53545 CONTAINING SYSTEMIC RISK: ARE REGULATORY REFORM PROPOSALS ON THE RIGHT TRACK? Augusto de la Torre and Alain Ize 1 October 26, 2009 Abstract This note questions two emerging views on ways to tackle systemic risk. As evidenced by the explosive growth of investment banks, which were regulated more lightly because they were assumed to be systemically less important, regulatory unevenness can trigger acutely destabilizing regulatory arbitrage. Hence, unless systemic footprints can be accurately measured and updated, something we think is unlikely, regulating differentially those institutions that are deemed to be the most systemically relevant looks like a perilous return to the past. Similarly, internalizing systemic liquidity risk by taxing maturity mismatches looks like a remnant of idiosyncratic thinking. Matching short liabilities with short assets can protect an individual intermediary's liquidity but at the expense of exacerbating systemic vulnerability. On both issues, the note proposes alternatives which we think are preferable. Progress is being made on the road to reforming prudential regulation. A number of important reports have already seen the light of day, a reform bill is before the U.S. Congress, and significant international coordination efforts are in motion under the G-20 process and the new Financial Stability Board. 2 While the debate remains heated, a strong consensus has emerged on the need to tackle systemic risk. This worthy goal is leading to specific proposals in at least two respects. The first concerns the question of how far to extend prudential regulation and to what institutions. The second addresses the question of how best to dissuade financial intermediaries from unduly relying on short-term wholesale funding. Regarding the first, views appear converging on the notion that systemically important financial intermediaries, whether commercial banks, investment banks, or any other, should be identified up-front and subjected to stricter prudential regulation depending on their "systemic footprint." 3 Regarding the second, agreement seems to be emerging on the view that excessive short-term funding can be avoided through a Pigovian systemic liquidity tax applied to the intermediary's mismatch between the maturity of assets and liabilities. 4 In our view, however, these two specific and increasingly 1 Augusto de la Torre (adelatorre@worldbank.org) is Chief Economist and Alain Ize (aize@worldbank.org) a senior consultant for the Latin America and Caribbean region at the World Bank. This note is based on a World Bank Policy Research Working Paper entitled "Regulatory Reform: Integrating Paradigms" (De la Torre and Ize, 2009), which reviews the Subprime crisis from the perspective of three basic market failures: agency problems, externalities, and mood swings. 2 See Financial Stability Forum (2008), Acharya and Richardson (2009), Brunnermeier et al (2009), FSA (2009), the April 2009 Declaration of the G-20 Summit, and US Treasury (2009), to cite just a few. 3 See Acharya and Richardson (2009), Brunnermeier et al (2009), and US Treasury (2009). 4 See Brunnermeier et al (2009). popular proposals suffer from serious problems because they fail to incorporate some key lessons from the crisis. 1. How to extend and apply prudential regulation? Consider first the view that regulation should vary across intermediaries based on their systemic footprints. The aim of such a differentiation is understandable. It seeks to limit the market advantages that intermediaries derive from being too-big-to-fail (TBTF) or too-interconnected-to- fail (TITF), and the related negative externalities they inflict on the system. Yet this proposal is bound to fail on account of two major interrelated drawbacks: (i) it takes a purely static view of systemic risk; and (ii) it assumes that systemic footprints can be measured in objective, uncontroversial, and permanent ways. A key lesson from financial history is that a regulatory unevenness that may seem inconsequential in the short run can trigger dynamics of regulatory arbitrage that turn out to be acutely destabilizing in the longer run. Financial flows will sooner or later find the line of least resistance, giving the lesser regulated intermediaries a competitive advantage and making them grow to the point where they become systemic behemoths. This became clear on the path to the Subprime crisis, as investors moved in droves to the world of the less regulated, highly leveraged, and short funded investment banks, rapidly increasing their relative size and boosting their systemic footprint to the point where they had to be admitted ex-post to the safety net, no questions asked. If systemic footprints could be easily, continuously, and accurately measured, a differential regulation based on that measure could work. A finely calibrated regulation that would exactly offset the competitive advantages deriving from TBTF or TITF while fully internalizing the impact of the failure of one institution on the rest of the system would limit systemic vulnerabilities at any point in time and ensure that these vulnerabilities do not grow over time. But this seems difficult to achieve in practice. For starters, if "systemic importance" is gauged on the basis of a simple objective criterion, such as size, intermediaries below the size threshold would benefit from lighter regulation. This would encourage them to multiply in numbers and engage in "systemic herding." As a group, they could become as systemically important as a single large intermediary. More elaborate, risk-based distinctions of systemic importance (the so-called "Basel III" approach) also face major hurdles. Measuring the risk exerted by one intermediary on the rest of the system is not an easy matter. It is subject to many competing interpretations of what is truly systemic and in the end prone to provide a false sense of security (just as Basel II did). Moreover, the attempts to measure network externalities would be chasing a moving target, perhaps even a wild goose, considering the continuous bouts of financial innovation, the permanently evolving market structures and linkages, and the incessant attempts by intermediaries to circumvent regulation. As a result, the list of systemically important institutions would likely require frequent changes. Yet, reclassifying institutions could easily become an operational and institutional nightmare, not the least in view of its signaling implications, which could result in a rapidly de facto frozen list. Thus, attempts to 2 introduce excessively fine-tuned differential regulation could well backfire and end up promoting regulatory arbitrage over time in ever more complex and less predictable ways. Rather than going for the high wire act of balancing one type of regulatory distortion with another, we think regulatory neutrality would be better served by simplicity. There are two possible solutions. One is to make all financial intermediaries fit within the same universal banking mode, something close to the European solution. However, unless entry requirements are set very low, which would increase supervisory costs by multiplying the number of supervised institutions, this could limit entry and promote the preponderance of TBTF conglomerates with limited creativity and large non-competitive rents. The alternative--which we find to be superior--is to apply the same prudential regulations to all intermediaries that take deposits and/or borrow in the market, yet allow for the existence of unregulated intermediaries that have no entry requirements (except for a basic license) but can only borrow from the regulated and have no access to the safety net. Our proposal ensures regulatory neutrality. Because the unregulated intermediaries could only fund themselves from the regulated, a dollar lent to a final borrower through an unregulated intermediary would end up paying the same Pigouvian tax as a dollar lent through a regulated intermediary. Hence, systemic risk would be evenly internalized across all possible paths of financial intermediation, whether they involve unregulated intermediaries or not. But, at the same time, and unlike the universal-banking-only solution, our proposed scheme would favor innovation and competition. Because the unregulated intermediaries would not need to meet entry requirements, they could start from scratch. This would facilitate the entry of the smaller and most innovative players, possibly into "niche" or "boutique" intermediation, where the value they add to intermediation is sufficient to offset their higher funding costs. The cost of oversight would remain low by effectively "delegating" supervision of the unregulated to the regulated intermediaries that lend to them under the general oversight of the supervisor. And the most successful of the unregulated intermediaries would eventually grow to become regulated, thereby gaining direct access to capital markets and retail deposits. Admittedly, our proposal does not directly address the TBTF-TITF problem. However, we do not think this is as severe a shortcoming as it may seem. First, it is simply wrong to equate systemic risk with TBTF-TITF. Systemic risk can be present even without TBTF-TITF. Second, systemic risk is not addressed by fueling regulatory arbitrage (which applying stricter prudential regulations to TBTF-TITF institutions would certainly do) but by removing (or at least lessening) the root causes leading to systemic events. A Pigouvian systemic liquidity tax along the lines proposed below, that makes the system sturdier without triggering dynamic instabilities, would be a major step towards this goal. Third, stricter regulation would be of little help in dealing with troubled or unviable TBTF-TITF institutions. That scenario calls for different instruments, particularly a suitable failure resolution framework--with powers to close, intervene, unwind positions, restructure, separate the "good" and "bad" parts of the balance sheet, etc. 3 2. How to curb undue reliance on short-term funding? Consider now the view that the systemic costs of excessive short-term funding could be curbed via a Pigovian tax applied to intermediaries' maturity mismatches. Why is it a wrong approach? Because it fails to adequately take externalities into account. Matching short liabilities with short assets can protect the liquidity of an individual intermediary but at the expense of exacerbating systemic vulnerability. Under systemic events, short loans become as illiquid as long loans, unless intermediaries press borrowers to repay the loans. But if they do so, they shift the liquidity pressure onto somebody else, the final borrower or another intermediary, thereby increasing default risk across the system and contributing to downward asset spirals. Indeed, this is exactly what happened with hedge funds and other institutional investors during the Subprime crisis. The short maturity of their assets allowed them to get, for the most part, home free quickly but jeopardized the stability of the system as a whole. A truly systemic liquidity tax should instead induce investors to retain more of the liquidity risk onto themselves (instead of pushing it on the financial intermediaries or the final borrowers), or intermediaries to hold more systemic liquidity (instead of relying on others to provide it). We therefore propose to turn the conventional maturity mismatch principle--it is ok to borrow short if you also lend short--on its head. The systemic liquidity tax should penalize short funding from uninsured wholesale investors and not allow the intermediaries to offset this tax with what amounts to a "tax credit" from short lending. 5 Instead, short lending--at least to the unregulated financial intermediaries--should itself be taxed (the shorter the loans, the higher the tax). The tax would aim at pricing, at least partially, the value of the option to "lend-short-and-run" that the deposit insurance and the LOLR implicitly provide. To be sure, the tax would increase the private costs of financial intermediation, but that is unavoidable in order to bring them closer to the social costs. Moreover, introducing the tax--we recognize--would involve tradeoffs. Penalizing short transactions would help internalize the systemic costs and benefits of liquidity but weaken the ability of investors and lenders to use a "short leash" (the option to exit quickly) as a disciplining device. We nonetheless think that the benefits of a reasonably calibrated systemic liquidity tax would significantly exceed these latter costs. 5 The tax base would be the uninsured debt funding adjusted for time to maturity minus the intermediary's holdings of systemically liquid assets (i.e., assets that would remain liquid under systemic turbulences). The tax could take various forms, including a capital surcharge or a risk-adjusted premium on the deposit insurance. 4 References Acharya, Viral and Matthew Richardson, 2009, Restoring Financial Stability, John Wiley & Sons. Brunnermeier, Markus, Andrew Crockett, Charles Goodhart, Avinash Persaud, and Hyun Shin, 2009, The Fundamental Principals of Financial Regulation: 11th Geneva Report on the World Economy. De la Torre, Augusto, and Alain Ize, 2009, "Regulatory Reform: Integrating Paradigms", Policy Research Working Paper 4842, World Bank. Financial Services Authority, 2009, "The Turner Review: A Regulatory Response to the Global Banking Crisis", London (March). Financial Stability Forum, 2008, "Report of the Financial Stability Forum on Enhancing Market and Institutional Resilience", Basel (April). G-20, 2009, "Declaration on Strengthening the Financial System", London (April). US Treasury, 2009, "Financial Regulatory Reform: A New Foundation", Washington (June). 5