Charles W. Calomiris

May 1999

Although many commentators discuss the need to reform the International Monetary Fund, few have offered detailed proposals to achieve that purpose. The following essay, however, offers a comprehensive plan that would end the IMF practice of bailing out troubled emerging markets and would return the fund to addressing real liquidity problems.

Economics normally provides rather dismal news—emphasizing necessary tradeoffs among objectives. In the attempt to redesign global financial architecture, however, such is not the case. It is not difficult to construct a set of mechanisms that concurrently resolve problems of illiquidity (by providing a responsive international lender of last resort facility alongside a domestic deposit insurance system) while avoiding the governance and incentive problems attendant to counterproductive bailouts of risk takers. Avoiding those problems entails establishing a mechanism that ensures credible market discipline of financial institutions.

The hurdles that must be overcome in designing an appropriate global financial architecture, then, are not those posed by economics, but rather by politics. The challenge lies with persuading those with vested interests in the current allocation of political power—including bankers, developing country oligarchs, and the U.S. Treasury—to relinquish some of the power they currently wield in order to make the global financial system more efficient, competitive, and democratic.

The details of the plan are summarized in the chart below. It must be emphasized that the success of each component of the proposal depends on all the other components. Without a reliable means of bringing credible market discipline to bear on banks to provide strong incentives for prudent risk management, government deposit insurance and IMF lending will spur excessive risk taking with its attendant costs. But in the presence of credible market discipline, deposit insurance and IMF lending (if structured properly) can strengthen the financial system by helping it avoid liquidity crises, which can result either from inadequate information or self-fulfilling expectations.

The Calomiris Plan in Brief

Membership Criteria for the IMF

Bank regulations:

Basle standards (but without restrictions on subordinated debt/tier 2 capital)

2% subordinated debt requirement (with rules on maturities, holders, and yields)

20% cash reserve requirement

20% "global securities" requirement

Free entry by domestic and foreign investors into banking

Bank recapitalizations are permitted, but strict guidelines must be met (and must follow preestablished rules, as in preferred stock matching program)

Domestic lenders of last resort avoid bank bailouts by following Bagehotian principles

Other membership criteria:

Limits on short-term government securities issues

If fixed exchange rate, 25% minimum central bank reserve requirement

If fixed exchange rate, banks offer accounts in domestic and foreign currencies

IMF Lending Rules

Loans are provided only to members in good standing (those following above rules)

If a member defaults, it may not borrow for 5 years, and then only after arrears paid

Loans are for 90 days

Supernumerary majority of members required to roll over loans for another 90 days

Loans are collateralized by 125% of value of loan in government securities

25% of the 125% collateral must be in foreign government securities

The interest rate on the loan is set at 2% above the value-weighted yield on the collateral observed one week before the loan request

The IMF reserves the right to refuse a loan to a member

No conditions are attached to IMF loans

IMF Funding

The IMF borrows from the discount windows of the Fed and other central banks

IMF borrowing from central banks is 100% collateralized by government securities issued by the government of the lending central bank

Government securities that serve as collateral for IMF borrowing from central banks is lent to the IMF by its member countries

Other Emergency Lending

IMF, World Bank, IDB, and others would make no other emergency lending available

The Exchange Stabilization Fund would be abolished

	EXISTING IMF	NEW IMF
GOALS	Bail outs, political aid, liquidity assistance	Bona fide liquidity assistance
WHO QUALIFIES?	Anyone!	Members meeting criteria
STRUCTURE OF LOANS	Varies, not clear	Collateralized, short-term
CONDITIONALITY	Negotiated	None, except membership
TIMING	Delayed, pending negotiation	Immediate
INTERSECTION WITH THE EXCHANGE STABILIZATION FUND	Frequent	None
INTERSECTION WITH WORLD BANK	Duplication, joint-effort	Clear boundaries: World Bank to assist in liberalization

In implementing such rules, the devil is in the details, hence my emphasis is on "blueprints" (specific concrete proposals) rather than simply on organizing principles. Slight differences in details can make the difference between a reform agenda that achieves both liquidity and proper incentives toward risk taking and one that achieves neither. Before discussing those details, let me review the problems my proposal is meant to address.

Why We Need a New IMF

Over the past twenty years, ninety banking crises have occurred in equal or greater magnitude (in terms of banking system losses) to the U.S. banking experience during the Great Depression. In at least ten cases, banking system losses exceed 20 percent of GDP—a staggering and unprecedented set of losses, which are occurring during a time of relatively stable and rapid global growth.

Banking system collapse due to excessive risk taking by banks has been a common feature of all the recent financial collapses. Bank losses precede and cause exchange rate collapses. In other words, banking systems have become the key source of financial instability. Economists have pointed to several core problems that feed that instability. First and foremost are incentive problems that encourage risk taking, particularly in response to adverse macroeconomic shocks.

Before banks were protected by government safety nets, economic downturns produced contractions of bank credit supply and cuts in bank dividends, as banks scrambled to reassure depositors that bank loan losses would not result in losses for depositors. Safety net protection has removed that important disciplinary check on bank behavior. Safety net protection (ultimately, taxpayer protection of banks and their claimants) relaxes market discipline on bank risk taking and subsidizes higher risk in banks. This effect is especially pronounced after banks experience initial losses to the value of their assets. In the wake of such losses, safety net protection encourages banks to consciously increase their asset risk. Those increases in risk often take the form of increased default risk and exchange rate risk after banks have already seen severe depletion of their capital.

Bailouts of developing economies’ banks and international bank lenders, orchestrated by domestic governments in cooperation with the IMF and the U.S. Treasury, must stop. Not only do they produce inefficient risk taking, fiscal disasters for domestic governments, and enormous distortionary taxes to pay the costs of the bailouts, they also undermine the core competitive and democratic processes on which successful financial systems depend (by supporting crony capitalism within both developed and developing economies).

The IMF didn’t invent bank bailouts, and IMF involvement in bailouts is mainly indirect. Nonetheless, it is quite destructive. The IMF provides only a small wealth transfer to its borrowers in the form of its loan subsidy, and so does not directly pay for much of the cost of the bailout. But the IMF pressures borrowers to bail out foreign bank lenders, and it lends support and legitimacy to domestic bailouts too by requiring government taxation to finance the repayment of IMF loans.

The destruction wrought by these bailouts has led many—including George Schultz and Anna Schwartz—to call for abolition of the IMF. Others argue, however, that liquidity assistance by the IMF could be useful if properly designed. Indeed, IMF liquidity assistance has sometimes been helpful. The most obvious case may be the March 1995 IMF loan to Argentina. Here the IMF lent to a government that had pursued significant, tangible fiscal and bank regulatory reforms and did so with the express goal of financing a defense of the Argentine currency board (not financing a bailout of banks or other government expenditures).

Addressing Liquidity Problems

The two potentially most important liquidity problems are (1) banking panics that result from temporary confusion on the part of bank debt holders about the incidence within the banking sector of losses attendant to an observable shock and (2) self-fulfilling collapses of fixed exchange rates that result from government illiquidity.

To solve the first problem, I propose a set of banking regulations that together would remove the threat of banking panics including: (1) capital standards founded on market discipline, achieved through a requirement that banks maintain a minimal proportion of uninsured, junior (or subordinated) debt; (2) credible deposit insurance for other bank debt claims; (3) a 20 percent "cash" (or equivalents) reserve requirement (relative to bank assets); (4) a 20 percent "global securities" requirement (relative to bank assets); (5) free entry by domestic and foreign competitors into banking; and (6) limitations on other government assistance to banks (which limit the transfer of bank losses to taxpayers).

My plan for designing effective, credible market discipline for banks is largely based on the Chicago Federal Reserve’s 1989 subordinated debt proposal, although there are some differences. It is important to emphasize that a broad consensus has emerged on the need to add some form of subordinated debt requirement to the Basle capital standards. Advocates of some form of such a requirement now include: the Bankers’ Roundtable, the U.S. Treasury, several Federal Reserve Banks, at least one Fed Governor, members of Congress, and the Shadow Financial Regulatory Committee. Last year, the Federal Reserve Board assembled a task force charged with exploring how best to design and implement a subordinated debt requirement. Outside the United States, several countries either have passed or are considering such a requirement.

The combination of domestic deposit insurance and market discipline (which prevents the abuse of deposit insurance) can resolve the threat of banking panics that result either from confusion about the incidence of shocks or from self-fulfilling concerns about the insufficiency of bank reserves. The IMF’s role would be mainly to address the other liquidity problem—liquidity crises that face member governments as the result of unwarranted speculative pressure on exchange rates. This was the original intent of the IMF’s founders, and it remains a legitimate objective of IMF policy.

Recent studies that emphasize the value of IMF liquidity protection argue that the current form of IMF assistance is inadequate—it is too little, too late, and with too many conditions and delays to be effective in short-circuiting self-fulfilling runs on currencies or government debt. But how does one provide effective liquidity protection without encouraging counterproductive bailouts of banks and/or governments?

My plan is to replace the current IMF and Exchange Stabilization Fund with a new IMF, which would offer a discount window lending facility. That facility would only be available to IMF members—and membership would require adherence to the aforementioned banking regulations, as well as some additional rules regarding government debt management, a 25 percent minimum reserve requirement for the central bank (if a fixed exchange rate is maintained), and a requirement that banks offer accounts denominated in both domestic and foreign currency.

By restricting access to the IMF window to members in good standing who conform to a few simple and easily verified rules, the IMF avoids free-riding on liquidity protection and the hazard of unwittingly financing bank bailouts in the guise of liquidity protection. The rules governing the discount window follow Walter Bagehot’s classic principles for ensuring liquidity, while avoiding free riding: lend freely on good collateral at a penalty rate. The specifics of membership rules, limits on collateral, and penalty lending rates (described above in "The Calomiris Plan in Brief") encourage member countries’ central banks (like their other banks) to diversify their securities portfolios and maintain adequate liquid reserves.

If a member is in good standing, loans are made available on good collateral using one-week-old prices to value collateral. The loan interest rate is set at 2 percent above the value-weighted yield to maturity on the collateral offered. That provides a fast and effective means to short-circuit a self-fulfilling "bad equilibrium."

Other Features

Why no additional conditions for loans? For liquidity assistance to be effective, it must be delivered quickly and supplied elastically. Why no more rules for members? There are lots of basic rules countries should meet to build effective domestic financial systems. These include accounting standards, procedures for registering collateral interests, court enforcement of creditors’ and stockholders’ rights, a transparent and efficient bankruptcy code, and many more. But these rules are more controversial (for example, I would argue that the Swedish bankruptcy code is far superior to the American) and are hard to specify in a simple way. I have not tried to construct a list of all desirable rules, but rather, a set of minimal rules that are important, simple, and verifiable.

Furthermore, additional rules (accounting, bankruptcy, and commercial laws) will arise endogenously if there is credible market discipline within the financial system, which the subordinated debt requirement and other rules will ensure. Market discipline is a lever for other reforms because it creates a strong constituency of banks and their debt-holders who will seek ways to improve transparency, contract enforceability, and sensible workout procedures.

How will the IMF actually operate the new window? IMF lenders would contribute (that is, lend) bonds to the IMF, which along with borrowers’ collateral would be used to access the discount window of the hard-currency central banks (which would lend cash to the IMF collateralized 100 percent by the government securities of that central bank’s government). The hard-currency central banks, thus, would lend without risk. They would also be free to sterilize the effects of IMF borrowing on the aggregate supply of hard currency.

To avoid the potential for costly bailouts, other redundant mechanisms would be abolished. That would include not only other IMF assistance, but also the Exchange Stabilization Fund and emergency assistance to banks via the World Bank and Inter-American Development Bank.

Political Feasibility

Having argued that this plan would achieve proper incentives in private banking and in government finance and would also protect against liquidity crises, I now turn to the more difficult question: whether it is politically feasible. Clearly, vested interests have reasons to oppose this approach because it would deprive them of valuable (though socially costly) subsidies. It may be possible, and worthwhile, to "buy off" those vested interests (particularly within developing countries’ banking systems) by offering a one-time injection of public funds to help recapitalize banks, and thereby make the pill of market discipline easier to swallow.

It is worth considering how domestic governments interested in implementing true reform might be helped by the World Bank and other development banks to achieve membership in the newly constituted IMF. Too often the World Bank has crowded out private lending and removed incentives for countries to adopt credible market discipline. World Bank loans to China are the clearest example of this problem. But in some cases (notably in Argentina recently) the World Bank has provided subsidies to make privatization and market discipline more achievable by rewarding cronies (e.g., provinces that control public banks) for giving up their instruments of patronage. More World Bank assistance should be directed toward that end.

A central principle of my proposal is that the functions of the IMF and the World Bank need to be clearly separated. The IMF should focus on liquidity protection for member countries that have achieved sound financial liberalization. The efficacy of that protection would be much enhanced by focusing on achieving that narrowly defined economic objective.

The World Bank would facilitate liberalization and hence help to expand the IMF’s membership list. Encouraging bona fide liberalization is a long-term process. The form and pace of assistance required is different from that of emergency liquidity assistance, and it is very counterproductive to confuse the two missions and the two kinds of assistance.

Possibly the greatest obstacle to my proposal will be the likely opposition of the U.S. Treasury to repealing the Exchange Stabilization Fund and focusing the IMF on providing bona fide liquidity assistance. The Treasury has used the Exchange Stabilization Fund and the IMF as slush funds to provide foreign aid (without the inconvenience of seeking congressional approval) in the guise of "liquidity" assistance. Getting the U.S. Treasury to forswear such activities is a formidable challenge—one that may require veto-proof support in Congress.

If all these challenges to reforming the IMF could be overcome, and something like this proposal were enacted, would the IMF abide by the new rules? Obviously, the goal of my plan has been to design rules that are transparent to outsiders and would make it harder for the IMF to "forbear" in enforcing those rules. The more we all talk about ways to further limit such forbearance, the better.

In summary, I think it is economically feasible to restructure the way the IMF does business to promote a more efficient and democratic financia l system—one that ensures market discipline while avoiding market chaos. If the G7 chose such a path, other countries would follow: The rewards to participating in an open, market-oriented, and stable global financial system would be irresistible. The transition process could be facilitated if the funds currently channeled through the World Bank and other development banks could be redirected toward helping countries to qualify for membership in the newly constituted IMF.

Charles W. Calomiris is codirector of AEI’s Financial Deregulation Project, Paul M. Montrone Professor of Finance and Economics at Columbia University Business School, and a member of the International Financial Institution Advisory Commission. A version of this essay appeared in the January/February 1999 issue of The International Economy, and material is reprinted with permission. The essay summarizes the IMF reform proposal described in detail in Mr. Calomiris’s "Blueprints for a New Global Financial Architecture," available at http://www.imfsite.org/reform/calomiris.html. See also the related paper by AEI Visiting Scholar Allan H. Meltzer, "What’s Wrong with the IMF? What Would Be Better?" at http://www.imfsite.org/reform/meltzer.html.