were good risk-free collateral, so Gorton does not appear to have found the magic elixir of financial security. Instead, he reminds us that taking risk out of investing is very hard to do.

But assume that the regulators can implement Gorton’s safe-collateral regime. The “safety” provided will need to be inexpensive because, otherwise, the incentives to experiment and leave the protected system would recreate the shadow banking problem again, especially as memories from the 2008 crisis fade.

Climate Change Economics

• “Fat Tails, Thin Tails, and Climate Change Policy,” by Robert S. Pindyck. September 2010. NBER #16353.

The most interesting intellectual debate in the economics of climate change is taking place between William Nordhaus (Yale), Martin Weitzman (Harvard), and Robert Pindyck (MIT). It involves technical discussions about the characteristics of the probability distributions of future temperature outcomes and the damages caused by such temperatures. This arcane discussion allegedly provides a scientific answer to the question of how much we should be willing to spend now to avoid catastrophic damages in the future.

The normal method of climate policy uncertainty evaluation (exemplified by Nordhaus’s work) is Monte Carlo simulation (repeated simulation of a model using random draws from the distribution of possible variable values) of integrated economic and climate models. In such simulations, the distribution of damages from global warming are assumed to be thin-tailed — that is, the costs of damages from extremely low-probability high-temperature warming declines to zero faster than exponentially. Under this assumption, the marginal benefit of avoiding very high future temperatures (which produce large damages and reduce future consumption by large amounts) is bounded, finite, and small.

In a series of papers, Weitzman asks what if the distribution of damages is fat-tailed? That is, what if the damages from high global warming temperatures decline to zero more slowly than exponentially? Then, the expected marginal benefits of any incremental reduction in the probability of high-temperature events would be infinite. The conclusion that follows is that we should devote all our income to preventing climate change and its possible catastrophic results.

Now this is an absurd result, which Weitzman concedes. But he proposes more modestly that standard expected value cost-benefit analysis is misleading and undervalues the probability of, and benefits from, preventing extreme outcomes.

Pindyck argues that Weitzman’s proposal still leaves us with a problem. Suppose we could pay 10 percent of current income for an insurance policy against extreme climate change outcomes. Weitzman implies that we should buy it because it certainly is less than the 100 percent of income demanded by his mathematical analysis. But that is true only if there are no other competing catastrophes. But nuclear war, a viral pandemic, and other equivalent events compete for our preventing-catastrophes budget. And if we are willing to pay 10 percent of our income to prevent each of them and there are ten such disasters, then we have once again reached the absurd result of using all our income to prevent catastrophes. So for Pindyck, we have come full-circle back to expected value cost-benefit analysis.

Financial Executive Compensation

• “Bank CEO Incentives and the Credit Crisis,” by Rüdiger Fahlenbrach and Rene M. Stulz. August 2010. SSRN #1439859.
• “Executive Compensation and Corporate Governance in Financial Firms: The Case for Convertible Equity-Based Pay,” by Jeffrey N. Gordon. July 2010. SSRN #1633906.
• “Pay for Banker Performance: Structuring Executive Compensation for Risk Regulation,” by Frederick Tung. July 2010. SSRN #1647025.

What role did the incentives provided by the compensation of financial executives play in the financial crisis? Rene Stulz has demonstrated that the more equity in an executive’s pay, the worse the subsequent performance of the financial institution. The larger the stock exposure of a chief executive officer in 2006, the worse the subsequent performance of the institution in 2007 and 2008. This evidence certainly contradicts the commonly held view that bank CEOs led us off a cliff to enrich themselves and their shareholders. But the evidence is consistent with the view that the more incentivized a CEO was to take shareholder interests into account, the worse the results for those shareholders, which would seem to contradict much modern compensation theory.

Recent papers by Jeffrey Gordon of Columbia Law School and Frederick Tung of Boston University Law School argue that equity compensation is just fine for non-financial firms because such firms do not pose a systemic threat to the rest of the economy. (The political system did not have the courage to test that theory in the case of General Motors.) But it is not efficient for financial institutions whose failure induces investors to lose confidence indiscriminately in other financial institutions.

They argue that the problem arises because executives of corporations are not diversified investors. Instead, their wealth is overwhelmingly invested in the firm they manage. Thus, unlike diversified investors who face systemic risk, executives do not really face correct incentives if their decisions have large spillover effects on all other assets in the economy, because they do not own any of those assets. Such executives win greatly if their bets are profitable but they do not suffer as much as diversified shareholders if their bets fail and have systemic consequences for the entire economy.

A second and related incentive problem arises when new capital must be raised to allow a financial firm to survive the reduction in asset values that follows a bad investment strategy such as subprime real estate loans. Non-diversified financial executives face too little incentive to secure new equity for their firms during financial stress because the infusion of new equity would disproportionately reduce their wealth relative to the wealth of an outside diversified shareholder.

Gordon proposes that the equity portion of executive pay be converted to subordinated debt automatically (with a haircut — a specified percentage penalty) if certain contractually specified events occur (e.g., debt ratings downgrades or stock price declines). He argues this would end excessive risk-taking at the very time when an institution needs less risk-taking and survival is important for the economy. Tung proposes that subordinated debt be a continuous component of financial executive pay.

The choice between the two proposals is directly analogous to the discussion in the previous section of equity versus catastrophe bonds, only in reverse — when to switch from equity to debt rather than debt to equity. That is, should the incentives arising from debt ownership be continuous or occur only during times of financial stress? If we can discover optimal switching rules, than switching would seem better than continuous debt.

Bank Capital Requirements

• “Fallacies, Irrelevant Facts, and Myths in the Discussion of Capital Regulation: Why Bank Equity Is Not Expensive,” by Anat Admati, Peter DeMarzo, Martin Hellwig, and Paul Pfleiderer. September 2010. SSRN #1669704.

Following financial crises, one of the recommended policy responses is to promote banks’ reliance on equity capital, instead of debt, so that shareholders absorb losses when future shocks occur. But financial institutions are often reluctant to increase their equity capital. They argue that it acts like a tax on lending and thus will reduce it. They also argue that it sends the market a negative signal about banks’ safety and soundness and thus may catalyze a negative confidence spiral, the exact opposite of the intended result. And bankers often argue that banks should be highly leveraged because equity capital is “expensive.”

Those and other components of the conventional wisdom about the appropriate role of bank equity capital are strongly critiqued in a recent paper by Anat Admati of the Stanford Business School and her colleagues. Their basic argument is that more equity is a cure for every problem in financial institutions.

They argue that banks have so much debt and so little equity because of the tax deductibility of debt interest (and the non-deductibility of dividend payouts to equity holders) in the corporate tax code rather than any inherent feature of lending. In their view, capital requirements are simply an administrative attempt to offset the bad capital-structure incentives created by the debt tax subsidy.

Many believe that more capital reduces returns in banking. While that is true during good times, the extra capital will allow the bank to survive bad times and thus the risk-adjusted return may actually increase with more capital.

The conventional wisdom in financial economics is that more equity rather than debt in financial institutions’ capital structures leads to excessive risk-taking by bank managers. In the conventional view, the predominant role of short-term debt in financial institutions (the short-term repurchase agreements at the heart of the shadow banking system, for example) is to constrain the investment practices of bank managers — if they become reckless, investors will go elsewhere. But the manner is which the discipline of debt manifests itself is a bank run, which creates such large collateral damage that political systems do not let runs occur, thus undermining the disciplinary effects of short-term debt. In fact, Gorton argues that the role of AAA-rated collateral in short-term repurchase agreements was to be informationally insensitive — that is, investment that no one had to think about. And something no one thinks about cannot discipline the market.

The currently faddish cure for the troubles created by short-term debt in the capital structure of financial institutions is catastrophe bonds — a hybrid combining characteristics of debt and equity. Catastrophe bonds pay interest like conventional bonds during normal times, but convert to equity during contractually specified stressful times like those experienced during the fall of 2008. Admati and her coauthors argue that catastrophe bonds are not a unique cure and are favored by most analysts because of the tax-deductible interest. Rather than catastrophe bonds, the authors argue that a new kind of equity with requirements on periodic payouts (dividends are mandatory rather than discretionary) to prevent managerial shirking should replace debt in the capital structure debt of financial institutions.

While the paper offers an important corrective to the conventional positive appraisal of debt and negative views toward equity in financial institutions, it does not address an important benefit of catastrophe bonds: their countercyclical nature. During good times, financial institutions have too much capital and during bad times they do not have enough. The paper argues correctly against the conventional view that additional equity is a drag on returns because such a view applies only to the good times when the cushion is not necessary but ignores the stressful times when the cushion allows survival. But the logical extension of this insight is that their prescription of more equity capital all the time would not be an efficient solution to this problem. Instead, catastrophe bonds that convert to equity exactly when you need it would seem to be preferable.