آیا صندوق بین المللی پول کمک کننده یا اسیب رسان است؟تاثیر برنامه های صندوق بین المللی پول در احتمال و پیامدهای بحران ارزی

We empirically analyze the effect of International Monetary Fund (IMF) involvement on the risk of entering a currency crisis and, respectively, the outcome of such a crisis. Specifically, we investigate whether countries with previous IMF intervention are more likely to experience currency crises. In a second step, we analyze the IMF’s impact on a country’s decision to adjust the exchange rate, once a crisis occurs. We find that IMF involvement reduces the probability of a crisis. Once in a crisis, IMF programs significantly increase the probability that the authorities devalue the exchange rate. The amount of loans and compliance with conditionality have no impact.

When the International Monetary Fund (IMF) was created in 1945 its founders envisioned a Fund that would promote exchange stability and would help its member countries to adjust to disequilibria in their balance of payments. Despite these high goals, the IMF has come under increased scrutiny and attack in recent years (e.g., Stiglitz, 2002). Some of the most intense criticisms aim at the ineffectiveness of the Fund’s programs and conditionality to promote good policy and economic outcomes in the recipient countries (e.g., Dreher, 2006, Przeworski and Vreeland, 2000 and Vreeland, 2003). A large amount of literature has emerged that investigates how IMF programs and their implementation affect countries’ balance of payments, the current account, inflation, and economic growth rates (for recent surveys see Bird, 2007, Joyce, 2004 and Steinwand and Stone, 2009). In face of this abundance of studies, it is surprising that few of them have investigated the Fund’s performance with regard to one of its most generic purposes: the promotion of a stable international exchange rate system.1 One of the rare exceptions is Muckherjee (2006), who reports that the IMF’s stabilization programs failed to prevent currency crises in countries with a high degree of state intervention in the financial sector, but not in others. Hutchison (2003, chapter 10) reports that 28% of currency crises were associated with a contemporaneous short-term IMF program, while 18% of such programs were associated with a contemporaneous currency crisis. However, he does not provide an analysis of the causal direction of this empirical relationship. Finally, Bird and Mandilaras (2009) find some evidence that countries that have had an IMF program hold higher levels of foreign reserves.2 To the extent that foreign reserves deter speculative attacks on currencies, this result suggests that IMF involvement mitigates crisis risk. Overall, we know little about whether IMF programs increase or decrease a country’s risk of experiencing a currency crisis or how programs affect a country’s strategies to resolve such a crisis. Given the paucity of evidence, it is not surprising that we know even less about the channels by which the IMF influences crisis risk and the outcome of currency crises. In theory, the Fund can influence economic policies and outcomes by its available or disbursed money, the policy conditions it attaches to its loans and, more generally, its policy advice. An equally important, but more indirect, channel is what we call the “scapegoat-channel.” By allowing policymakers to shift the blame for unpopular policies onto the Fund and thus increasing their chances of political survival, the IMF can enhance the chances that economically sensible policies will in fact be implemented (Vreeland, 1999). Finally, the second indirect channel is the “moral-hazard” potentially associated with IMF lending, which might affect macroeconomic policies negatively (Vaubel, 1983). As IMF lending may be interpreted as income insurance against adverse shocks, this insurance cover might induce the potential recipients to lower their precautions against such damages. The overall effect of the IMF depends on the net effect of these channels. In this paper we therefore examine how IMF programs, disbursed loans, and compliance with conditionality affect the risk of currency crises and the outcome of such crises.3 Specifically, we investigate whether countries with previous IMF intervention are more or less likely to experience currency crises. In a second step, we test for the IMF’s impact on a country’s decision to adjust the exchange rate, once a crisis occurs. Even though the IMF aims to prevent currency crises in the first place, these crises have been a regular feature of the international exchange system. Once crises occur, the Fund’s goal is to limit their severity, resolve them quickly, and thus prevent them from having systemic implications. Protracted crises often result from the authorities’ attempt to delay a necessary adjustment of the exchange rate for too long. One of the most frequent pieces of advice the IMF gives to countries experiencing such crises therefore is an adjustment of the exchange rate. Even though IMF loans bolster countries’ reserves, this advice, or even conditionality, coupled with the opportunity to blame the IMF for a devaluation, should lead to an increased propensity for exchange rate adjustment caused by IMF crisis involvement. To anticipate our main results, we find that IMF involvement reduces the probability of a crisis. Once in a crisis, IMF programs significantly increase the probability that the exchange rate devalues. The next section discusses the various channels by which the IMF can influence crises; section three describes the method and data employed. Section 4 presents the empirical analysis, while Section 5 provides extensions. The final section concludes.

This paper has examined how IMF programs affect the risk of currency crises and the outcome of such crises. This is an important question, as the preservation of stability in the global financial system constitutes one of the Fund’s prime functions. To evaluate whether the Fund fulfills this function in the context of speculative pressure in international currency markets, we used panel data for 68 countries over the period 1975–2002 and investigated whether countries with previous IMF intervention are more likely to experience currency crises and how IMF programs impact a country’s decision to adjust the exchange rate once a crisis occurs. Our results suggest that the IMF—contrary to the Fund’s critics—does indeed fulfill its functions of promoting exchange rate stability and helping its members to correct macroeconomic imbalances. The existence of an IMF program significantly decreases the risk of a currency crisis and increases the likelihood that the exchange rate will be adjusted once a crisis is underway. Most interestingly, in both cases the existence of an IMF program drives this result, rather than money in terms of disbursed loans or compliance with conditionality. This suggests that the more indirect aspects of IMF programs, such as IMF advice, its function as a “seal of approval” and its ability to reduce the political costs of implementing unpopular policies might be more relevant than the amount of money the IMF places at countries’ disposal or countries’ compliance with IMF conditions. Future research should concentrate on further disentangling the effects of these different channels and investigating more deeply the mechanisms by which different IMF programs affect policy outcomes. Clearly, better proxies for the different channels are needed. While we can accurately measure the amount of IMF money received, an alternative interpretation of our results regarding compliance might arguably be that the proxy employed here is too crude to lead to significant results. With this grain of salt, our finding has implications for the design of conditionality. Whether or not the IMF should impose conditions on sovereign countries has been highly debated upon from the very beginning of the IMF’s operations. The empirical results of this paper have shown that compliance with conditionality does not have a statistically significant effect on currency crisis risk or the government’s decision to devalue its currency. This finding complements other studies, which have shown that the Funds’ conditions do not (or only marginally) affect economic policies and outcomes (Dreher, 2005 and Dreher, 2006). One interpretation of these results is that conditions imposed by outside factors can be circumvented, even if the officially agreed criteria have been met. In order to lend more effectively, it would therefore be most important for the IMF to detect factors influencing ownership and thus the willingness to reform. Arguably, if the IMF would support reform-minded governments, its loans might make a difference (even if its advice might not) by helping governments to implement these reforms against political opposition (by acting as a scapegoat for unpopular policies) and by giving a “seal of approval” to these governments. The results also allow a different interpretation, however. According to the IMF, conditions are the outcome of a bargaining process between the government and the Fund. They might therefore reflect the government’s agenda instead of being imposed by the IMF. As a consequence, compliance with conditionality does not make a difference with respect to economic policies, because the same policies would have been implemented without the Fund’s conditions. Whatever be the underlying causal mechanism, conditionality would not be necessary. In terms of policy advice, our results therefore suggest that the IMF’s surveillance and technical assistance might be more important than its lending and conditionality. Placing greater emphasis on the former might thus well be worthwhile.36 To some extent this is in line with the route recently chosen, in particular with the creation of the Flexible Credit Line in March 2009.