50347 NOTE NUMBER 6 PUBLIC POLICY FOR THE PRIVATE SECTOR JULY 2009 Macro-Prudential Regulation FINANCIAL AND PRIVATE SECTOR DEVELOPMENT VICE PRESIDENCY Avinash Persaud Fixing Fundamental Market (and Regulatory) Failures Avinash Persaud This is not the fir s t int e r na t io na l b a nk ing c r is is t he wo r ld h a s s e e n . (avinash@intelligence -capital.com) is chairman The p re vious on e s o c c ur r e d wit ho ut c r e d it d e f a ult s wa p s , s pe c i a l of Intelligence Capital investme nt ve hi c le s , o r e v e n c r e d it r a t ing s . I f c r is e s k e e p r e p e a t i n g Limited and emeritus the mselve s, it s e e m s r e a s o na b le t o a r g ue t ha t p o lic y m a k e r s n e e d t o professor at Gresham College in London. caref ully consid e r wha t t he y a r e d o ing a nd no t jus t " d o ub le u p " b y sup e rf icially re a c t ing t o t he s p e c if ic f e a t ur e s o f t o d a y ' s c r is i s . W h i l e This is the sixth in a series of policy briefs on we cannot hop e t o p r e v e nt c r is e s , we c a n p e r ha p s m a k e t he m f e w e r the crisis--assessing the and mild er b y a d o p t ing a nd im p le m e nt ing b e t t e r r e g ula t io n -- i n policy responses, shedding light on financial reforms p articular, more m a c r o - p r ud e nt ia l r e g ula t io n. currently under debate, There is a widely held view that the current finan- 85th.1 If crises keep repeating themselves, it seems and providing insights cial crisis resulted from an insufficient reach of reasonable to argue that policy makers need to for emerging-market policy makers. regulation and that the solution is to take existing carefully consider what they are doing and not regulation and spread it without gaps across insti- just "double up." It also means that policy makers tutions and jurisdictions. If this were to be the should not superficially react to the characters main policy response, it would be a mistake for and colors of the current crisis. The last 84 crises several reasons. The most important one is that occurred without credit default swaps and special at the heart of the crisis lay highly regulated insti- investment vehicles. The last 80-something had tutions in sophisticated jurisdictions--Northern nothing to do with credit ratings. The solution THE WORLD BANK GROUP Rock, IKB, Fortis, Royal Bank of Scotland, UBS, to the crisis is not more regulation, though more Citigroup. If there were no mortgage fraud, no comprehensive regulation may be required in tax secrecy, and no conflicts of interest, a crisis some areas. Instead, it is better regulation--in would still have occurred. And while risk did shift particular, regulation with a greater macro- outside the capital adequacy regime, the special prudential orientation, as recommended by investment vehicles were not secret and supervi- numerous recent official reports.2 sors had the discretion to look at how regulated institutions were managing risks and to respond What is macro-prudential regulation? if necessary. It seems banal today to point out that the reason This is not the first international banking crisis we try to prevent financial crises is that the costs to the world has seen. By some estimates it is the society are invariably enormous and exceed the MACRO-PRUDENTIAL REGULATION FIXING FUNDAMENTAL MARKET (AND REGULATORY) FAILURES private cost to individual financial institutions. degree of leverage, and interconnectedness with We regulate to internalize these externalities in the rest of the system. the behavior of such institutions. One of the main The existing framework of banking regulation tools regulators use to do this is capital adequacy was insufficiently macro-prudential and had been requirements. But the current approach to capi- recognized as such by commentators for some tal adequacy is too narrow. Capital adequacy lev- time (see Borio 2005; Borio and White 2004; and els are set on the implicit assumption that we can Persaud 2000). Moreover, the emphasis on micro- make the system as a whole safe by ensuring that prudential regulation may have contributed to individual banks are safe. This represents a fal- the buildup of some macro risks. 2 lacy of composition. In trying to make themselves Through many avenues, some regulatory safer, banks and other highly leveraged financial and some not, often in the name of prudence, intermediaries can behave in ways that collec- transparency, and sensitivity to risk, the growing tively undermine the system. This is in essence influence of current market prices has intensi- what differentiates macro-prudential from micro- fied homogeneous behavior in financial systems. prudential concerns. These avenues include mark-to-market valuation Here is an example of macro-prudential of assets; regulator-mandated market-based mea- concerns. Selling an asset when it appears to be sures of risk, such as the use of credit spreads risky may be considered a prudent response for in internal credit models or price volatility in an individual bank and is supported by much market risk models; and the increasing use of current regulation. But if many banks do this, credit ratings, where the signals are slower mov- the asset price will collapse, forcing risk-averse ing but positively correlated with financial mar- institutions to sell more and leading to general kets. Where measured risk is based on market declines in asset prices, higher correlations and prices, or on variables correlated with market volatility across markets, spiraling losses, and prices, it can contribute to systemic risk as market collapsing liquidity. Micro-prudential behavior participants herd into areas that appear to be can cause or worsen systemic risks. A macro- safe.3 And measured risk can be highly procycli- prudential approach to an increase in risk is to cal, because it falls in the buildup to booms and consider systemic behavior in the management rises in volatile busts. of that risk: who should hold it, and do they have the incentive to do so? If it is liquidity risk, is it Macro-prudential regulation and the cycle in the interests of the system if all institutions, The economic cycle is a major source of homo- regardless of their liquidity conditions, sell the geneous behavior, so addressing it is a critical same asset at the same time? Risk in a financial macro-prudential concern. In the up phase of system is more than an aggregation of risks in the cycle, price-based measures of asset values individual institutions; it is also about endog- rise, price-based measures of risk fall, and com- enous risks that arise as a result of the collective petition to increase bank profits grows. Most behavior of institutions. financial institutions spontaneously respond by Macro-prudential regulation concerns itself expanding their balance sheets to take advan- with the stability of the financial system as a whole. tage of the fixed costs of banking franchises and By contrast, micro-prudential regulation, con- regulation; trying to lower the cost of funding by sisting of such measures as the certification of using short-term funding from money markets; those working in the financial sector and rules and increasing leverage. Those that do not do so on how financial institutions operate, concerns are seen as underleveraging their equity and are itself with the stability of individual entities and punished by stock markets. In the more prosaic the protection of individuals. Micro-prudential words of former Citigroup CEO Chuck Prince, regulation examines the responses of an indi- in a July 2007 interview with the Financial Times, vidual bank to exogenous risks. By construction, "when the music is playing, you have to get up it does not incorporate endogenous risk. It also and dance." By contrast, when the boom ends, ignores the systemic importance of individual asset prices begin to fall and short-term fund- institutions resulting from such factors as size, ing to institutions with impaired and uncertain assets or high leverage dries up. Forced sales of Politicians want to reap electoral benefit from assets drive up their measured risk, and the boom the sense of well-being and prosperity during inevitably turns to bust. a boom. Policy officials convince themselves, One of the key lessons of this crisis is that and try to convince others, that the boom is not market discipline is little defense against the an unsustainable credit binge but the positive macro-prudential risks that come with the eco- result of structural reforms that they have put nomic cycle. The institutions that have been into place. Booms have social benefits. They are most resilient to the crisis, such as HSBC and associated with a higher appetite for risk and a J.P. Morgan, had lower equity "ratings" (lower perception that risks have fallen, and this often 3 price-earnings ratios) than those that proved to means greater access to finance for the previ- be less resilient, such as Northern Rock, Bear ously unbanked and underinsured. Booms are Stearns, Fortis, and Lehman Brothers. Market not quite a conspiracy of silence, but there are discipline has an important role to play in the few who gain from their early demise. So booms efficiency of the financial sector, but it cannot be tend to be explained away, excused, and accom- on the front line of defense against crises. modated, allowing them to grow larger and larger One reason that market discipline was seen as and thus to cause more damage when they even- such an important pillar in the precrisis approach tually collapse. to banking regulation was the implicit model that regulators had in mind: financial crashes occur Countercyclical charges and buffers randomly as a result of a bad institution failing, In light of the observations above, there is a grow- and that failure becomes systemic. The histori- ing consensus around three ideas: Capital require- cal experience is rather different: crashes follow ments need to have a countercyclical element in booms. In the boom almost all financial institu- order to, in the words of the G-20 communiqué tions look good, and in the bust almost all look of April 2, "dampen rather than amplify the finan- bad. Differentiation is poor. The current crisis cial and economic cycle" by "requiring buffers of is another instance of this all-too-familiar cycle. resources to be built up in good times." There But if crises repeat themselves and follow booms, should be greater emphasis on rules rather than banning the products, players, and jurisdictions supervisory discretion to counterbalance the politi- that were merely the symptoms of the latest boom cal pressures on supervisors. And these rules should will do little to prevent the next one. include leverage limits and liquidity buffers. Moreover, the notion that some financial The references in the G-20 communiqué echo products are safe and some are not, and that a statement by the Basel Committee on Banking the use of unsafe products is the problem, also Supervision following its March 2009 meeting, looks suspect in a boom-bust world. The booms recommending the "introduction of standards to are often a result of things appearing to be safer promote the buildup of capital buffers that can be than they are. Securitization was viewed as a way drawn down in periods of stress." These statements of making banks safer. Diversified portfolios of by the G-20 and the Basel Committee, coupled with subprime mortgages were viewed as having low similar conclusions by other official reports, sug- delinquency rates. Micro-prudential regulation gest that the argument in favor of macro-prudential is necessary to weed out the truly reckless insti- regulation has been won. But how countercycli- tutions and behavior. But it needs to be supple- cal capital charges and liquidity buffers are to be mented with macro-prudential regulation aimed implemented has not yet been addressed in great in part at acting as a countervailing force against detail. Given the politics of booms, the how is almost the decline of measured risk in a boom (and thus as important as the whether. excessive levels and interconnectivity of risk tak- In practical terms, Goodhart and Persaud ing) and against the rise of measured risk in the have recommended that regulators increase the subsequent collapse. existing or base capital adequacy requirements Supervisors have plenty of discretion, but (based on an assessment of inherent risks) by two they find it hard to use because of the politics multiples calculated using a few simple, transpar- of booms. Almost everyone wants a boom to last. ent rules.4 MACRO-PRUDENTIAL REGULATION FIXING FUNDAMENTAL MARKET (AND REGULATORY) FAILURES The first multiple would be a function of the which increases systemic fragility and intercon- growth of credit and leverage. Regulators should nectedness. This private incentive to create sys- meet with monetary policy officials (where they temic risk can be offset through new capital or are separate) in a financial stability committee. reserve requirements. It is partly this notion that This meeting would produce a forecast of the the G-20 communiqué refers to when stating that growth of aggregate bank assets that is consis- the G-20 leaders have agreed to introduce mea- tent with the central bank's target for inflation sures "to reduce the reliance on inappropriately (or other macroeconomic nominal target). The risky sources of funding." Liquidity buffers, with forecast would have a reasonable band around their size related to maturity mismatches between 4 it reflecting uncertainty. If a bank's assets grow assets and liabilities, would have similar effect. above this band, the bank would have to put aside But once again there is little discussion of meth- a higher multiple of its capital for this new lend- odology and implementation. Measuring the true ing. If its assets grow less than the lower bound, maturity of bank assets and liabilities is not a it may put aside a lower multiple. straightforward exercise. For example, suppose that the financial stabil- In the framework set out in the Geneva Report ity committee concluded that growth in aggre- (Brunnermeier and others 2009), assets that can- gate bank assets of between 7.5 percent and 12.5 not be posted at the central bank for liquidity percent was consistent with its inflation target of 3 can be assumed to have a minimum maturity percent. Growth in a bank's assets by 25 percent, of two years or more. If a pool of these assets or twice the upper range, may lead to a doubling was funded by a pool of two-year term deposits, of the minimum capital adequacy level from 8 there would be no liquidity risk and no liquidity percent to 16 percent of risk-weighted assets. A charge. But if the pool of funding had a maturity related approach is to have one minimum capital of one month and so had to be rolled over every adequacy requirement for "bad" times and one month, the liquidity multiple on the base capital that is twice that level for "good" times, with good charge would be near its maximum--say 2, so the and bad times being determined by bank prof- minimum capital adequacy requirement would itability. Of course it is impossible to ascertain rise from 8 percent to 16 percent. whether these capital levels would have made In a boom in which the first countercycli- the system safe, but the consensus today is that cal multiple is also 2, the final capital adequacy they would have at least made it safer. requirement would be 32 percent of risk-weighted Financial stability committees exist in many assets (8 percent 2 2). Liquidity multiples countries. But they generally work poorly because would make lending costlier, since banks tradi- their deliberations have no consequence. tionally fund themselves short and lend long. Requiring such committees to agree on a sus- But the liquidity multiples would give banks an tainable level of growth in bank assets could incentive to find longer-term funding, and where make their work more penetrating and action they cannot do so, a liquidity buffer or liquidity oriented. reserve that could be drawn down in times of The second multiple on capital requirements stress would buy time for institutions to deal with would be related to the mismatch in the matu- a liquidity problem. rity of bank assets and liabilities. One significant lesson of the crisis is that the risk of an asset is Can the cycle be measured? determined largely by the maturity of its funding. Many people, most notably former U.S. Federal Northern Rock and other casualties of the crash Reserve Chairman Alan Greenspan, voice the might well have survived with the same assets if concern that it is very hard to know when we the average maturity of their funding had been are in a boom. Of course, measuring the cycle longer. The liquidity of banks' assets has fallen far is what inflation-targeting central banks do on more than the credit quality of those assets. a daily basis. But this misses the point a little. If If regulators make little distinction on how the purpose of countercyclical capital charges assets are funded, however, financial institu- were to end boom-bust cycles, we would need tions will rely on cheaper, short-term funding, to be more confident about the calibration of booms than we are today. But if the purpose is were still performing from a credit point of view, to lean against the wind, our calibrations can be but had become highly illiquid, had long-term less precise. funding. In the absence of fair-value accounting Recall that without countercyclical charges, standards, they would not have joined the selling the natural inclination in a boom is to lend even frenzy that compounded the crisis. Second, with- more because measured risks fall. The precrisis out the mark-to-market volatility, institutions with regulatory approach took the economic cycle long-term funding would have been more willing and amplified it. The goal instead should be to to buy these assets. That would have provided moderate the worst excesses of the cycle, not to greater price support, limiting the spiral of losses 5 kill it. Indeed, the cycle is an important source that endangered so many banking institutions. of creative destruction in our economic system. Compensation Valuation and mark-to-funding accounting In the G-20 communiqué and elsewhere, great Many commentators consider accounting issues attention is given to dealing with the incentives to be central in the crisis. They argue that the use of individual bankers and traders. But there are of fair-value accounting has added to the spiral clear limits to how much governments should of sales. But suspending fair-value accounting is be involved in private firms' decisions on execu- not helpful in an environment made worse by tive pay. While measures to lengthen bankers' uncertainty. Instead, financial institutions should horizons are necessary, greater hopes should be complement mark-to-market accounting with placed in macro-prudential regulation pushing mark-to-funding valuations (see Brunnermeier banks to develop incentive packages that better and others 2009). promote through-the-cycle behavior. If that failed, Under mark-to-funding valuations there are however, regulators should certainly do more to essentially two alternative prices for an asset: address the important issue of incentives. today's market price and the discounted present value of the future earnings stream. In normal Macro-prudential regulation beyond the cycle times these two prices are nearly the same. But The other dimension of macro-prudential regu- in a liquidity crisis the market price falls substan- lation is the cross-sectional one: how to manage tially below the present value. If an institution the buildup of risks arising from the structure of has short-term funding, the realistic price to use the financial system. is the market price. If it has long-term funding, the present-value price is a better measure of the Risk assignment risks faced by the institution. Under a mark-to- Requiring the banking system to hold more capi- funding accounting framework, a weighted aver- tal on average will not improve the resilience of age of the market price and present-value price the financial system as a whole unless there is would be used whose weights would depend on also a better match of risk taking to risk capac- the weighted average maturity of the institution's ity. Indeed, piling up capital requirements may funding. The combination of liquidity charges act as an anticompetitive barrier, reinforcing the and mark-to-funding value accounting would cre- specter of a few banks holding a government ate incentives for institutions to seek longer-term hostage because they are too big to fail. funding and would encourage a tendency for Micro-prudential regulation was often accom- illiquid assets to be owned by institutions with panied by a misguided view of risk as an absolute, longer-term funding. constant property of an asset that can be mea- At first sight, mark-to-funding would not sured, sliced, diced, and transferred. This is an appear to alleviate the problem facing banks elegant view of risk and has the merit of allowing today--in fact, it could make matters worse-- banks to build highly complex valuation mod- because they have short-term funding. But this els and to sell highly complex risk management proposal would have had two ameliorating effects products to handle and distribute risk. But it is in the crisis. First, many of the bank-owned spe- also an artificial construct that has little bearing cial investment vehicles that managed assets that on the nature of risk. MACRO-PRUDENTIAL REGULATION FIXING FUNDAMENTAL MARKET (AND REGULATORY) FAILURES In reality, there is not one constant risk. The rent system, the natural risk absorbers behave like three broad financial risks--credit risk, liquid- risk traders, selling and buying when everyone ity risk, and market risk--are very different. else is doing so. Moreover, the potential spillover risk from some- Capital requirements encouraging those with one holding an asset depends as much on who is a capacity to absorb a type of risk to hold that holding the asset as on what it is. Different hold- risk not only will make the system safer without ers have different capacities for different risks. destroying the risk taking that is vital for eco- The distinction between "safe" and "risky" assets nomic prosperity; they also will introduce new is deceptive: one can do a lot of damage with a players with risk capacities. This would both 6 simple mortgage, for example. strengthen the resilience of the financial system The capacity for holding a risk is best assessed and reduce our dependence in a crisis on a few by considering how that risk is hedged. Liquidity banks that appeared to be well capitalized during risk--the risk that an immediate sale would lead the previous boom. to a large discount in the price--is best hedged over time and is best held by institutions that do Systemic institutions not need to respond to an immediate fall in price. Not all financial institutions pose systemic risks. A bank funded with short-term money market Regulation should acknowledge that some banks deposits has little capacity for liquidity risk. Credit are systemically important, and others less so. risk--the risk that someone holding a loan will In each country supervisors establish a list of default--is not hedged by having more time for systemically important institutions that receive the default to happen but by having offsetting closer scrutiny and require greater containment credit risks. Banks, with access to a wide range of behavior. Critical factors that determine sys- of credits, have a far greater capacity than most temic importance for an institution, instrument, to diversify and hedge credit risks. or trade are size of exposures, especially with The way to reduce systemic risk is to encourage respect to the core banking system and retail individual risks to flow to where there is a capacity consumers; degree of leverage and maturity for them. Unintentionally, much micro-prudential mismatches; and correlation or interconnectiv- regulation did the opposite. By not requiring firms ity with the financial system. to put aside capital for maturity mismatches and by In the past, interconnectivity has been under- encouraging mark-to-market valuation and daily stood to include issues such as payment and set- risk management of assets by everyone, regulators tlement systems, and these remain vital. Today, encouraged liquidity risk to flow to banks even interconnectivity may also include institutions though they had little capacity for it. By requiring that behave in a highly correlated manner even banks to hold capital against credit risks, regula- if individually they appear small relative to the tors encouraged credit risk to flow to those that size of the financial system. were seeking the extra yield, were not required to Goodhart and Persaud, as members of the set aside capital for credit risks, and had limited UN Commission of Experts on Reforms of the capacity to hedge that risk. No reasonable amount International Monetary and Financial System, of capital can remedy a system that inadvertently have urged the commission to recommend leads to risk-bearing assets being held by those establishing a list of systemically important instru- without a capacity to hold them. ments. And where instruments are declared sys- What can regulators do? They need to differ- temically important because of their volume, entiate institutions less by what they are called link to leverage, or interconnectivity, they rec- and more by how they are funded. They should ommend requiring that the instruments be reg- require more capital to be set aside for risks where istered and, where appropriate, exchange traded there is no natural hedging capacity. This will and centrally cleared. draw risks to where they can be best absorbed. They also must work to make value accounting Host and home country regulation and risk management techniques sensitive to A gathering view is that financial institutions are funding and risk capacity. Instead, under the cur- global and so financial regulation needs to be global. But reality does not rhyme so easily. The for individual institutions to lend more. Micro- crisis would not have been averted by more inter- prudential regulation is not enough; it must be national meetings, and it has taught us that there supplemented by macro-prudential regulation is much that needs to be done at the national level that catches the systemic consequences of all to strengthen regulation. Countercyclical and institutions acting in a similar manner. While we liquidity charges cannot be set or implemented cannot hope to prevent crises, we can perhaps globally but need to be handled nationally in make them fewer and milder by adopting and accordance with national cycles. implementing better regulation--in particular, Although there is a clear need for cross- more macro-prudential regulation. 7 border sharing of information and coordination of regulatory actions and principles (particularly in micro-prudential regulation), the setting of capital rules and banking supervision is likely to switch Notes back from "home country" to "host country." This The author would like to thank Constantinos Stepha- should not be resisted because it would have two nou, Aquiles Almansi, and Damodaran Krishnamurti additional benefits, particularly for emerging for their comments, although the views expressed in economies. First, if foreign banks were required to this policy brief remain those of the author. set up their local presence as independent subsid- 1. For a discussion on the history of financial crises, see iaries that could withstand the default of an inter- Reinhart and Rogoff (2008). national parent, it would reduce exposure to lax 2. These include the April 2 communiqué of the G-20 jurisdictions more effectively than trying to force leaders, the Turner Review (FSA 2009), the G-30 report everyone to follow a standard that could be inap- (2009), the de Larosiere Group report (2009), the UN propriate and would in any case be enforced with Commission of Experts recommendations (2009), and different degrees of intensity. Second, nationally the 11th Geneva Report (Brunnermeier and others set countercyclical charges could give common- 2009). currency areas or countries with fixed or managed 3. See Persaud (2000) for a discussion on how, through exchange rates a much-needed additional policy the financial sector's use of value-at-risk models, "the instrument--one that could provide a more dif- observation of safety creates risk and the observation of ferentiated response than a single interest rate risk creates safety." The late economist Hyman Minsky could to a boom in one member state and defla- also argued in more general terms, and long before tion in another. This policy instrument may also the advent of value-at-risk models, that risks are born in be important in emerging economies, where, per- periods of stability. haps as a result of the absence of developed bond 4. The original ideas were published in Goodhart and and currency markets, interest rates are not an Persaud (2008a, b) and expanded in Brunnermeier and effective regulator of the economic cycle. others (2009). Conclusion References Warren Buffett famously remarked that you Borio, C. 2005. "Monetary and Financial Stability: So see who is swimming naked only when the tide Close and Yet So Far." National Institute Economic runs out. By this, he probably means that while Review 192 (1): 84­101. fraud and unethical practices are going on all Borio, C., and W. White. 2004. "Whither Monetary and the time, they become visible only when the veil Financial Stability? The Implications of Evolving of rising market prices is removed. They are not Policy Regimes." BIS Working Paper 147, Bank for the cause of the tide going out; they are merely International Settlements, Basel. revealed by it. We must continue to clamp down Brunnermeier, M., A. Crockett, C. A. E. Goodhart, A. on fraud and ethical abuses and promote trans- D. Persaud, and H. Shin. 2009. The Fundamental Prin- parency, but this is not enough to avoid crises. ciples of Financial Regulation. Geneva Report on the We cannot avoid crises without avoiding the World Economy 11. Geneva: International Center booms--booms that are always underpinned for Monetary and Banking Studies; London: Centre by a good story explaining why it is prudent for Economic Policy Research. MACRO-PRUDENTIAL REGULATION FIXING FUNDAMENTAL MARKET (AND REGULATORY) FAILURES de Larosiere Group. 2009. Report of the High-Level Group on Financial Supervision in the EU. Brussels. FSA (U.K. Financial Services Authority). 2009. The Turner Review: A Regulatory Response to the Global Bank- ing Crisis. London. Goodhart, C. A. E., and A. D. Persaud. 2008a. "How to crisisresponse Avoid the Next Crash." Financial Times, January 30. ------. 2008b. "A Party Pooper's Guide to Financial The views published here Stability." Financial Times, June 5. are those of the authors and G-30 (Group of Thirty). 2009. Financial Reform: A Frame- should not be attributed work for Financial Stability. Washington, DC. to the World Bank Group. Persaud, A. 2000. "Sending the Herd off the Cliff Edge: Nor do any of the conclusions The Disturbing Interaction between Herding and represent official policy of Market-Sensitive Risk Management Systems." First the World Bank Group or Prize Essay, Jacques de Larosiere Award in Global of its Executive Directors or Finance, Institute of International Finance, Wash- the countries they represent. ington, DC. Reinhart, C., and K. Rogoff. 2008. This Time Is Different: To order additional copies A Panoramic View of Eight Centuries of Financial Crises. contact Suzanne Smith, NBER Working Paper 13882. Cambridge, MA: Na- managing editor, tional Bureau of Economic Research. The World Bank, UN Commission of Experts on Reforms of the Interna- 1818 H Street, NW, tional Monetary and Financial System. 2009. Recom- Washington, DC 20433. mendations. New York: United Nations. Telephone: 001 202 458 7281 Fax: 001 202 522 3480 Email: ssmith7@worldbank.org Produced by Grammarians, Inc. Printed on recycled paper This Note is available online: http://rru.worldbank.org/PublicPolicyJournal