The Procedure for Increasing IMF Quotas

The IMF was created in 1944 to support world trade and economic growth by stabilizing the international financial system—a global public good.¹¹

To do so, the Fund provides assistance to countries facing balance-of-payments difficulties. When a country spends more abroad on goods and services than it receives, it incurs a current account deficit. Selling assets or borrowing can finance this shortfall and involves a private capital inflow into the country. But when private sources do not cover the current account deficit, a country's government must finance it through the sale of its official reserves of foreign currencies. The core responsibility of the IMF is to provide loans to deficit countries when they exhaust their reserves. This assistance helps countries rebuild their reserves, stabilize their currencies, and continue paying for imports, while they adjust policies and make reforms to correct the payments problems.

The Fund's approach to financial assistance has two main components—financing and conditionality—that jointly address the payments crisis and the underlying factors that contributed to it. Access to Fund assistance is conditioned on the adoption and pursuit of economic and structural policy measures negotiated by the IMF with the recipient country.¹²

This “conditionality” usually takes the form of performance criteria (for example, inflation and spending targets) and policy benchmarks (or example, tax reform and privatization), and the aim is to alleviate the economic difficulties that led to balance-of-payments problem.

The IMF's financial resources come from members' subscriptions, which are known as “quotas.” Each country's quota is calculated by a formula reflecting the relative size of its economy, using various measures of output and trade. Quotas also determine members' voting power in the organization. Each member has 250 “basic” votes, plus one additional vote for each part of its quota equal to SDR 100,000. As basic votes make up only a small fraction of total votes, control of the IMF is heavily weighted toward its larger members. To illustrate, the United States, with its quota of $54.2 billion has 371,743 votes (17.1 percent of the total), while Palau has 281 votes (0.013 percent of the total). Large members have even greater influence because important decisions are subject to special majorities. The United States, with more than 17 percent of the votes, has veto power over decisions, such as quota increases, that require 85 percent approval.¹³

These voting rules may be an endogenous solution to the global free-rider problem. By tying contributions to voting power, large nations are able to use their influence to promote policies with nation-specific benefits.¹⁴

Stone, for example, finds that the United States and other large members use IMF loans to advance their foreign policy goals in Africa.¹⁵ Such benefits create incentives for large members to contribute to organizations that produce both global public and nation-specific joint products.

In the case of the largest member, however, the foreign policy executive does not control contributions to the IMF; increases in the U.S. quota require legislative consent. This suggests that diplomatic influence may not be the only, or even the most important, nation-specific benefit driving U.S. support for the IMF.

The procedure for increasing quotas begins at the IMF and the main method is to establish an equiproportional increase for all members under a “General Review of Quotas.” General reviews are held about every five years and have produced eight large quota increases since 1946 (see Table 1). On each occasion—including those when the IMF chose not to propose any increase—a major factor affecting the outcomes has been the difficulty in obtaining authorization from Congress for an increase in the U.S. quota.

Every member country must consent to its quota increase. In the United States, this means legislative approval and appropriations. Since the United States is predominant at the IMF, Congress commands extraordinary leverage in the process of changing quotas. According to Pauly, “Quota increases, although strongly preferred by the Fund, sometimes entail legislative affirmation within member-states. They certainly do in the United States, a reality which has complicated the life of the Fund since the beginning.”¹⁶

Boughton cites several cases of general reviews that were tailored to expedite congressional approval and notes that no general increase in quotas has taken effect without Congress consenting to the U.S. increase.¹⁷ Woods argues that a “recalcitrant” Congress increases the influence of the United States:

Congress, however, is not a single entity. It is composed of individual legislators, who hold varied positions on funding the IMF. While there are members within Congress who are obstacles to quota increases, there are also members who are allies—those who want to give the Fund more resources to stabilize world financial markets. We take a closer look at the battle that occurs in Congress because, depending on who wins, Congress can be just as much an ally as an obstacle to the IMF. Furthermore, analyzing the factors that shape member votes is important to understanding the nature of the private benefits the United States obtains from the IMF. While it is reasonable to expect the U.S. executive branch to use its voting power and informal influence at the IMF to advance foreign policy goals, when it comes to funding the IMF, we expect members of Congress to be motivated by local concerns relevant to their electoral prospects.

In summary, increases in IMF quotas require broad support within the IMF, because 85 percent of the votes are required to approve changes. With more than 17 percent of the votes, the United States is the pivotal actor. But U.S. officials cannot act independently of Congress. Congress must formally approve changes in the U.S. quota, which means that anyone seeking an increase—the president, the treasury secretary, the U.S. executive director to the IMF, other member governments—must be sensitive to diverse congressional sentiment.

Approach and Arguments

Which members of Congress will vote in favor of (or against) quota increases? Legislator positions are influenced by many factors, including partisan identity, political ideology, and expectations about the future consequences of IMF rescues (such as the moral hazard problem). We make the standard assumption that legislator behavior is self-interested and derives, at least in part, from the desire to remain in office. Thus positions on quota increases reflect how the policy affects members electorally. This means we need to know something about the interests of voters and special interest groups with respect to the IMF. To derive these interests, we ask: Who benefits and who loses from IMF policies? We look to the literature on international finance and trade to derive such distributional effects.

Our first argument is that “money-center” banks make up a key constituency for the IMF and lobby on its behalf.¹⁹

IMF financial rescues provide de facto insurance to these banks, allowing them to retain the gains from international lending while distributing losses, when they occur, to the public sector. Thus we expect campaign contributions from money-center banks to have a positive impact on the propensity of a member of Congress to vote in favor of increasing the U.S. quota. Our second argument is that members representing districts with greater proportions of net “winners” from economic globalization are more likely to favor increasing the IMF's resources. The IMF, by pursuing its mandate to protect the world economy from financial shocks, encourages globalization, which has predictable distributional consequences.

We are not the first to identify banks as an important constituency for the IMF. A radical “dependencista” version of the argument has been around since at least the 1970s, and a more orthodox variant is currently circulating.²⁰

Bhagwati speaks of a “Wall Street–Treasury complex,” in which bankers rotate in and out of executive roles, influencing IMF policy to the benefit of Wall Street.²¹ Other scholars examine the extent to which commercial banks exert a systematic bias on IMF lending.²² Gould argues that Wall Street has influence because the IMF depends on private financial flows to make its conditionality policies effective.²³ For Oatley, commercial banks have “privileged” political access to the IMF because (1) international and national regulators see them as “too big to fail,” (2) they are few in number and have large, individual stakes in lobbying the IMF, and (3) because the IMF understands that commercial bank finance is crucial to the success of their programs.²⁴

We also treat international financiers as central actors in the politics of the IMF. However, we are less concerned with the reasons why banks have influence on the details of individual IMF loan programs than with the reasons why such banks might endorse granting more resources to the Fund. A quota increase gives the IMF greater resources to support the international payments system. It does not dictate how these resources are used or to whom they are allocated. Hence we emphasize a broader reason why money-center banks endorse quota increases: they benefit from the moral hazard created by Fund financial rescues.

Even if intended to stabilize the international financial system, IMF rescues are a form of insurance for private creditors and thus a source of “moral hazard.”²⁵

A moral hazard is an action that encourages the very behavior that it seeks to prevent. With respect to the Fund, moral hazard arises when its crisis assistance encourages banks to take on risks that they might otherwise spurn, in an attempt to reap greater financial returns. Creditors may overlend to emerging economies because of the expectation that the IMF will provide the liquidity that will allow them to exit the country in time of crisis, without bearing their full losses.

While there is a vigorous ongoing debate on the extent of the moral hazard problem, there is agreement that it exists.²⁶

IMF Deputy Managing Director Anne Krueger sees it as a major concern: “Private institutions may be encouraged to lend and invest recklessly by the belief that the Fund will ensure that their creditors can repay them.”²⁷ The International Financial Institutions Advisory Commission (the “Meltzer Commission”) also emphasized the problem.²⁸ Rogoff holds the pragmatic view that some moral hazard is an inevitable consequence of stabilizing international finance, a view reflected in policy circles, as well.²⁹ According to Thomas Dawson, former executive director for the United States at the IMF, “The problem of moral hazard is nobody has figured how you save the system without bailing out at least some investors.”³⁰

Our argument is that money-center banks are among the most direct private beneficiaries of IMF-created moral hazard. Bird finds that IMF financial assistance to debtor countries is often used to repay loans to commercial banks.³¹

In some instances, debt service is an explicit component of IMF programs.³² More generally, unanticipated increases in U.S. financial commitments to the IMF are associated with increases in the stock market capitalization of the exposed money-center banks.³³ To state this in terms of the joint products model, moral hazard is the complementary output of IMF financial rescues—the private benefit that induces money center banks to support funding the IMF. As a concentrated industry, international banks should have little problem organizing for political purposes. We therefore expect international banks to lobby Congress to expand the resources of the IMF.

Other constituencies in the United States are affected by IMF policies, but the effects are more diffuse than with banks. Our insight is that IMF rescues affect district constituencies indirectly, by way of their effect on the openness of international trade and capital markets. The primary rationale for the IMF is to preserve the openness of the world economy.³⁴

Rescues are thus the means to an end—maintenance of an economically integrated world economy—and the end is what drives constituency preferences on rescues. The inference is that members of Congress oppose (support) financial rescues because their constituents are harmed (gain) by economic globalization. International trade theory provides the basis for specific predictions.

Stolper and Samuelson, and Mundell identified the winners and losers from economic globalization in terms of factors of production from which factor owners, like high-skilled and low-skilled labor, derive their incomes.³⁵

Owners of locally abundant factors tend to gain more than average from globalization, while owners of scarce factors tend to lose. In the United States, the relatively scarce factor is low-skilled labor, and thus the group most likely to lose from globalization is low-skilled labor.³⁶ As trade has increased with nations where low-skilled labor is relatively abundant, organized labor in the United States has mobilized against globalization and received protection in less-skilled intensive industries in return.³⁷ By contrast, highly skilled labor is abundant in the United States relative to the rest of the world and thereby benefits from globalization. Indeed, individual-level data from public opinion surveys provides consistent, strong evidence: workers with college degrees or high skills residing in nations abundant in skills support further liberalization of trade and factor flows, while those with less education and fewer skills resist such initiatives.³⁸

Our extension of trade theory to IMF funding recognizes that the IMF's mandate to protect the world economy from financial disorder confers individual benefits to people who gain income from globalization. We thus expect voters with high (low) skills to support (oppose) the IMF. As diffuse interests, we do not expect lobbying from these actors, but we do expect their interests to find expression in Congress by way of the electoral calculations of legislators. Legislators calculate the distributional effects of a policy on voting constituencies within their districts and take positions that reflect these districts interests.³⁹

These calculations occur even in the absence of direct influence, meaning that diffuse groups of high- and low-skilled workers do not actually have to organize for this mechanism to be effective.

From a joint products perspective, our arguments imply that the IMF's stabilization of world financial markets creates private benefits for two sets of actors within the United States: money-center banks and high-skilled citizens. The benefit to banks is subsidized risk insurance; the benefit to the high-skilled is increased real income. The interests of these actors are reflected in the voting behavior of members of Congress via the electoral connection and induce some members to support funding the IMF. In this way, joint production of private and public goods reduces the usual problem of underprovision of public goods.

Data and Analysis: Congressional Votes on IMF Quota Increases

The IMF conducts a general review of the adequacy of quota resources at least once every five years. If it determines that a quota increase is needed, the U.S. Congress must first ratify the U.S. increase. Roll-call voting provides an opportunity to test our arguments.

We analyze congressional votes in 1983 and 1998 on quota increases. These are the only quota increases for which “clean” roll-call votes were taken. When Congress considers a quota increase, it usually does so by including the issue in an omnibus spending bill, which makes it impossible to isolate legislator positions' on the IMF issue. However, we identified three amendments and one motion to the 1983 and 1998 spending bills that dealt exclusively with IMF quotas. Table 2 summarizes these votes.

These are “clean” votes in the sense that a vote for or against captures only a member's position on increasing U.S. contributions to the IMF. Even though these votes took place during the Latin American debt crisis and the Asian currency crisis, we are not overly concerned with bias. On one hand, the crisis context might work against our arguments if members tend to ignore antiglobalization constituents during a crisis, for fear of being held responsible if a global economic meltdown follows a “no” vote on funding the IMF. On the other hand, it may work in favor of our claims if banks and proglobalization workers lobby harder during periods of global instability. Fortunately, the data we use for our variables of interest make the sampling issue moot, as they come from noncrisis periods. We measure district skill levels with census data from 1980 and 1990, well before the onset of either crisis. With respect to bank campaign contributions, we lag the data one legislative cycle (two years), which places our measure of lobbying pressure well ahead of each crisis.

Three of the votes (V286, V287, and V313) occurred in 1983 following the IMF's eighth General Review. The debt crisis provoked worries among some members of Congress that the quota increase would fund a bailout of commercial banks, and these members were reluctant to approve the quota increase without also tightening regulatory control over banks.⁴⁰

This Congress did with the International Lending Supervision Act of 1983. This act was conjoined in a single bill (H.R. 2957) which, in addition to funding the IMF, also extended the authority of the Export-Import Bank, encouraged worldwide economic growth, and provided for continued U.S. participation in multilateral development banks. Just before this omnibus bill passed the House by a close vote of 217 to 211, three members proposed amendments that would strip the bill of the IMF quota increase. We analyze votes on these amendments.

The fourth vote (V109) came in 1998 during the Asian crisis and involved a motion to an emergency supplemental spending bill (H.R. 3579). What prompted the motion was that the House and the Senate were considering two different versions of the same bill. The Senate version included funding for the U.S. peacekeeping missions in Bosnia and the Middle East, disaster relief for storm victims in the United States, as well as $18 billion for an IMF quota increase and funding the establishment of the IMF's New Arrangements to Borrow (NAB). However, the House separated these funding requests into two bills: H.R. 3579 included funding for Bosnia, the Middle East, and disaster relief; while H.R. 3580 funded $18 billon for the IMF/NAB and provided $500 million to pay down U.S. arrears to the UN.

With the House bill diverging from the Senate's, IMF funding was under threat. Procedure requires that bills pass both Houses in identical form. In an attempt to reconcile the bills, David Obey (D-Wis.), ranking member of the House Appropriations Committee, offered a motion to instruct the conference committee to put the IMF's $18 billion back in the emergency bill. On 23 April, Congress defeated Obey's motion by a vote of 186 to 222, stalling the appropriation of funds for the IMF and the NAB for another six months. The spread of the crisis to Russia and Brazil, along with President Bill Clinton's admonishment of congressional foot-dragging as “irresponsible,” finally helped convince opponents that they would be held responsible if a global recession were to take place.⁴¹

We have two hypotheses. First, we expect the probability of a vote in favor of funding the IMF to increase with a member's affinity with money-center banks. This affinity is proxied by the amount of campaign contributions each member receives from money-center banks. Second, we expect variation in skill levels across House districts to affect member voting. Specifically, we anticipate that the higher the skill level of constituents, the more likely a member will be to vote for the IMF quota increase. This illustrates our argument that members see the IMF as an organization that promotes global economic integration and take positions that reflect the impact of globalization on their constituencies.

To identify money-center banks, we use the regulatory classification in the U.S. Federal Financial Institutions Examination Council's (FFIEC) Country Exposure Lending Survey. The FFIEC identifies the specific banks that make up the “money-center” group the list on which to base our collection of campaign contribution data.⁴²

For campaign contributions, we use the Federal Election Commission's data on contributions from political action committees (PACs). Our constructed variable is bank_pac—the sum total of money-center bank contributions to each House member as a percentage of that member's total receipts in the previous electoral cycle.⁴³

We measure constituent skill levels in two ways: by educational attainment and by occupational classification. The variable college is the share of district population with four years of college. skills is the percentage of district workers in executive, administrative, managerial, professional, and professional specialty occupations.⁴⁴

In all our models, we control for member “ideology” and party affiliation. Our proxy for ideology is the first dimension of the dw-nominate score.⁴⁵

Higher values denote a more “conservative” ideology and we expect a negative sign: more economically conservative members should oppose increasing the quota because they see the IMF as an unnecessary and remote bureaucracy whose interventions in the market create moral hazard.⁴⁶ We also control for party, because political party affiliation may color the way members respond to presidential appeals for funding the IMF.

Table 3 presents results of Probit analyses of the three 1983 votes. While we find strong ideological and partisan effects, our variables of interest—bank_pac and college—are correctly signed and highly significant. For all three votes (Models 1 to 3), the more campaign contributions from international banks and the higher the education level in a district, the more likely a member is to vote in favor of an IMF quota increase. In Model 4, we include controls for district income (median household income) and mexican origins (share of district population of Mexican ancestry). The latter is intended to capture any effect that proximity to Mexico—the first victim of the debt crisis—might have on member voting. Our core results are not affected by the inclusion of these controls.

Table 4 contains results after substituting the variable skills (share of district population working in high-skilled industries) for college attainment. Our findings are robust to this alternative specification.

The vote on Obey's 1998 motion (V109) should be difficult for our argument because House members voted strongly along party lines. This should not come as a surprise because the division between the parties was wider during the 106th Congress than at any time since before World War I.⁴⁷

With the Clinton administration standing behind the IMF quota increase, all but twenty-two Republican members opposed the measure while only twenty-eight Democrats took positions against the president. Yet, despite the strong partisan character of this vote, our main variables are signed correctly and significant in several alternative models, as shown in Table 5. Model 1 includes our two variables of interest. Model 2 substitutes skills for college, and Model 3 controls for other potentially relevant district characteristics. mexican+korean+thai is the share of district population of ethnic groups originally from three countries that suffered major currency crisis in the 1990s. Our estimates do not support a relationship. We control for the possibility that the Ricardo-Viner “specific-factors” model predicts voting on IMF funding with net imports and net exports, which measure district industrial characteristics. Since the IMF pursues a protrade mandate, members representing districts that face strong import competition should oppose funding the IMF, while members with export-oriented industries in their districts should support it. These results are only suggestive; the coefficients are correctly signed but not significant.

In Table 6, we provide a substantive interpretation of these findings. Using models from Tables 3 and 5, we simulated the predicted probability of observing a vote in favor of an IMF quota increase for both Democrats and Republicans, then we examined how these probabilities change as each explanatory variable is increased by one standard deviation above its mean.⁴⁸

The effects are substantively large. For example, increasing college (share of district population with four years of college) by one standard deviation increases the probability a Democrat will support IMF funding by as much as 14 percentage points (V313, Table 3, Model 4). Although the effect is evident for members of both parties, Democrats are about twice as sensitive to district skill levels as Republicans. Similarly, increasing bank-pac (campaign contributions from international banks) by one standard deviation increases the probability that a Democrat will support the IMF by 12 percentage points on average, but the same change in bank contributions to a Republican yields a 6 percentage point increase in the likelihood of voting in favor of funding the IMF. These partisan differences in the responsiveness to district skill levels and campaign contributions from money center banks probably reflect the fact that Democrats have had strong ties to antiglobalization trade unions since the 1970s and a traditional distrust of Wall Street that extends back to the Populist era. This would suggest that increases in proglobalization workers and campaign contributions from banks would have a greater impact on Democrats than on Republicans.

IMF Lending Patterns

We established a relationship between money-center bank contributions and congressional voting on IMF funding. In this section, we check to see if money-center bank interests are reflected in IMF policy decisions. Our findings suggest that IMF lending decisions are correlated with the size of outstanding U.S. commercial bank loans in IMF member countries.

Two questions about IMF behavior motivate this analysis. First, does the extent of commercial bank exposure make the IMF more likely to bail out a country should it face a currency or debt crisis? Second, does greater commercial bank exposure induce the IMF to provide larger loans to a country, all else equal? In order to examine these questions, we adopt a two-stage approach. In the first stage, we look solely at the decision by the IMF to offer assistance; in the second, we examine the amount of assistance approved by the IMF.⁶⁰

Our data covers the period 1983–2002, during which the IMF approved 369 loans under the Stand-By and Extended Fund Facilities (EFF) programs. In the first stage of our analysis, the dependent variable is binary, representing whether or not a member country received an IMF loan in a given year. In the second stage, we analyze the size of IMF loans approved for member countries.

As our prior analysis focused on the ties between money-center banks and Congress, the chief explanatory variable for this part of the analysis is the country exposures of these same banks abroad. Individual banks do not disclose the geographic profile of their foreign loans, but the FFIEC collects, aggregates, and publishes this information for the group of money-center banks. Thus our key independent variable is us_banks, the amount owed to U.S. money-center banks by each IMF member country.⁶¹

Our argument does not lead us to predict that increased bank lending will necessarily cause a country to require IMF assistance, but rather that of those countries experiencing debt or currency crises in a given year, the IMF will be more likely to provide assistance to those with larger debts to U.S. banks. Consequently, our model includes the principal variables used to predict and identify sovereign debt and currency crises. Economists at the IMF and elsewhere have modeled currency and debt crises to establish an Early-Warning-System (EWS) that can be used by the Fund in its surveillance of the world's economies. A recent review of EWS models identified the economic indicators that yield the best predictive power.⁶²

We draw on this review and include several indicators related to countries' overall debt, debt profile, international reserves, and overall economy. We also include an indicator of financial crises (economic crisis) from another study.⁶³ Finally, because instability may persist beyond the duration of the Fund's assistance, we also include a dummy variable prior imf loans for countries that have received any Stand-By or EFF loans over the previous decade.

U.S. foreign policy goals may also affect IMF decision making. We follow standard practice and include the variable un_affinity, UN voting affinity scores for debtor countries vis-à-vis the United States.⁶⁴

Similarly, we include loans from the World Bank and International Development Association (ibrd loans), on the grounds that the IMF might be more willing to lend to countries that are receiving development assistance from the World Bank. A set of additional controls round out the model: year dummies, a time trend, and dummy variables for regions and economic groupings.