The International Monetary Fund (IMF) is a financial institution founded in 1944 with the main purpose of assisting members facing temporary balance of payments problems. From an initial membership of 44 states, today almost all the countries in the world participate in it. Members of the IMF do not have equal power. They contribute a quota subscription of financial resources, and this quota is the basis for determining voting power. Historically, quota allocations have been based mainly on economic size and external trade volume. Heterogeneity among members in terms of population, wealth and integration to international markets has recently produced a debate about the methods used for quota determination. In response to this, and also in the face of mounting criticism from academics and policymakers, the IMF embarked in September 2005 on a large-scale program of modernization. Salient among its objectives was governance reform, including adjusting quota shares to “reflect better the relative weight of members in the world economy”. In April 2008, a reform proposal representing a step in this direction was approved. In the past decade there have been many reform proposals that focused on different aspects of IMF governance, but none of these, and other, reform proposals has been founded on a model of expected utility maximization. With the aim of contributing to this debate, we adapt the model of Barberà and Jackson (2006) of optimal voting rules in a heterogeneous union. In the model, votes at the IMF take place over two alternatives: whether or not to bailout a member country in crisis. An optimal voting rule seeks to maximize a welfare function that takes into account the utilities of all citizens represented in the IMF. When making a decision for a particular vote, representatives at the IMF weight the benefits and costs that a bailout would have on their respective countries' citizens. Benefits are assumed to relate to trade linkages, and costs are assumed to originate from moral hazard ineficiencies affecting net factor income from abroad. When benefits outweigh costs a country will vote in favor of a bailout. To determine how a crisis abroad affects the welfare of a member country’s citizens, we use a static model of aggregate demand and have the foreign crisis triggering a real exchange rate shock to each member’s current account due to trade linkages with the crisis country. The direct effect is proportional to the size of trade with the crisis country as exports to it would decrease, and imports from it increase. Since a country can dampen the effect that a crisis abroad has on its real exchange rate through the use of international reserves, the welfare impact would be decreasing in the stock of foreign reserves held. Under the assumption that the marginal utility of income is decreasing, welfare effects are larger for poorer countries. In this context, optimal weights are proportional to a country’s volume of trade. Given participation in world trade, quotas decrease with per-capita income and with holdings of foreign reserves. The model further predicts that the IMF would be more likely to provide assistance to bigger, more open countries. Another prediction of the model is that voting thresholds should be increasing in the importance of capital flows relative to trade flows. The model has implication for the reform proposals that have been presented to improve the legitimacy of the IMF (see, for example, Cottarelli, 2005 and Rapkin and Strand, 2006). In particular under the assumptions of the model, there is no rationale for a double majority system as the “count and account” proposal of O’Neill and Peleg (2000). Further research will be aimed at developing a more sophisticated model of income and consumption determination, incorporating moral hazard, and a more detailed formulation of bailout costs, relating them to capital injections in the Fund.
