From Vox:Thomas Philippon15 February 2009
Proposals for financial regulatory reform are everywhere. This column presents an opinionated synthesis of the key issues and proposals with the aim of focusing and stimulating the debate.
The global crisis has scared the public, captivated policy makers, and fascinated the academic community. A consensus has emerged that the financial system is broken and must be fixed. This column puts forth an opinionated overview of the various reports and proposals for new financial regulations in an attempt to stimulate and focus discussion.
I do not describe the crisis itself since much has already been written about it. Brunnermeier (2008) and Hellwig (2008) provide excellent analysis of the early part of the crisis, the Policy Recommendations from NYU Stern (2009) has a complete coverage, and Blanchard (2008) offers a clear and concise macroeconomic perspective.
Three key reports
I will mostly focus on three reports:
· The “Geneva Report” (Brunnermeier, Crocket, Goodhart, Persaud, and Shin;
· The “G30 Report” by the Group of Thirty, chaired by Paul Volcker; and
· The “NYU-Stern Report” by a group of professors from NYU’s Stern School of which I am one.
These three reports cover most, if not all, areas of financial regulation. I also discuss the capital insurance proposals of Kashyap, Rajan and Stein (2008), and the proposals of Zingales (2008, 2009).
SECTION 1: The failures of current regulations
There is much agreement regarding the shortcomings of current regulations. Let me focus on five specific areas.
Systemic risk. All reports agree that the financial regulatory frameworks around the world pay too little attention to “systemic risk”. For instance, Acharya, Pedersen, Philippon and Richardson (2009) argue that “Current financial regulations seek to limit each institution’s risk seen in isolation; they are not sufficiently focused on systemic risk. As a result, while individual risks are properly dealt with in normal times, the system itself remains, or is induced to be, fragile and vulnerable to large macroeconomic shocks.” Or, as the Geneva Report puts it: “regulation implicitly assumes that we can make the system as a whole safe by simply trying to make sure that individual banks are safe. … a fallacy of composition.”
Pro-cyclical risk taking. All reports agree that current financial regulations tend to encourage pro-cyclical risking taking which increases the likelihood of financial crises, and their severity when they occur. Under current regulations, prolonged periods of low volatility reduce statistical measures of risk and thus encourage excessive risk taking. In bad times, the pendulum swings back producing excessive risk aversion.
Large Complex Financial Institutions (LCFIs). All reports agree that current regulations do not deal adequately with LCFIs, defining LCFIs as “financial intermediaries engaged in some combination of commercial banking, investment banking, asset management and insurance, whose failure poses a systemic risk or `externality’ to the financial system as a whole.” (Saunders, Smith and Walter 2009). All reports also insist on the danger induced by implicit Too-Big-To-Fail guarantees.
Capital requirements. All of the reports struggle with a paradox of financial regulation; capital held to meet minimum requirements cannot be used as a buffer against unexpected losses. As such, fixed capital requirements can only ensure that losses do not immediately make banks insolvent. They might give regulators enough time to intervene, but they are ineffective against systemic risk. The real buffer can only come from equity in excess of the requirements.
Liquidity and maturity mismatch. The traditional view of systemic risk focuses on sequences of bank failures, for example a domino-like spread of counterparty failures. Today’s financial system, where many banks finance their investments in credit markets, faces a different type of systemic risk. As the Geneva Report points out, “a key avenue through which systemic risk flows today is via funding liquidity combined with adverse asset price movements due to low market liquidity.” Acharya and Schnabl (2009) also explain why regulators should take liquidity into consideration when assessing capital adequacy ratios, and Hellwig (2008) insists on the maturity mismatch in vehicles that relied on short term market financing to fund long term assets (real estate assets in particular).
SECTION 2: Propositions for regulatory reforms
The various reports agree broadly on the principles that new regulations should follow – they should be based on rules rather than discretion, and they should address well-identified externalities (see Hart and Zingales, 2008). When it comes to concrete proposals, however, there are fewer practical ideas, and less agreement.
My goal here is not to be exhaustive, focusing rather on the issues that I view as most essential, and on the proposals that are sufficiently spelled out to be evaluated.
Systemic risk. The first step towards regulating systemic risk is to measure it. But how should we do that? In particular, how do we define how much a particular firm contributes to systemic risk? How can we improve Value at Risk (VaR) measures?
There is some good news on the VaR front. There are currently two proposals to incorporate systemic risk into the standard measures. The Geneva report argues for CoVaR, based on the work of Adrian and Brunnermeier (2008), where CoVaR is the VaR of financial institutions conditional on other institutions being in distress. The NYU-Stern report proposes a systemic capital requirement based on the individual firm’s contribution to aggregate tail risk. These two measures have much in common, and are specifically tailored to deal with systemic as opposed to individual risks.
My view is that a combination of systemic and liquidity risk measures proposed by the NYU-Stern and Geneva Reports can considerably improve risk management practices inside financial firms, and, just as importantly, create the basis for a constructive dialogue between these firms and their regulator. At this stage, we need more empirical work to show that the proposed measures can indeed be used to identify institutions that pose systemic risk before a crisis hits.
As far as institutions are concerned, all reports also agree that Central Banks should be explicitly in charge of what the NYU-Stern Report calls “systemic regulations” and the Geneva Report calls “macro-prudential regulation”.
LCFIs, Moral Hazard and the Scope of Regulation. A key problem in improving LCFI regulation is that we do not have a good definition of LCFIs. The Geneva report argues that the best measures are leverage, maturity mismatch, and asset growth. These measures are certainly useful, but it is difficult to argue that they differentiate systemic from individual risks: a firm with high leverage and maturity mismatch would already be classified as individually risky. Saunders, Smith and Walter (2009) argue that LCFIs should be identified based on measures of size in combination with measures of complexity or interconnectedness. The difficulty is that we do not have readily available measures of complexity and interconnectedness.
LCFIs are particularly problematic because they are too-big-to-fail. This is in part because we do not have the procedures to deal with their failures. Altman and Philippon (2009) “advocate the creation of specific Bankruptcy procedures to deal with LCFIs.” Zingales (2008, 2009) argues that we need a “new piece of legislation introducing a new form of bankruptcy for banks, where derivative contracts are kept in place and the long-term debt is swapped into equity.” The G30 report argues that “legislation should establish a process for managing the resolution of failed non depository financial intuitions comparable to the process for depository institutions.” Thus, there is broad agreement on what is needed, but, as far as I am aware, there are few concrete proposals about how to deal with the mind-boggling complexity of the issue. Chapter 11 was deemed too risky for the standard (if large-scale) bankruptcy of General Motors. How far are we, then, from a procedure that we could use to deal with the failure of financial Godzillas such as AIG or Citigroup?
Hedge funds and similar. The reports do not present a consensus on the critical issue of what to do with the largely unregulated sector of hedge funds and private equity. Should we impose systemic regulations to a “group of institutions” if they are interconnected and can collectively pose systemic risks even though they appear individually small? My view is that we should, but I would not know how to start.
Capital requirements and cyclical risk taking. The reports suggest that capital adequacy requirements should incorporate liquidity risk, and that they should be tightened in good times – when systemic risk is building up but has not yet been realized – and should be loosened in bad times when banks need a breathing space to weather the crisis.
I see mostly good news on the liquidity front. The G30 Report proposes “norms for maintaining a sizeable diversified mix of long term funding and an available cushion of highly liquid unencumbered assets.” The NYU-Stern report argues that, in order to limit regulatory arbitrage, “regulation should not be narrowly focused on a single ratio of bank balance-sheet,” and should take into account “liquidity to assets ratio” (measured only through stress-time liquidity). The Geneva report has a well-articulated proposal for regulating liquidity and maturity mismatch (their Chapter 5 is a must-read in my opinion).
Unfortunately, I have not seen as much progress on the issue of cyclical risk taking. As Blanchard (2008) explains, “pro-cyclical capital ratios, in which capital ratios increase either in response to activity or to some index of systemic risk, sound like an attractive automatic stabilizer. […] The challenge is clearly in the details of the design, the choice of an index, the degree of pro-cyclicality.” The Geneva report argues that “macro-prudential regulation should be countercyclical and lean against bubbles,” but does not propose an index against which the cycle should be measured.
The Group of Thirty report is even vaguer, since it simply advocates tighter benchmarks when “markets are exuberant and tendencies for underestimating risk are great.” This hardly sounds like the starting point of a useful regulatory debate. The NYU-Stern report argues that stress tests for systemic risk capital can be used to construct a-cyclical risk measures. This is a more precise and practical idea, but it is not clear that this will be enough to create pro-cyclical regulations.
I very much doubt that we can agree on a set of objective measures of ‘excessive’ credit expansion (let alone bubbles). I think that the best we can expect is a powerful regulator running systemic stress tests based partly on historical data and partly on subjective forward looking scenarios. The critical issue in my view does not lay in the construction of an appropriate cyclical index, but rather in making sure that the regulator is powerful enough to enforce tighter prudential regulations based in part on subjective and debatable interpretations of economic data. The financial industry will not like it, and it has a strong track record of capturing its regulators, so this will not be easy.
Recapitalization during crises. This is probably the most controversial and most interesting topic. In times of crisis, asset values decline, and banks tend to curtail lending and liquidate assets in order to control their leverage. These actions increase systemic risk. It would be more efficient to recapitalize the banks automatically at the first sign of crisis.
An interesting idea is to create recapitalization requirements, in addition to capital requirements. One way to do so is to force levered financial institutions to issue securities that provide automatic recapitalization if the firm’s value decreases. Wall (1989) proposed subordinated debentures with an embedded put option. Doherty and Harrington (1997) and Flannery (2005) proposed reverse convertible debentures. These securities limit financial distress costs ex-post without distorting bank managers’ ex-ante incentives.
In a thought-provoking paper, Kashyap, Rajan and Stein (2008) argue that the idea of automatic recapitalization can be applied to systemic risk. They propose a capital insurance scheme based on systemic risk. Each bank would buy capital insurance policies that would pay off when the overall banking sector is in bad shape. The insurer would be a pension fund or a sovereign wealth fund that would essentially provide fully funded `banking-industry catastrophe insurance’. Kashyap, Rajan and Stein explain that “a bank with $500 billion in risk-weighted assets could be given the following option by regulators: it could either accept a capital requirement that is 2% higher, meaning that the bank would have to raise $10 billion in new equity. Or it could acquire an insurance policy that pays off $10 billion upon the occurrence of a systemic event.”
The issue with this proposal is that it does not provide a link between a firm’s own contribution to aggregate losses and the insurance it must get. The policy pays off $10 billion regardless of the health of the bank at that point. The financial institution still has the incentive to lever up, take concentrated bets, and build illiquid positions which may improve the risk/return profile of the firm but nevertheless increase the systemic risk in the system. The crisis has shown that this is a first order concern. Another limitation of this sort of proposal is that if the crisis is large enough, no amount of private money will ever be enough, and the Fed is always going to be the lender of last resort. The mere existence of a LOLR creates moral hazard unless LOLR services are properly priced ex-ante.
The NYU-Stern report proposes solutions to these problems (Acharya, Pedersen, Philippon and Richardson 2009). One is to make the size of the required insurance policy proportional to the estimated systemic risk capital charge defined above in the section on systemic risk. Another is to specify that the insurance must cover the short fall from a pre-specified target: the bank would have to buy an insurance contract such that its equity is at least $50 billion upon the occurrence of a systemic event. If the bank had only $30 billion, the policy would pay off $20 billion, but if the bank had $55 billion, the policy would pay nothing. The bank would have an incentive to limit its systemic exposure in order to decrease its insurance premium.
The Geneva Report is sceptical: “We doubt whether additional private insurance can then help much on occasions when market and funding liquidity vanishes; the examples of the mono-lines and of AIG confirm our doubts.” The NYU-Stern Report argues that the scheme could be implemented with a mixture of private capital (if only for price discovery) and public capital.
Another important caveat is that capital insurance dominates capital requirements only to the extent that it is expensive to keep equity on banks’ balance. This is indeed the core motivation of Kashyap, Rajan and Stein (2008). But one can take the opposite view, in which case higher equity ratios would be a much simpler solution. Hellwig (2008) puts it eloquently: “At this point, the institutions concerned will protest that equity capital is expensive. I have yet to see a convincing argument showing that this protest is referring to social costs, rather than just the private costs to the bank manager of having to go to outside financiers and having to explain to them what he is doing and why his activities should merit their entrusting him with their money.”
My view is that having some insurance and, perhaps more importantly, some price discovery for the costs of systemic risk would be invaluable. It is therefore worth implementing such a system even if its scale is somewhat limited. I would also argue that an imperfect system is still preferable to ad-hoc LOLR interventions that create incentives for reckless risk taking ex-ante, and leave large liabilities for tax-payers ex-post.
SECTION 3: Conclusion
Let me offer two concluding thoughts.
1. Regarding financial regulations, the devil is in the details – and the successive failures of TARP versions 1.0, 2.0, etc. prove this point more than ever. We have enough agreement on the broad principles of financial regulations, and we need to get down to specifics. I would therefore consider any future report that does not include tables, figures, numbers, equations, and specific proposals to be useless rhetoric.
2. We need to be ready to take a tough stand on future regulations for institutions that are too-big-to-fail.
This issue reminds me of the paradox of free trade. The benefits of free trade are widespread and difficult to grasp, while its costs are concentrated and easily publicized. Public support for free trade is therefore structurally weak. Moral hazard created by implicit guarantees is also widespread and difficult to grasp. It shows up in spreads lowered by a few basis points here and there, in slight distortions of comparative advantages, and in overall weaker governance. But the costs of LCFI failures are large and concentrated. It is therefore tempting for regulators to focus too much on bailouts, and too little on incentives. But this is clearly the wrong policy for the long run. Incentives and accountability must be improved, even if it means fighting a regulatory battle with the industry.
Sir Winston Churchill famously remarked that “Britain and France had to choose between war and dishonour. They chose dishonour. They will have war.” If in the hope of ending the crisis quickly, we choose to bail out the banks without making their managers, shareholders and creditors accountable, then we choose dishonour, and we will have more devastating crises."
Me:
Causes Of This Crisis On February 16th, 2009 Don the liberta... says:
"Moral hazard created by implicit guarantees is also widespread and difficult to grasp."
It's not that hard to grasp. Through lobbying, the large banks have purchased an insurance policy that implicitly guarantees that they will be bailed out in a financial crisis. Does anyone doubt that now? It is the main cause of this crisis, in that it allowed large and systemically important banks to forgo prudence and take ludicrous risks. The insurance premiums for systemic risk would either be huge or ineffective. The government has to guarantee certain banks to avoid Calling Runs and Bank Runs. Perhaps a Narrow/Limited Banking sector could be put in place that is guaranteed, to be complemented by other financial concerns which are not guaranteed, but cannot cause either Calling or Banks Runs.
You also don't mention Fraud, Mismanagement, Negligence, and Collusion, the second most important cause of this crisis. Do you really believe that subprime loans being handed out at the height of the housing bubble, when prices were at their peak, was completely honest and aboveboard? Please.
Posts
