TY - JOUR
AU - Kern,, Andreas
AB - Abstract Do international lenders of last resort create financial instability by generating moral hazard? The evidence is thin and plagued with measurement error. We use the number of American troops hosted by third countries to measure the strength of American commitment to ensuring the countries’ economic health. We test several hypotheses against a dataset covering about sixty-eight countries between 1960 and 2009. Using evidence from fixed-effects and instrumental-variable models, we find that increasing the number of US troops by one standard deviation above the mean raises the probability of a financial crisis in the host country by up to 13 percentage points. We also investigate the channels through which moral hazard materializes. Countries with more US troops conduct more expansionary fiscal and monetary policies, implement riskier financial regulations, and receive more capital, especially from US banks. While many scholars of international relations view the American overseas military presence as a source of stability, we identify an underexplored mechanism by which it generates instability. Introduction Why do financial crises occur over and over again? Scholarship often points to weak financial regulations and reckless spending, both public and private (Reinhart and Rogoff 2009, chaps. 2 and 17; Schularick and Taylor 2012). At times, governments may even engage in risky policies as a way of increasing their chances of staying in power (Rajan 2010, 21; Steinberg, Koesel, and Thompson 2013). But this raises a puzzle. Why do potential victims, such as domestic and international investors, continue feeding a system that seemingly produces recurrent and inevitable crises? One would expect to see markets ceasing to provide capital to bad borrowers until those borrowers credibly implement the reforms needed to safeguard investments (Tomz 2007, chap. 1; Stasavage 2011, chap. 1). In practice, we know that this rarely happens (Reinhart and Rogoff 2009). Many observers of international finance place the blame on moral hazard. A country's ability to draw on an international lender of last resort can encourage its government to adopt ex ante risky expansionary policies that undermine financial stability. Investors, knowing that they can rely on this lender to protect their investments, are also more likely to indulge such a country's demand for capital. This results in a higher probability of financial crises. To illustrate this mechanism, commentators often point at the American bailout of Mexico in 1995 and argue that it set into motion a “new era of . . . risk taking” (Lane and Phillips 2001, 1), which, a few years later, contributed to financial crises in East Asia, Russia, and Brazil (Lane and Phillips 2000; Dreher 2004, 1). More recently, the controversial bailout of Greece triggered a spirited debate on the risks of moral hazard (Lauricella 2010; Dixon 2011; Davidson 2015; Smale and Yardley 2015). While this reasoning is intuitive, in practice it proves difficult to identify the effects of an international lender of last resort. Indeed, scholarly findings remain inconclusive (Vaubel 1983; Calomiris 2000; Dreher 2004; Jeanne and Zettelmeyer 2005; Reynaud and Vauday 2008; Dreher, Sturm, and Vreeland 2014). One reason, we argue, is that this literature focuses on the International Monetary Fund (IMF). It proves extremely hard to quantify ex ante IMF bailout expectations. Thus, measurement error threatens the accurate identification of moral hazard. Moreover, countries such as the United States regularly bypass the IMF altogether when providing support to ailing countries. Therefore, IMF bailout data may omit important cases of bi- or multilateral bailouts. We address these concerns by studying the strategies of the IMF's most influential donor: the United States (Oatley and Yackee 2004; Copelovitch 2010; McDowell 2012; Dreher et al. 2014). We ask whether American support encourages governments and markets to act recklessly. At any given time, the United States has a strategic interest in the stability of a number of countries (McManus and Yarhi-Milo 2017). Foreign governments and international investors who recognize this interest may take advantage of the situation out of an expectation that Washington will bail them out in the event of a crisis. Extensive scholarship demonstrates that countries that the United States considers strategically important are more likely to receive loans and bailouts from the IMF (Thacker 1999; Oatley and Yackee 2004; Copelovitch 2010) or directly from the United States (Kahler 1985; McDowell 2012; Schneider and Tobin 2013; Henke 2018). We predict that costly signals of US support to other countries increases their odds of suffering from financial instability. We use US troop presence as a measure of a credible commitment by the American government to third countries. The deployment of American troops is among the most credible signals that Washington can send to show its commitment to an ally or client (Reynaud and Vauday 2008; Kane 2012; Machain and Morgan 2013). It is both politically costly and highly visible. From a research perspective, troop deployments prove relatively easy to quantify.1 We find that host countries are 5 percentage points more likely to experience a financial crisis when the number of American troops on their soil doubles. Raising the number of US troops from the sample mean to one within-country standard deviation above it increases the likelihood of a financial crisis by up to 13 percentage points. When looking solely at the onset of crises, we find that the same change in troops increases the probability of a crisis by about 3 percentage points. The results are robust to country and year fixed effects; they are not affected by a range of potential confounders. To rule out spurious correlations, we complete two additional tests. First, we take advantage of the timing of the conflicts in Afghanistan and Iraq to use an instrumental variable (IV) approach. After the first Gulf War in 1991, the United States deployed troops along its medical evacuation chain between Iraq and Germany to transport severely injured soldiers to its European military medical center. The increase in the relative importance of the countries along the corridor after the conflicts started is exogenous and unlikely to have unmodeled effects on financial stability. We find that our results are robust: US troops still have a negative effect on financial stability. Second, we study the mechanisms that underpin moral hazard. We predict that hosting US troops leads to the adoption of riskier domestic economic policies, a form of debtor moral hazard. We also expect capital markets to react positively to their presence, constituting creditor moral hazard. On the domestic policy front, we show that countries adopt more expansionary fiscal and monetary policies as they host more troops. Furthermore, they are more likely to implement a deposit insurance scheme, a policy that transfers liability from the private to the public sector. On the private market front, we show that American troops are accompanied by more private debt, stronger exchange rates, higher levels of foreign liabilities, and stronger US banking ties. Taken together, these findings help us to reject the idea that our main results are due to reverse-causality or alternative explanations. They are, however, consistent with an explanation rooted in moral hazard. This article sheds light on an important question of financial politics: why don't investors and other creditors make more demands for reforms that might prevent future crises? Recent work tends to focus on domestic political failures (Mian, Sufi, and Trebbi 2010; McCarty, Poole, and Rosenthal 2013; Walter 2013), but this does not explain why international capital fails to punish governments that implement poor financial policies. Our arguments help answer this question. Our findings also complement important current debates in the study of international political economy (Kahler 1985; Keefer 2007; Drezner and McNamara 2013). Our argument draws on the “second image reversed” literature, which holds that international shocks shape domestic political outcomes (Gourevitch 1978). Recent extensions of this paradigm include the “new interdependence” approach and theories of domestic-international interactions (Gourevitch 2013; Chaudoin, Milner, and Pang 2015; Farrell and Newman 2016). We endogenize a crucial component of the global financial architecture, namely domestic prudential regulation, and model it as a joint process of international pressures and domestic political incentives. Moreover, this article contributes to the growing scholarship that ties security to economic outcomes (Schneider and Troeger 2006; Berger et al. 2013; Göktepe and Satyanath 2013). Oatley (2015, 27) argues that US security interests initially generate booms that eventually lead to busts. We focus on the consequences of US foreign policy on other countries. In a related research program, Lake (2007, 2009, 2013) develops the notion of international hierarchy in which international relations are structured by powerful countries (see also Mattern and Zarakol 2016). We qualify his belief that “[t]he hierarchies created and nurtured by the United States over the last century have been a source of order and, in turn, peace and prosperity for both the United States and its subordinate states,” a result consistent with hegemonic stability theory (Lake 2007, 79; Kindleberger 1973, chap. 14). Instead, we show that hegemonic instability more accurately describes the American role (Snidal 1985; Eichengreen 1996; Cooley 2008; Oatley et al. 2013). Finally, in contrast to previous findings (for example, Berger et al. 2013), our results show that Washington's foreign policy interests can prove costly to the United States. By insuring foreign countries and allowing them to take excessive macrofinancial risks, the United States becomes liable for costs it would rather avoid. In the next section, we develop our theoretical argument, explain how American foreign policy interests generate moral hazard, and show why US troops are a good way to measure commitment to other countries. After briefly presenting the data, we report our headline results. We then investigate the channels linking US presence and financial instability. In the conclusion, we speculate on ways in which moral hazard can be reduced. US Troops, Bailout Expectations, and Moral Hazard Argument Our argument rests on a simple idea: American commitment to other states increases the incentives of their government and markets to take bigger risks. US support to another country has two consequences. First, it enables the latter's government to adopt riskier macroeconomic and financial policies—policies that have appealing political payoffs. For instance, fiscal policy can be used to buy support from key constituencies (Alesina, Roubini, and Cohen 1997, 15; Bueno de Mesquita et al. 2003, 147). Likewise, riskier financial policies can—up to a point—unleash capital and raise the disposable income of potential supporters (Rajan 2010, 21). The downside of these policies is a higher chance of financial instability. This is why US support is crucial: the behavior of the American government can strengthen beliefs that it will use its resources to prevent or reduce the impact of a potential financial crisis. The US government may do so itself or indirectly via international institutions such as the IMF. Take Pakistan for instance. At the outset of the 2008 financial crisis, the head of the IMF mission to Pakistan, in a private conversation with US officials, noted that “the [government of Pakistan] is relying too heavily on their geo-strategic importance and still believes that, in the end, international assistance will come to help overcome the current budget shortfalls” (Wikileaks 2011a). The governor of the State Bank of Pakistan bluntly admitted that “despite the constant discussion and public announcements of plans A, B, and C, the government has only one plan to get money” (Wikileaks 2011b). Second, US support can incentivize capital markets to channel resources toward the country in question. Typically, markets punish misbehaving states (Mosley 2003, 102). However, if investors believe that they will be bailed out by the United States in a case of hardship, then the risks that they are facing will be greatly reduced. Investors, especially from the United States, can use a range of resources provided by the US government to invest abroad. To name just a few examples, capital markets can rely on investment insurances provided by institutions such as the Ex-Im Bank or the Overseas Private Investment Corporation (OPIC). These insurances increase the risk of moral hazard on the supply side of credit. Investors relying on these institutions are explicitly insured against potential political threats. CalEnergy's expansion in Indonesia illustrates this point. After Indonesia canceled several projects in 1998, CalEnergy's CEO told Congress “I can tell you, we would have not invested in Indonesia if we'd not been able to attain OPIC insurance, because the transparency issues were clear to us” (House of Representatives 1998a, 62). US support therefore generates two kinds of moral hazard. It creates debtor moral hazard by encouraging countries to adopt risky policies. It also creates creditor moral hazard by incentivizing capital markets to invest in them. The literature, by and large, mostly focused on the latter (Dreher 2004; Dreher and Vaubel 2004; Kamin 2004; Noy 2008). Both channels can operate at the same time or independently of each other. We summarize the main mechanisms of our argument in Figure 1. This figure represents a stylized illustration of the channels through which US troop deployments affect domestic macroeconomic policy-making and the behavior of capital markets. The channels are separated for illustrative purposes—reality is more complex. There could be additional causal paths between |${Z_1}$| and |${Z_2}$|⁠, for instance. We leave for future research why some countries experience moral hazard through, say, |${Z_1}$| rather than |${Z_2}$|⁠. These complex and unknown causal paths make mediation analysis unreliable (Imai et al. 2011). However, we verify that our results are robust to correlations across these different outcomes in Appendix A10. Figure 1. View largeDownload slide theoretical pathway Note: Simplified illustration of the causal path. We report the (reduced form) main results Y1 = f(X) in Table 1 and the mechanism path Z1 = f(X) and Z2 = f(X) in Table 3 (Panel A and B, respectively). Note that there could be additional causal paths (for example between Z1 and Z2) that we do not list here. Table A6 reports the estimates of the rationality of bailout expectations Y2 = f(X). US financial support (Y2) includes support by debtors or creditors; it can be partial or complete, and it can be direct or indirect (for example through the IMF). Figure 1. View largeDownload slide theoretical pathway Note: Simplified illustration of the causal path. We report the (reduced form) main results Y1 = f(X) in Table 1 and the mechanism path Z1 = f(X) and Z2 = f(X) in Table 3 (Panel A and B, respectively). Note that there could be additional causal paths (for example between Z1 and Z2) that we do not list here. Table A6 reports the estimates of the rationality of bailout expectations Y2 = f(X). US financial support (Y2) includes support by debtors or creditors; it can be partial or complete, and it can be direct or indirect (for example through the IMF). Before we discuss how we measure US support, we must answer two questions. The first is this: why would a government engage in risky policy-making if it could compromise its survival in the future? Frankel (2005), for instance, claims that politicians in developing countries often lose their jobs after a financial crisis. To understand this decision, we need to investigate its benefits and potential costs. We noted above that aggressive macroeconomic policy yields sizable payoffs (Alesina et al. 1997; Drazen 2000). Governments regularly spend, print money, or facilitate private lending for political gain (Khwaja and Mian 2005; Laeven and Levine 2009; Ansell 2014). Several country case studies show that incumbents rely on these measures to boost investment and household consumption, especially around elections (Dinç 2005; Rajan 2010, 43). Antoniades and Calomiris (2015) argue that American voters are less likely to vote for the incumbent president when credit for mortgages tightens. Furthermore, governments only control half of the moral hazard problem. Banks have few reasons to look too closely at laxer domestic regulations if they believe that they are insured by the government. Their profits increase alongside their ability to take risks. What is crucial here is that this makes it harder for domestic governments to reduce risks. Policies to tackle the inflow of resources, such as tighter capital controls, are corrosive for output growth. Governments are known to be unwilling or unable to burst the bubble of an overheating economy (Haggard 2000; MacIntyre 2001). Brown and Dinc (2005) show that governments strategically keep failing banks alive until after elections in order not to threaten their reelection chances. In Turkey, the government was unwilling to address rising financial risks from ballooning public debt levels. In a private memo, US officials note that “this government has clearly demonstrated that it will only implement sound policies and structural reforms when it is under clear pressure from its debt structure, the markets, the IFIs or the international authorities” (Wikileaks 2011c). MacIntyre (2001) finds that disagreement between critical veto players frequently prevents timely policy interventions. Thus, even governments that want to implement the best prudential regulations are often constrained in their ability to do so. Next, consider the limited cost of financial crises. To begin with, financial mismanagement does not always lead to a full-blown crisis. Crises are probabilistic events. Thus, risks taken by the ruling government may pay off. In some cases, risky policies are followed by a soft landing. To the extent that the likelihood of crisis remains subjectively low enough, then a simple benefit-cost calculation predicts that risky policies will be adopted. Even if a crisis does occur, people do not always blame the incumbent (Hellwig 2007). In open economies, the link between a government's actions and the crisis can be tenuous. Keefer (2007, 617) argues that there exists a considerable gap between a government's prudential regulations and the onset of a financial crisis. He notes that the “incentives of all governments, elected or non elected, to pursue strict prudential regulation are therefore low.” Crespo-Tenorio, Jensen, and Rosas (2013, 1063) show that, in countries with high levels of capital mobility, a banking crisis has little effect on a government's odds of losing power. They note that “openness might shield governments from electoral backlash following a banking crisis” (Crespo-Tenorio et al. 2013, 1049). Pepinsky (2012, 135) shows that incumbents in countries that suffer from the effect of external crises are often able to avoid the blame altogether. He observes that most “crisis-affected economies have experienced neither political turnover nor regime change.” Even in the epicenter of a crisis, partisanship can reduce the political cost of a crisis. In the case of the UK after 2007, Labour voters tended to blame banks instead of their own government (Hellwig and Coffey 2011). Sometimes, people blame foreign agents such as the IMF or the European Union (EU) (Magalhaes 2014). Thus, voters do not always punish their leaders. We examined whether financial crises are predictors of electoral transitions and we find no such evidence (Table A58). Let us be clear here: we do not argue that financial crises never have consequences. The global crisis of 2007, which saw many incumbents lose their jobs, is a reminder of the electoral implications of financial breakdowns (Bartels 2013). However, even this crisis did not radically change the political game (Indriason 2014; Magalhaes 2014; Marsh and Mikhaylov 2014). Ultimately, politicians continuously take risky decisions when they believe that the expected benefits outweigh the expected costs. Financial crises are not different: governments are willing to risk a hypothetical crisis that possibly could hurt their reelection prospects if the returns are high enough. And, as we noted, even those that remain cautious might be at the mercy of the incentives created for creditors. The second question is this: why would the United States let itself become exposed to this risk? We believe that the United States does it for two reasons. First, its foreign policy interests may override concerns over moral hazard and financial stability. For instance, in the aftermath of 9/11, the Bush administration's desire to intervene in Iraq increased the need for the support of Turkey (Henke 2018). As we show in our case studies (Section A3), the United States forced the IMF to hand out generous loans to the latter. Second, US military presence is often followed by broader American investments (Biglaiser and DeRouen 2007). Thus, the United States’ economic vulnerability to the host country also increases. We document several cases in Table A1. Measurement Our main challenge is to find a good measure of US support to other countries. We gauge US commitment to another country by the number of American troops deployed in it (Meernik, Krueger, and Poe 1998; Thacker 1999; Biglaiser and DeRouen 2007). Our approach relates to Reynaud and Vauday (2008) who use US troop deployments to construct an index of geopolitical importance. Incidentally, they report a significant effect of the size of US troops on the likelihood of a country entering a loan agreement with the IMF. We believe that this approach is an improvement over alternative measures. For instance, voting patterns at the United Nations (UN) General Assembly are often used to identify countries that are close to the United States (Voeten 2000). While this measure generates valuable insights, it does not reveal much about US commitment to foreign countries. Instead, it tells us something about the commitment of foreign countries to the United States. Furthermore, assembly votes are not necessarily costly enough to be interpreted as such by markets. Using troop deployments has several advantages. It is well established that US troops can be regarded as the most credible commitment device of the US administration toward allies and important countries (Reynaud and Vauday 2008; Kane 2012).2 The military itself shares this view. Secretary of the Army Caldera and Chief of Staff Reimer noted that “[putting American soldiers on the ground] is . . . the most tangible evidence of the nation's commitment to both allies and adversaries” (Caldera and Reimer 1999, viii). For the most part, American troops are stationed in countries that are not involved in violent conflicts. They are important contributors in delivering foreign aid, technical assistance, and capacity-building in different economic sectors (Kane 2012; Biglaiser and DeRouen 2009). Troop deployments depend on broader strategic interests. Generally, these interests are rooted in national security considerations, whether it is the combat on global terrorism (for instance, Afghanistan), drug trafficking (for instance, Ecuador), a country's proximity to an ongoing war (for instance, Turkey), or a country's regional importance (for instance, Taiwan). Yet, troops are more than a security commitment. Policy makers often tie military presence with broader global interests. For instance, Admiral Prueher pointed out during a congressional hearing about the 1997 East Asian crisis that “the last thing that is very much on our plate is trying to deal with the impact of the East Asian economic crisis on security matters, and those things, as have also been pointed out, are very closely intertwined” House of Representatives 1998b, 28). Likewise, State Department official Stuart Eizenstat said the following: Our engagement at this time in the financial problems of our friends and allies will speak volumes about our capacity to mobilize their support in the future for a whole range of issues. . . . We have [one hundred thousand] troops in the Asia-Pacific area. . . . [L]eadership is not divisible. We can't lead on critical security issues or on opening markets, while at the same time thinking we can abdicate our lead in maintaining the international financial system. (quoted in House of Representatives 1998a, 22) Troops are a costly signal that the United States prefers the status quo and is willing to pay for it. Indeed, the United States often provides financial resources to countries that are important to its greater interests. For instance, in the run-up to the Iraq War in 2002, Turkey's importance in the eyes of the United States increased considerably. Simultaneously, its government engaged in risky expansionary policies. In response, US representatives offered a generous macroeconomic support package worth about |${\$}$|3.5 billion in exchange for military cooperation. A year later, a private US memo mentioned the need for “early and substantial US financial support in the form of a ‘standby’ arrangement, mentioning a figure of |${\$}$|20 billion . . . to positively influence the perception of the markets that the [United States] would ‘not let Turkey go down the drain’ and that the Turkish economy would ‘stay afloat’” (Wikileaks 2011d). This undoubtedly had an effect on Turkish policy makers: as another US memo noted, “Turkish expectations about US economic support remain high” (Wikileaks 2011e). In some cases, the United States enlisted the IMF to provide support to important allies, as in Ecuador in the 1990s. Calomiris (2000, 88) argues the following: Ecuador has been suffering a deepening financial crisis for several years. . . As yet, there is no consensus for reform in Ecuador, and there is no reason to believe that reforms will be produced by a few hundreds of millions of IMF dollars. Why in the world is the IMF sending money to Ecuador? Some observes claim that IMF aid to Ecuador is best understood as a means of sending political payola to the Ecuadorian government at a time when the United States wishes to ensure continuing use of its military bases there monitoring drug trafficking. US military presence also shapes the behavior of international investors. Gray (2009; 2013, chap. 2) shows that markets take cues from costly signals sent by international actors. Biglaiser and DeRouen (2009, 251) argue that the presence of US troops operates as a “US Seal of Approval” to attract foreign investors and identify a “follow the flag” effect of American troop deployments on capital inflows from the United States. Garner (2014) reports a positive effect of US troop deployments on international FDI inflows. He argues that this result is primarily driven by a signaling effect, ensuring foreign investors that the United States would intervene to restore stability if turmoil starts. Troop deployments come with substantial government support from OPIC's political insurance scheme and other bilateral measures, such as economic support funds (ESF) and export credit guarantees of the EX-IM Bank. The catalytic effect of US troops in attracting foreign capital allows governments and the private sector in host countries to tap into foreign capital markets and potentially expand beyond their economic means.3 US commitment creates expectations. Internal and public memos reveal a major side effect of US military presence abroad: the increased expectation of US support in case of financial hardship. For instance, another internal report on the Turkish situation indicates the following: Other banking and market analysts shared Citibank's belief that the expectation of US assistance has created a moral hazard to one degree or another and is encouraging the markets to look beyond such key issues as the sustainability of real interest rates. There was a general consensus from our interlocutors that the markets believe that the [United States] and IMF will take a more charitable view of Turkey's actions in the event of an Iraq operation. (Wikileaks 2003, emphasis added) Similar patterns have been observed in several other countries during the Asian financial crises (Cronin 1998) and more recently in Kyrgyzstan, Pakistan (Cohen and Chollet 2007), Afghanistan (Sopko 2014), and Tajikistan (Nichol 2003). And such expectations are often justified. Table A6 shows that countries with more troops are more likely to receive IMF bailouts. Given the prior discussion, we introduce several scope conditions to our theory. We believe that we are more likely to find evidence of moral hazard among emerging countries. A country needs at least some degree of financial sophistication to be tied to global capital markets. The poorest countries and those with underdeveloped financial sectors are unlikely to be able to take advantage of a US commitment. At the other end of the spectrum, moral hazard is unlikely to be detectable in countries that are highly developed and have easy access to capital. We analyze all countries first and then reduce the sample to non–Organisation for Economic Co-operation and Development (OECD) countries and to countries with a private credit-to-GDP ratio between 20 and 80 percent, a typical measure of intermediate financial development. Data and Models We built a dataset covering 1960 to 2009.4 All variables are summarized in Table A2 and introduced in A4. Our main independent variable is the log number of US troops (⁠|$+1 $|⁠) located in a given country-year. The data are compiled by the Department of Defense (Kane 2004, 2012).5 By taking the log, we follow the literature for two reasons (Biglaiser and DeRouen 2007). First, theoretically we expect the effect of troops to decrease at the margin. Second, the data are extremely right-skewed. Taking the log reduces the influence of extreme outliers.6 Because we control for population separately and focus on within-country variation (by including country fixed effects), we do not further normalize the variable.7 In A6, we discuss how we dealt with possible measurement issues. We verify that the results are not driven by embassy security personnel and that they are robust to alternative ways to measure troops. For instance, we use Lake's index of security alignment to the United States as our main independent variable and find very similar results (Lake 2007, 2009, 2013).8 Finally, note that troop deployments vary considerably across and within countries; we take advantage of this to derive estimates based on within-country variation. Removing countries with few troops (for example, fewer than ten) does not affect the results.9 We claim that the presence of American troops increases moral hazard because it provides a credible signal to markets that the United States is committed to the stability and welfare of this country. We argue that this encourages riskier expansionary macroeconomic policies. We test these hypotheses in two steps. First, we examine whether countries that have more American troops on their soil are more likely to suffer from financial distress. Second, we analyze whether these same countries are more likely to implement risky domestic policies and are more likely to receive capital. To test the first hypothesis, we draw on data on financial crises from Reinhart and Rogoff (2009). They collected data on five kinds of crises: banking crises, currency crises, inflation crises, domestic public debt crises, and external public debt crises. We pool these crises together and create a dummy variable that takes the value 1 if any crisis occurs in a given country-year and 0 otherwise.10 We cannot make clear predictions as to the kind of crisis that follows moral hazard. For instance, if the government boosts public debt, then we may expect a sovereign debt crisis. However, if the government encourages private debt, then a banking crisis might be more likely.11 The baseline likelihood of a crisis in any given year is 36 percent. To test the second hypothesis, we investigate nine channels through which moral hazard may operate. Five are related to domestic policies and four to the behavior of capital markets. On the domestic front, we expect American troops to increase public spending and public debt. Governments might be inclined to carry favor to their supporters by following an expansive fiscal policy. Both are measured in logged US dollars. We also expect troops to encourage expansionary monetary policy. The presence of troops should decrease interest rates.12 Two mechanisms lead to this prediction. First, on the lender side, investors will charge lower risk premia as they bank on a US financial rescue in case of a financial crisis. The mispricing of risk is the classic case of debtor moral hazard (Mian et al. 2010). Second, from a borrower's perspective, policy makers who anticipate a US rescue will use all tools at their disposal to stimulate the economy. Their toolkit includes expansionary monetary policy. Taking these effects together, we expect that US troops will lead to lower interest rates that, in turn, boost credit growth. We use the IFS’ Central Bank Policy Rate as the closest indicator for tight or lax monetary policy.13 Besides fiscal and monetary policies, we model financial regulations, an area that received increased attention in recent years (Wilf 2016; Guisinger and Brune 2017; Pond, 2018). Governments can induce risk-taking by loosening financial regulations. The likelihood of doing so increases when liabilities can be transferred to the United States. The first piece of regulation we examine are deposit insurance systems. Deposit insurances reduce the risk of bank runs but can create excessive risk-taking by banks (Diamond and Dybvig 1983; Demirgüç-Kunt and Detragiache 2002). Customers and creditors have little reason to scrutinize their bank if they know that their deposits are safe. This allows banks to venture into riskier investments that may be very lucrative in the short run (Dam and Koetter 2012). Deposit insurances are thus a transfer of liability from the private sector to the government. If the United States operates as a lender of last resort, the weight of this liability is again (partly) transferred from the host government to the United States. The presence of US troops therefore increases the likelihood of governments implementing a deposit insurance. The second piece of regulation is the degree of capital account openness. Governments benefit from capital influx, but in the absence of a lender of last resort they are vulnerable to sudden shifts of investors’ sentiments (Passari and Rey 2015; Rey 2016). Such volatility in international capital flows is often associated with financial crises in emerging market economies, like in Southeast Asia around 1997 (Haggard 2000; Satyanath 2005). American support may reduce these concerns and encourage local policy makers to weaken safeguards. The data come from Karcher and Steinberg (2013) and range from –2 to 2, with higher values denoting more openness. The last four outcomes are related to capital markets. We begin with private demand for capital. Households and firms should be more willing to take debt if they believe that their investments are safe. Private debt is measured in logged US dollars. On the supply side of capital, we expect US banks to increase their exposure to countries that are protected by the US government. If American lenders expect their government to help if their investments turn sour, then they may be more willing to invest in these countries. Exposure is measured in the log of the claims by US banks in the host country (measured in millions of dollars) (McDowell 2017). For similar reasons, we model the amount of total foreign liabilities of the country (logged) and expect a positive correlation with troops. Likewise, the influx of capital should strengthen the host country's exchange rate. We use the World Bank's exchange rate index, which is based on a weighted combination of a country's exchange rate with various trade partners divided by a cost index. The series is normalized to equal 100 in 2010. We estimate variants of the following model: \begin{equation*} \ \begin{array}{*{20}{r}} {{{\rm{Y}}_{i,t}}}&{ = \beta {\rm{log}}{{\left( {{\rm{US\ Troops}} + 1} \right)}_{i,t - 1}} + \gamma '{{{\bf C}}_{i,t - 1}} + {\phi _i} + {\psi _t} + {\varepsilon _{i,t}},} \end{array}\end{equation*} where i denotes a country and t a year, Y is one of the outcomes listed above, |${{\bf C}}$| is a vector of control variables, |$\phi $| are country fixed effects, |$\psi $| are year fixed effects, and |${\varepsilon _{}}$| is the error term.14 We lag all independent variables by one year in the results reported here. We discuss other lag structures, especially for US troops, in Appendix A6. We expect |$\beta $|⁠, the key parameter, to be positive when we model financial crises. Standard errors are clustered by country.15 Estimating fixed effects in nonlinear model can run into the incidental parameter problem (Greene 2003). To avoid bias, we estimate the parameters with least squares. We also report the estimates from a logit model with country fixed effects. Table A25 replicates all models with logit fixed effects and A26 with conditional logit, with very little consequences for our results (Beck 2015). Given the observational nature of the data, we rely on well-specified models to obtain credible estimates. First, by using country fixed effects, we focus on within-country variation to eliminate unobserved country-specific effects. This matters because it controls for factors such as geographic proximity to the United States, which could be correlated with troops and the main outcomes. An increased presence of US troops within a country likely indicates an increase of the importance of this country in the eyes of the American government.16 Second, year fixed effects capture global macroeconomic shocks. Including them also captures shifts in American policy-making and the general tendency of the United States to station troops abroad (for instance during the Cold War). Admittedly, this is a crude approach given that international macroeconomic forces shape domestic outcomes in complicated ways (Caballero, Farhi, and Gourinchas 2006; Forbes and Warnock 2012). Still, since we want to study one particular source of international instability—moral hazard—we are making this simplifying assumption. Third, we control for a number of confounding factors. The socioeconomic variables that we control for are the log of gross domestic product (GDP) per capita, GDP growth, logged population, and the size of the service sector. Output and growth are key determinants of financial crises, since they capture variation in aggregate demand. Population helps us control for the relative size of the US contingent compared to the host country. Besides these variables, our results may also be driven by a country's similarity to the United States. We control for this effect in two ways. Economically, the size of the service sector proxies the development level of the local economy. Politically, we use a measure of distance between the political preferences of a given country and those of the United States. In line with previous research, we use the ideal point estimates from UN General Assembly voting (Voeten 2000; Bailey, Strezhnev, and Voeten 2015). Finally, we control for a country's regime type (Cheibub, Gandhi, and Vreeland 2010). Results Risk of Crisis Our main hypothesis states that the risk of a financial crisis increases in the presence of US troops on a country's soil. The results are reported in Table 1. Models 1 to 3 use the entire sample. Models 4 to 9 only include non-OECD countries, a sample that we believe is more appropriate. Models 10 to 12 use data from countries that have intermediate financial systems. Each model differs with respect to the inclusion (or not) of control variables, country fixed effects, and year fixed effects. Table 1. Likelihood of a financial crisis in a country-year depending on the number of US troops (log) stationed in a country Explaining financial crises All countries Non-OECD Credit 20–80% of GDP (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) FE FE FE FE FE FE FE FE Logit FE FE FE US troops (log) 0.031** 0.073 *** 0.043 * 0.070 ** 0.044*** 0.050*** 0.074 *** 0.054 *** 0.534*** 0.048 0.091 *** 0.065 *** (0.014) (0.023) (0.022) (0.027) (0.015) (0.016) (0.025) (0.019) (0.116) (0.031) (0.018) (0.020) GDP/Cap (log) –0.127 –0.173* –0.242*** 0.049 –0.054 0.000 –0.149 –0.039 (0.096) (0.091) (0.078) (0.108) (0.092) (0.473) (0.240) (0.234) Population (log) 0.260 0.179 1.415*** 0.697 *** 7.657 *** 0.186 0.524 * (0.190) (0.183) (0.257) (0.195) (1.916) (0.342) (0.303) GDP growth (%) –0.020 *** –0.016 *** –0.015 *** –0.011 *** –0.120 *** –0.023 *** –0.017 ** (0.003) (0.004) (0.004) (0.004) (0.026) (0.006) (0.007) Service sector (% of GDP) –0.001 –0.002 –0.005 –0.005 –0.089 *** –0.003 –0.003 (0.004) (0.004) (0.004) (0.005) (0.025) (0.006) (0.006) Democracy –0.095 –0.111 * –0.024 –0.051 –0.230 0.028 0.017 (0.082) (0.060) (0.080) (0.062) (0.358) (0.092) (0.083) Distance to US ideal point 0.135 * 0.017 0.216 *** 0.032 0.423 0.253*** 0.066 (0.073) (0.065) (0.059) (0.071) (0.395) (0.090) (0.117) Country FE ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Time trend ✓ ✓ ✓ Year FE ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Observations 2904 1942 1942 1585 1585 1463 1117 1117 1080 1194 893 893 R2 0.18 0.11 0.23 0.03 0.31 0.33 0.23 0.36 0.20 0.16 0.25 # Countries 68 63 63 41 41 40 39 39 57 52 52 Explaining financial crises All countries Non-OECD Credit 20–80% of GDP (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) FE FE FE FE FE FE FE FE Logit FE FE FE US troops (log) 0.031** 0.073 *** 0.043 * 0.070 ** 0.044*** 0.050*** 0.074 *** 0.054 *** 0.534*** 0.048 0.091 *** 0.065 *** (0.014) (0.023) (0.022) (0.027) (0.015) (0.016) (0.025) (0.019) (0.116) (0.031) (0.018) (0.020) GDP/Cap (log) –0.127 –0.173* –0.242*** 0.049 –0.054 0.000 –0.149 –0.039 (0.096) (0.091) (0.078) (0.108) (0.092) (0.473) (0.240) (0.234) Population (log) 0.260 0.179 1.415*** 0.697 *** 7.657 *** 0.186 0.524 * (0.190) (0.183) (0.257) (0.195) (1.916) (0.342) (0.303) GDP growth (%) –0.020 *** –0.016 *** –0.015 *** –0.011 *** –0.120 *** –0.023 *** –0.017 ** (0.003) (0.004) (0.004) (0.004) (0.026) (0.006) (0.007) Service sector (% of GDP) –0.001 –0.002 –0.005 –0.005 –0.089 *** –0.003 –0.003 (0.004) (0.004) (0.004) (0.005) (0.025) (0.006) (0.006) Democracy –0.095 –0.111 * –0.024 –0.051 –0.230 0.028 0.017 (0.082) (0.060) (0.080) (0.062) (0.358) (0.092) (0.083) Distance to US ideal point 0.135 * 0.017 0.216 *** 0.032 0.423 0.253*** 0.066 (0.073) (0.065) (0.059) (0.071) (0.395) (0.090) (0.117) Country FE ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Time trend ✓ ✓ ✓ Year FE ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Observations 2904 1942 1942 1585 1585 1463 1117 1117 1080 1194 893 893 R2 0.18 0.11 0.23 0.03 0.31 0.33 0.23 0.36 0.20 0.16 0.25 # Countries 68 63 63 41 41 40 39 39 57 52 52 Notes: (1) All explanatory variables are lagged by one year. (2) Credit 20–80 percent of GDP means that the sample only contains country-years that have a private credit to GDP ratio that is between 20 and 80 percent. (3) Within-unit R2 reported. (4) Statistical significance: *p < 0.1; **p < 0.05; ***p < 0.01. View Large Table 1. Likelihood of a financial crisis in a country-year depending on the number of US troops (log) stationed in a country Explaining financial crises All countries Non-OECD Credit 20–80% of GDP (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) FE FE FE FE FE FE FE FE Logit FE FE FE US troops (log) 0.031** 0.073 *** 0.043 * 0.070 ** 0.044*** 0.050*** 0.074 *** 0.054 *** 0.534*** 0.048 0.091 *** 0.065 *** (0.014) (0.023) (0.022) (0.027) (0.015) (0.016) (0.025) (0.019) (0.116) (0.031) (0.018) (0.020) GDP/Cap (log) –0.127 –0.173* –0.242*** 0.049 –0.054 0.000 –0.149 –0.039 (0.096) (0.091) (0.078) (0.108) (0.092) (0.473) (0.240) (0.234) Population (log) 0.260 0.179 1.415*** 0.697 *** 7.657 *** 0.186 0.524 * (0.190) (0.183) (0.257) (0.195) (1.916) (0.342) (0.303) GDP growth (%) –0.020 *** –0.016 *** –0.015 *** –0.011 *** –0.120 *** –0.023 *** –0.017 ** (0.003) (0.004) (0.004) (0.004) (0.026) (0.006) (0.007) Service sector (% of GDP) –0.001 –0.002 –0.005 –0.005 –0.089 *** –0.003 –0.003 (0.004) (0.004) (0.004) (0.005) (0.025) (0.006) (0.006) Democracy –0.095 –0.111 * –0.024 –0.051 –0.230 0.028 0.017 (0.082) (0.060) (0.080) (0.062) (0.358) (0.092) (0.083) Distance to US ideal point 0.135 * 0.017 0.216 *** 0.032 0.423 0.253*** 0.066 (0.073) (0.065) (0.059) (0.071) (0.395) (0.090) (0.117) Country FE ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Time trend ✓ ✓ ✓ Year FE ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Observations 2904 1942 1942 1585 1585 1463 1117 1117 1080 1194 893 893 R2 0.18 0.11 0.23 0.03 0.31 0.33 0.23 0.36 0.20 0.16 0.25 # Countries 68 63 63 41 41 40 39 39 57 52 52 Explaining financial crises All countries Non-OECD Credit 20–80% of GDP (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) FE FE FE FE FE FE FE FE Logit FE FE FE US troops (log) 0.031** 0.073 *** 0.043 * 0.070 ** 0.044*** 0.050*** 0.074 *** 0.054 *** 0.534*** 0.048 0.091 *** 0.065 *** (0.014) (0.023) (0.022) (0.027) (0.015) (0.016) (0.025) (0.019) (0.116) (0.031) (0.018) (0.020) GDP/Cap (log) –0.127 –0.173* –0.242*** 0.049 –0.054 0.000 –0.149 –0.039 (0.096) (0.091) (0.078) (0.108) (0.092) (0.473) (0.240) (0.234) Population (log) 0.260 0.179 1.415*** 0.697 *** 7.657 *** 0.186 0.524 * (0.190) (0.183) (0.257) (0.195) (1.916) (0.342) (0.303) GDP growth (%) –0.020 *** –0.016 *** –0.015 *** –0.011 *** –0.120 *** –0.023 *** –0.017 ** (0.003) (0.004) (0.004) (0.004) (0.026) (0.006) (0.007) Service sector (% of GDP) –0.001 –0.002 –0.005 –0.005 –0.089 *** –0.003 –0.003 (0.004) (0.004) (0.004) (0.005) (0.025) (0.006) (0.006) Democracy –0.095 –0.111 * –0.024 –0.051 –0.230 0.028 0.017 (0.082) (0.060) (0.080) (0.062) (0.358) (0.092) (0.083) Distance to US ideal point 0.135 * 0.017 0.216 *** 0.032 0.423 0.253*** 0.066 (0.073) (0.065) (0.059) (0.071) (0.395) (0.090) (0.117) Country FE ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Time trend ✓ ✓ ✓ Year FE ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Observations 2904 1942 1942 1585 1585 1463 1117 1117 1080 1194 893 893 R2 0.18 0.11 0.23 0.03 0.31 0.33 0.23 0.36 0.20 0.16 0.25 # Countries 68 63 63 41 41 40 39 39 57 52 52 Notes: (1) All explanatory variables are lagged by one year. (2) Credit 20–80 percent of GDP means that the sample only contains country-years that have a private credit to GDP ratio that is between 20 and 80 percent. (3) Within-unit R2 reported. (4) Statistical significance: *p < 0.1; **p < 0.05; ***p < 0.01. View Large Hosting American troops increases the likelihood of experiencing a financial crisis. The effect is significant and fairly stable in all specifications, regardless of the inclusion of covariates, country, or year fixed effects. The point estimates are slightly higher when we restrict the sample to non-OECD countries and to countries with intermediate levels of financial development. This matters because these are the samples of countries for which we expect the theory to apply most forcefully. Substantively, host countries are about 5 percentage points more likely to experience a financial crisis when the number of US troops doubles. Such changes are rare on a year-to-year basis, but they happen regularly over longer periods of time. In fact, the magnitude of the effect becomes larger when we use typical changes in the number of troops. Here and below, we illustrate the effects of troops in the following manner. We report the change in outcomes if the number of troops goes from the (nonlogged) sample mean to one within-country standard deviation above the mean. The actual number varies across samples, but this generally means that we increase troops by a factor three or four. Using such a change, we find that the likelihood of a financial crisis goes up by up to 13 percentage points (depending on the model). Remember that the unconditional likelihood of experiencing a crisis in any country-year is about 36 percent. The role of moral hazard, to the extent that we capture it with US troops, is therefore quite significant. We subjected our results to a wide range of robustness tests. For sake of space, we refer the reader to Appendix A6 for a review of all robustness tests conducted. These include using various subsamples, different model specifications, and robustness tests with different dependent and independent variables. Instrumental Variable Approach Endogeneity potentially threatens our estimates. For instance, American troops could be sent to financially unstable countries—a case of reverse causality. If this were the case, the positive effect of US troops on crises would be spurious. We believe this not to be the case for three reasons. First, we examine the number of troops before and after a financial crisis. If troops were a function of financial instability, then we would expect their number to increase in the aftermath of a crisis. Figure A4 shows the opposite: the number of US troops declines in the aftermath of a financial crisis. We offer two more robust pieces of evidence for our analysis. Below, we test whether we observe patterns that are consistent with an increase in moral hazard and inconsistent with the story suggested by reverse causality. Beforehand we present results from an instrumental variable approach.17 To identify a plausible instrument, we start from the following idea: we need to find a shock that increased the relative importance of a given country in the eyes of the United States for exogenous reasons. We take advantage of the first and second Gulf War as well as the conflict in Afghanistan to identify such a shock. After the conflicts started, the United States needed a safe path to send injured troops from the conflict zone to its main military hospitals. For operations in the Middle East and Afghanistan, the most severely injured soldiers are flown to Landstuhl in Germany. The Landstuhl Medical Center (LRMC) states that “more than 90,000 Wounded Warriors from Afghanistan and Iraq have been treated at LRMC as they make their way through the medical evacuation system” (Hennessy 2016, 135). To ensure the safety of these operations, the United States needed to build new bases and station troops in countries that were geographically located along that path. Placing troops along this route was also essential for having alternative delivery options for supplies. In fact, the United States built new military bases in Bulgaria, Romania, and Hungary. Following military terminology, we refer to countries along the corridor between Germany and Afghanistan and Iraq as Echelon countries for the name of the program's official name. We rely on this information to build a dichotomous instrument that indicates if a country is located in the Echelon corridor. Countries in the corridor are coded as 1 after the start of the relevant conflict (Iraq or Afghanistan) and 0 beforehand. This concerns all countries along this path and not only those that built new bases (as this would create a different kind of selection issue). All other countries are coded as 0 throughout. We argue that the relative importance of the Echelon countries to the United States increased significantly after the first Gulf War in 1991 (Matthews and Holt 1991; Department of State 2008). The same happened to a set of central Asian countries after the Afghan war started in 2001. Providing the shortest path to Germany proved crucial to provide care to injured troops. Furthermore, the Iraqi and Afghan conflicts are plausibly exogenous to these countries. To reduce concerns about the spillover effects of the conflicts, we remove countries that are immediately neighboring Iraq and Afghanistan from our analysis. The remaining countries are too far from the war operations to be economically affected by these conflicts, at least in a different manner from the rest of the world. We have therefore good theoretical reasons to believe that (a) these countries are more likely to host troops after 1991 (or 2001 for the central Asian countries) and (b) that there is no reason to suspect that the Echelon status correlates with the main error term. The Echelon countries are shaded in the map in Figure 2. Figure 2. View largeDownload slide map Note: Map of countries included in the Echelon program (in grey) in the aftermath of war operations in Iraq and Afghanistan (in black). Countries neighboring Iraq and Afghanistan omitted in the main analysis; see test for more details. Figure 2. View largeDownload slide map Note: Map of countries included in the Echelon program (in grey) in the aftermath of war operations in Iraq and Afghanistan (in black). Countries neighboring Iraq and Afghanistan omitted in the main analysis; see test for more details. To check the plausibility of the instrument, we first regress the number of troops on two time-period dummies (1991–2001 and post-2002), which we interact with a dichotomous indicator for Echelon countries, as well as country fixed effects. The results are reported in Table A52. We see that indeed the number of troops increased in these countries.18 We implement the instrumental variable strategy with two-stage least squares: \begin{eqnarray*} {{\rm{log}}{{\left( {{\rm{US\ Troops}} + 1} \right)}_{i,t}}} &=& {\psi _i} + {\eta _t} + \delta {\rm{Echelon\ Countr}}{{\rm{y}}_{i,t}}\nonumber\\ &&+ \,\phi '{{{\bf X}}_{i,t}} + {\mu _{i,t}}\,{\left( {1{\rm{st\ stage}}} \right)}\nonumber\\[3pt] {{{\rm{Y}}_{i,t + 1}}} &=& {\alpha _i} + {\gamma _t} + \beta \widehat {{\rm{log}}{{\left( {{\rm{US\ Troops}} + 1} \right)}_{i,t}}}\nonumber\\[3pt] && +\, \lambda '{{{\bf X}}_{i,t}} + {\varepsilon _{i,t}}. \,{\left( {2{\rm{nd\ stage}}} \right)} \end{eqnarray*} The results are reported in Table 2.19 The sample for the first four models contains countries for Europe, the Middle East, and Asia—that is, countries that could plausibly have been on the path for emergency planes. These are our preferred models. The last four models use the earlier set of non-OECD countries.20 Each model differs on specification and whether a quadratic time trend or year fixed effects are included.21 Table 2. Likelihood of a financial crisis in a country-year depending on the number of US troops (log) stationed in a country Instrumental variable approach Regional subset Non-OECD (1) (2) (3) (4) (5) (6) (7) (8) IV IV IV IV IV IV IV IV Second stage US troops (log) 0.12 * 0.45 ** 0.56 * 0.13 ** 0.49 *** 0.63 ** –1.37 0.29 *** (0.07) (0.21) (0.29) (0.05) (0.12) (0.26) (2.46) (0.10) GDP/cap (log) –0.37 *** –0.71 ** –0.97 ** –0.26 ** –0.39 *** –0.15 0.39 –0.11 * (0.13) (0.33) (0.48) (0.12) (0.09) (0.11) (0.81) (0.06) Population (log) 0.12 –0.35 –0.64 0.14 –0.46 –1.27 5.63 –0.08 (0.12) (0.29) (0.49) (0.18) (0.34) (0.88) (8.63) (0.31) GDP growth (%) –0.01 * –0.01 –0.01 ** –0.01 ** –0.01 –0.01 *** (0.01) (0.01) (0.01) (0.01) (0.01) (0.00) Service sector (% of GDP) –0.02 ** –0.02 –0.01 –0.02 *** 0.03 –0.01*** (0.01) (0.01) (0.00) (0.01) (0.05) (0.00) Democracy 0.50 * 0.69 0.03 0.22* –0.85 0.09 (0.30) (0.47) (0.09) (0.12) (1.41) (0.07) Distance to US ideal point 0.32 ** 0.51 * 0.23 ** –1.18 (0.14) (0.28) (0.10) (2.04) Country FE ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Quadratic time ✓ ✓ Year FE ✓ ✓ ✓ ✓ ✓ ✓ First stage Echelon 1.02 *** 0.57 *** 0.44 *** 1.57 *** 1.49 *** 0.74 ** –0.25 1.19 *** (0.18) (0.16) (0.15) (0.24) (0.21) (0.32) (0.45) (0.30) F Statistic 31.34 12.61 8.49 42.68 51.90 5.28 0.30 15.41 Observations 1448 1039 1039 1039 1463 1117 1117 1117 # Countries 38 35 35 35 40 39 39 39 Instrumental variable approach Regional subset Non-OECD (1) (2) (3) (4) (5) (6) (7) (8) IV IV IV IV IV IV IV IV Second stage US troops (log) 0.12 * 0.45 ** 0.56 * 0.13 ** 0.49 *** 0.63 ** –1.37 0.29 *** (0.07) (0.21) (0.29) (0.05) (0.12) (0.26) (2.46) (0.10) GDP/cap (log) –0.37 *** –0.71 ** –0.97 ** –0.26 ** –0.39 *** –0.15 0.39 –0.11 * (0.13) (0.33) (0.48) (0.12) (0.09) (0.11) (0.81) (0.06) Population (log) 0.12 –0.35 –0.64 0.14 –0.46 –1.27 5.63 –0.08 (0.12) (0.29) (0.49) (0.18) (0.34) (0.88) (8.63) (0.31) GDP growth (%) –0.01 * –0.01 –0.01 ** –0.01 ** –0.01 –0.01 *** (0.01) (0.01) (0.01) (0.01) (0.01) (0.00) Service sector (% of GDP) –0.02 ** –0.02 –0.01 –0.02 *** 0.03 –0.01*** (0.01) (0.01) (0.00) (0.01) (0.05) (0.00) Democracy 0.50 * 0.69 0.03 0.22* –0.85 0.09 (0.30) (0.47) (0.09) (0.12) (1.41) (0.07) Distance to US ideal point 0.32 ** 0.51 * 0.23 ** –1.18 (0.14) (0.28) (0.10) (2.04) Country FE ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Quadratic time ✓ ✓ Year FE ✓ ✓ ✓ ✓ ✓ ✓ First stage Echelon 1.02 *** 0.57 *** 0.44 *** 1.57 *** 1.49 *** 0.74 ** –0.25 1.19 *** (0.18) (0.16) (0.15) (0.24) (0.21) (0.32) (0.45) (0.30) F Statistic 31.34 12.61 8.49 42.68 51.90 5.28 0.30 15.41 Observations 1448 1039 1039 1039 1463 1117 1117 1117 # Countries 38 35 35 35 40 39 39 39 Notes: (1) See text for details about the instrument. (2) All explanatory variables are lagged by one year. (3) Statistical significance: *p < 0.1; **p < 0.05; ***p < 0.01. View Large Table 2. Likelihood of a financial crisis in a country-year depending on the number of US troops (log) stationed in a country Instrumental variable approach Regional subset Non-OECD (1) (2) (3) (4) (5) (6) (7) (8) IV IV IV IV IV IV IV IV Second stage US troops (log) 0.12 * 0.45 ** 0.56 * 0.13 ** 0.49 *** 0.63 ** –1.37 0.29 *** (0.07) (0.21) (0.29) (0.05) (0.12) (0.26) (2.46) (0.10) GDP/cap (log) –0.37 *** –0.71 ** –0.97 ** –0.26 ** –0.39 *** –0.15 0.39 –0.11 * (0.13) (0.33) (0.48) (0.12) (0.09) (0.11) (0.81) (0.06) Population (log) 0.12 –0.35 –0.64 0.14 –0.46 –1.27 5.63 –0.08 (0.12) (0.29) (0.49) (0.18) (0.34) (0.88) (8.63) (0.31) GDP growth (%) –0.01 * –0.01 –0.01 ** –0.01 ** –0.01 –0.01 *** (0.01) (0.01) (0.01) (0.01) (0.01) (0.00) Service sector (% of GDP) –0.02 ** –0.02 –0.01 –0.02 *** 0.03 –0.01*** (0.01) (0.01) (0.00) (0.01) (0.05) (0.00) Democracy 0.50 * 0.69 0.03 0.22* –0.85 0.09 (0.30) (0.47) (0.09) (0.12) (1.41) (0.07) Distance to US ideal point 0.32 ** 0.51 * 0.23 ** –1.18 (0.14) (0.28) (0.10) (2.04) Country FE ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Quadratic time ✓ ✓ Year FE ✓ ✓ ✓ ✓ ✓ ✓ First stage Echelon 1.02 *** 0.57 *** 0.44 *** 1.57 *** 1.49 *** 0.74 ** –0.25 1.19 *** (0.18) (0.16) (0.15) (0.24) (0.21) (0.32) (0.45) (0.30) F Statistic 31.34 12.61 8.49 42.68 51.90 5.28 0.30 15.41 Observations 1448 1039 1039 1039 1463 1117 1117 1117 # Countries 38 35 35 35 40 39 39 39 Instrumental variable approach Regional subset Non-OECD (1) (2) (3) (4) (5) (6) (7) (8) IV IV IV IV IV IV IV IV Second stage US troops (log) 0.12 * 0.45 ** 0.56 * 0.13 ** 0.49 *** 0.63 ** –1.37 0.29 *** (0.07) (0.21) (0.29) (0.05) (0.12) (0.26) (2.46) (0.10) GDP/cap (log) –0.37 *** –0.71 ** –0.97 ** –0.26 ** –0.39 *** –0.15 0.39 –0.11 * (0.13) (0.33) (0.48) (0.12) (0.09) (0.11) (0.81) (0.06) Population (log) 0.12 –0.35 –0.64 0.14 –0.46 –1.27 5.63 –0.08 (0.12) (0.29) (0.49) (0.18) (0.34) (0.88) (8.63) (0.31) GDP growth (%) –0.01 * –0.01 –0.01 ** –0.01 ** –0.01 –0.01 *** (0.01) (0.01) (0.01) (0.01) (0.01) (0.00) Service sector (% of GDP) –0.02 ** –0.02 –0.01 –0.02 *** 0.03 –0.01*** (0.01) (0.01) (0.00) (0.01) (0.05) (0.00) Democracy 0.50 * 0.69 0.03 0.22* –0.85 0.09 (0.30) (0.47) (0.09) (0.12) (1.41) (0.07) Distance to US ideal point 0.32 ** 0.51 * 0.23 ** –1.18 (0.14) (0.28) (0.10) (2.04) Country FE ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Quadratic time ✓ ✓ Year FE ✓ ✓ ✓ ✓ ✓ ✓ First stage Echelon 1.02 *** 0.57 *** 0.44 *** 1.57 *** 1.49 *** 0.74 ** –0.25 1.19 *** (0.18) (0.16) (0.15) (0.24) (0.21) (0.32) (0.45) (0.30) F Statistic 31.34 12.61 8.49 42.68 51.90 5.28 0.30 15.41 Observations 1448 1039 1039 1039 1463 1117 1117 1117 # Countries 38 35 35 35 40 39 39 39 Notes: (1) See text for details about the instrument. (2) All explanatory variables are lagged by one year. (3) Statistical significance: *p < 0.1; **p < 0.05; ***p < 0.01. View Large Starting with the first stage, the point estimate of the instrument remains positive throughout (with one exception), in line with our theoretical expectation. Countries along the Echelon rescue-flight route host more American troops once the program started. The instrument's F statistic lies above 10 in all but three cases, suggesting that our main estimates are not threatened by bias due to weak instruments (Staiger and Stock 1997). The effect of Echelon status changes sign and its F statistic declines dramatically in Model 7. The instrument becomes extremely weak, which precludes us from drawing conclusion from it. The weakness of the instrument may be caused by collinearity between Echelon status and political distance to the United States. Countries that sign agreements with the United States for the Echelon program are also likely to move politically toward it. We verify this conjecture in Models 4 and 8, in which we remove the ideal point distance to the United States. Indeed, we find that the instrument becomes much stronger. Turning to the second stage, we find that the effect of US troops remains positive and statistically significant in almost all models. Troops continue to increase the likelihood of experiencing a financial crisis, with point estimates ranging from 0.2 to 1.05, all of which are statistically significant, except for Model 7. As noted, it is difficult to interpret this result because of the weakness of the instrument. We note that the estimates are larger in absolute terms than those found earlier. There could be several reasons for this. One is that troops imperfectly measure US support to other countries. This creates measurement error, which biases our ordinary least squares (OLS) estimates toward zero. In that sense, the IV estimates correct for this downward bias. Likewise, an omitted variable might be driving OLS estimation results toward zero as well. Yet another reason is that the single-equation models and the IV estimates measure different quantities. The latter estimates the local average treatment effect. Based on our scope conditions, countries in the Echelon corridor are among those most likely to be responsive to US interests. They are middle-income economies with decent (but not perfect) capital markets. As a result, they might be the most likely to be affected by positive US signals. In that case, our IV estimates represent an upper bound, in the sense that these estimates identify the effect of troops among countries that are particularly likely to react to their presence. The last reason for the difference between OLS and IV is more worrisome. The effect of the instrument may materialize through some other unmodeled third variable. The exclusion restriction assumption would thus be violated, threatening our IV estimates. The exclusion restriction assumption cannot be tested. However, Conley, Hansen, and Rossi (2010) provide a sensitivity test of the reliability of the instrument if the exclusion restriction assumption is not met. We report the results in Figure A7 and show that even large violations of the exclusion restriction would not affect our results. Channels of Moral Hazard We now leverage our theory to test the channels through which moral hazard could operate. Earlier, we made two broad conjectures. First, the presence of US troops would embolden host governments to adopt expansive fiscal, monetary, and financial policies. The results are reported in Table 3, Panel A. Second, capital markets should react favorably to US troops, as we show in Panel B. Table 3. Outcomes are listed at the top of each column Mechanisms Panel A: Policies Public spending Public debt CB policy rate Deposit insurance Capital openness (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) OLS FE OLS FE OLS FE OLS FE OLS FE US troops (log) 0.382*** 0.023*** 0.124** 0.135** –0.909*** –0.703** 0.031*** 0.018** 0.196*** 0.007 (0.015) (0.005) (0.048) (0.06) (0.116) (0.28) (0.004) (0.009 (0.012) (0.017) Control var. ✓ ✓ ✓ ✓ ✓ Country FE ✓ ✓ ✓ ✓ ✓ Year FE ✓ ✓ ✓ ✓ ✓ Observations 2819 2128 3682 2719 569 331 2910 1134 3817 2891 R2 0.18 0.82 0.00 0.19 0.10 0.33 0.02 0.65 0.07 0.14 # Countries 111 116 42 48 120 Panel B: Capital Markets Private debt Liabilities (log) US bank Exchange rate (1) (2) (3) (4) (5) (6) (7) (8) OLS FE OLS FE OLS FE OLS FE US troops (log) 0.181*** 0.040** 0.505*** 0.050*** 0.792*** 0.036 4.466 9.413** (0.024) (0.02) (0.015) (0.01) (0.026) (0.026) (24.601) (4.545) Control var. ✓ ✓ ✓ ✓ Country FE ✓ ✓ ✓ ✓ Year FE ✓ ✓ ✓ ✓ Observations 3701 2727 3723 2787 2596 1991 1594 1289 R2 0.01 0.29 0.22 0.81 0.26 0.29 0.00 0.20 # Countries 117 117 119 56 Mechanisms Panel A: Policies Public spending Public debt CB policy rate Deposit insurance Capital openness (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) OLS FE OLS FE OLS FE OLS FE OLS FE US troops (log) 0.382*** 0.023*** 0.124** 0.135** –0.909*** –0.703** 0.031*** 0.018** 0.196*** 0.007 (0.015) (0.005) (0.048) (0.06) (0.116) (0.28) (0.004) (0.009 (0.012) (0.017) Control var. ✓ ✓ ✓ ✓ ✓ Country FE ✓ ✓ ✓ ✓ ✓ Year FE ✓ ✓ ✓ ✓ ✓ Observations 2819 2128 3682 2719 569 331 2910 1134 3817 2891 R2 0.18 0.82 0.00 0.19 0.10 0.33 0.02 0.65 0.07 0.14 # Countries 111 116 42 48 120 Panel B: Capital Markets Private debt Liabilities (log) US bank Exchange rate (1) (2) (3) (4) (5) (6) (7) (8) OLS FE OLS FE OLS FE OLS FE US troops (log) 0.181*** 0.040** 0.505*** 0.050*** 0.792*** 0.036 4.466 9.413** (0.024) (0.02) (0.015) (0.01) (0.026) (0.026) (24.601) (4.545) Control var. ✓ ✓ ✓ ✓ Country FE ✓ ✓ ✓ ✓ Year FE ✓ ✓ ✓ ✓ Observations 3701 2727 3723 2787 2596 1991 1594 1289 R2 0.01 0.29 0.22 0.81 0.26 0.29 0.00 0.20 # Countries 117 117 119 56 Notes: (1) The sample is restricted to non-OECD countries. (2) The unit of analysis is a country-year. (3) All explanatory variables are lagged by one year. (4) The control variables are identical to those used in Table 1. (5) For the analysis of interest rates, we removed country-years with rate spreads above the ninety-fifth percentile. (6) Statistical significance: *p < 0.1; **p < 0.05; ***p < 0.01. View Large Table 3. Outcomes are listed at the top of each column Mechanisms Panel A: Policies Public spending Public debt CB policy rate Deposit insurance Capital openness (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) OLS FE OLS FE OLS FE OLS FE OLS FE US troops (log) 0.382*** 0.023*** 0.124** 0.135** –0.909*** –0.703** 0.031*** 0.018** 0.196*** 0.007 (0.015) (0.005) (0.048) (0.06) (0.116) (0.28) (0.004) (0.009 (0.012) (0.017) Control var. ✓ ✓ ✓ ✓ ✓ Country FE ✓ ✓ ✓ ✓ ✓ Year FE ✓ ✓ ✓ ✓ ✓ Observations 2819 2128 3682 2719 569 331 2910 1134 3817 2891 R2 0.18 0.82 0.00 0.19 0.10 0.33 0.02 0.65 0.07 0.14 # Countries 111 116 42 48 120 Panel B: Capital Markets Private debt Liabilities (log) US bank Exchange rate (1) (2) (3) (4) (5) (6) (7) (8) OLS FE OLS FE OLS FE OLS FE US troops (log) 0.181*** 0.040** 0.505*** 0.050*** 0.792*** 0.036 4.466 9.413** (0.024) (0.02) (0.015) (0.01) (0.026) (0.026) (24.601) (4.545) Control var. ✓ ✓ ✓ ✓ Country FE ✓ ✓ ✓ ✓ Year FE ✓ ✓ ✓ ✓ Observations 3701 2727 3723 2787 2596 1991 1594 1289 R2 0.01 0.29 0.22 0.81 0.26 0.29 0.00 0.20 # Countries 117 117 119 56 Mechanisms Panel A: Policies Public spending Public debt CB policy rate Deposit insurance Capital openness (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) OLS FE OLS FE OLS FE OLS FE OLS FE US troops (log) 0.382*** 0.023*** 0.124** 0.135** –0.909*** –0.703** 0.031*** 0.018** 0.196*** 0.007 (0.015) (0.005) (0.048) (0.06) (0.116) (0.28) (0.004) (0.009 (0.012) (0.017) Control var. ✓ ✓ ✓ ✓ ✓ Country FE ✓ ✓ ✓ ✓ ✓ Year FE ✓ ✓ ✓ ✓ ✓ Observations 2819 2128 3682 2719 569 331 2910 1134 3817 2891 R2 0.18 0.82 0.00 0.19 0.10 0.33 0.02 0.65 0.07 0.14 # Countries 111 116 42 48 120 Panel B: Capital Markets Private debt Liabilities (log) US bank Exchange rate (1) (2) (3) (4) (5) (6) (7) (8) OLS FE OLS FE OLS FE OLS FE US troops (log) 0.181*** 0.040** 0.505*** 0.050*** 0.792*** 0.036 4.466 9.413** (0.024) (0.02) (0.015) (0.01) (0.026) (0.026) (24.601) (4.545) Control var. ✓ ✓ ✓ ✓ Country FE ✓ ✓ ✓ ✓ Year FE ✓ ✓ ✓ ✓ Observations 3701 2727 3723 2787 2596 1991 1594 1289 R2 0.01 0.29 0.22 0.81 0.26 0.29 0.00 0.20 # Countries 117 117 119 56 Notes: (1) The sample is restricted to non-OECD countries. (2) The unit of analysis is a country-year. (3) All explanatory variables are lagged by one year. (4) The control variables are identical to those used in Table 1. (5) For the analysis of interest rates, we removed country-years with rate spreads above the ninety-fifth percentile. (6) Statistical significance: *p < 0.1; **p < 0.05; ***p < 0.01. View Large These tests are important because they help us rule out spurious correlations caused by endogeneity. If troops were deployed in countries that were imminently at risk of a financial breakdown, we would expect troops to be negatively correlated with total foreign liabilities and US bank exposure, because international investors would avoid these areas. Furthermore, we would observe a weakening of the exchange rate as capital would leave these countries for safer places. Domestic policy makers would have to increase the interest rate to stop the bleeding. As we show, none of these predictions materialize. Instead, foreign liabilities and US bank exposure increase with US troops, the exchange rate strengthens, and the policy rates decreases as US troops increase. The combination of lower interest rates and capital inflows is particularly interesting because it goes against conventional macroeconomic wisdom and can only be explained by a powerful pull effect coming from these countries. The results confirm our expectations. When troops depart from the sample mean by one standard deviation, government consumption increases by 0.05 to 0.8 percent. The former estimate is probably more reliable, suggesting that the effect is fairly small, though statistically significant. Besides consumption, we find that public debt also goes up by a small amount: 0.24 to 0.28 percent. The effect depends less on model specification and remains consistent with our expectations. Turning our attention to monetary policy, we find a substantial decrease in interest rates. The Central Bank's headline policy rate declines by about 1.2 to 1.4 percentage points. The magnitude of this effect suggests that US support matters just as much in term of its implications on monetary compared to financial policy. The presence of American troops encourages an expansive monetary policy. This makes sense if countries were previously constrained in their access to international capital markets. Next, we look at two risk-inducing regulations. We find a positive effect of troops on the adoption of a deposit insurance. An increase of troops by one standard deviation above the mean raises the likelihood of adopting a deposit insurance by 3 to 6 percent. Evidence is weaker on capital openness. We find a positive effect, but it becomes statistically insignificant when all the control variables are included. Continuing with Panel B, we find that private debt increases by 0.1 to 0.3 percent. Thus, US support generates an increase in both public and in private spending. The point estimates in both cases are small but consistent with the theory. Total foreign liabilities of the country also increase by an estimated 0.1 to 1.1 percent. The effects of US bank exposure are similar, though with a broader range, since it increases by 0.1 to 1.7 percent. Finally, the exchange rate strengthens by 10 to 23 points. As we noted in the theory section, the causal models for each of these outcomes could be considerably more complicated than reported here. To ensure that our results are robust, we estimate a set of systems of equations that allows for correlation between each of these nine outcomes. The results are reported in Table A47. In addition, we used principal-component factor analysis to generate generic policy and market reactions to troops using the variables in Table 3, Panel A and B respectively. The results are reported in Table A48. Markets seem more reactive to the insurance provided by troops than policies. However, given that there may exist other policies that mattered for some countries, the policy effect could be understated. Besides these tests, we examine additional outcomes that could plausibly be affected by US support. We confirm that private and public financial leverage increases, and yet the quality of borrowing decreases. Specifically, the economy-wide z-score (which measures the probability of insolvency of the domestic financial system) is positively associated with the numbers of US troops. At the same time, the share of nonperforming loans also increases. These findings suggest the presence of a market failure, which arises from moral hazard. We report these results in Table A50. In light of the importance of international capital inflows and indebtedness in the context of a financial crisis, we expand our analysis and find that “hot money” inflows—such as portfolio, derivative, and debt liabilities—significantly increase (Noy 2008). We report these findings in Table A49. Overall, we can detect a substantial buildup of macrofinancial vulnerabilities. Conclusion The international moral hazard “fairy,” often invoked but rarely seen, forms the cornerstone of most opposition to bailouts. In theory, interventions by the IMF or other lenders of last resort may create incentives to engage in poor economic policy-making. In practice, studies so far mostly struggled to isolate the presence of moral hazard coming from international financial rescues. This lack of clear-cut evidence stems, in part, from measurement issues. Using US military personnel as a proxy for political support to local hosts, we find that moral hazard significantly contributes to financial instability. Increasing the number of US troops by one standard deviation above the mean raises the probability of experiencing a financial crisis by up to 13 percentage points. Furthermore, countries that host more US troops are more likely to adopt risky policies. Capital markets appear happy to follow suit. In light of these findings, the key challenge faced by the United States is to promise credibly not to bail out “hazardous” governments and investors. In doing so, the United States faces a time inconsistency problem: if its interests align with those of a foreign country or its creditors, then it will have little choice but to intervene in the event of an economic crises. To contain borrower moral hazard, the United States attempted to establish a credible no-bailout policy stance and thus did not come to the rescue of Russia in 2000 or Argentina in 2002. But this triggered, from Washington's perspective, unacceptable levels of turbulence in international financial markets. Another solution is to impose tough reforms ex post on countries and financial institutions that receive bailouts. The United States and the IMF have both made efforts to do so in the past. Governments in East Asia (which implicitly believed that their close ties to the United States would save them) were incensed when the IMF imposed demanding structural adjustment programs (Walter 2013; Rickard and Caraway 2018). These experiences motivated a policy shift in these economies toward rapidly accumulating foreign reserves. Although this shift likely made these countries financially more resilient and better equipped to deal with the outbreak of a financial crisis, it also contributed to the rise of global imbalances during the 2000s (Alfaro, Kalemli-Ozcan, and Volosovych 2015). Against this background, strengthening the monitoring and surveillance mandate of the IMF combined with better coordination between the US administration and the international financial community might represent the only feasible option. From an international governance perspective, this might prove preferable to a disengagement of America's overseas presence. Indeed, we note that our results do not answer questions concerning the net welfare effect of US troop deployments. US troops generate other side benefits that may or may not exceed the cost of financial instability. We simply emphasize that American foreign policy might be intertwined with international financial instability. We conclude with two thoughts about the broader implications of this article. First, a hegemonic system may inherently produce financial instability. This stands in sharp contrast with the dominant streams of the literature, which tend to conclude that hegemony enhances international stability (Ba 2018) or, at the very least, that it is a stable system. Norrlof (2010, 10) and Norrlof and Wohlforth (2016, 2) argue that the United States will be able to maintain its position for the foreseeable future. This may well be the case, but our concern is that the United States’ international position creates external instability, especially in the so-called periphery. In doing so, we join several voices that express concerns about a hegemon's ability to maintain external stability (Johnson 2000; Monteiro 2012). We argue that, without targeted policies to prevent moral hazard, a hierarchical system may have built-in sources of instability coming from the top. Crises are not only the result of a decline of hegemonic power; they are intrinsically part of such a system as long as local authorities possess a degree of autonomy. Second, and paradoxically, a potential weakening of the United States’ primacy in international financial markets might even worsen the moral hazard problem (Woods 2008; Brautigam 2009; Cooley and Nexon 2013; Bauerle Danzman, Winecoff, and Oatley 2017). The recent inception of the BRICS Bank possibly allows financial institutions and especially sovereign borrowers to choose between competing lenders of last resort. As lending and bailout decisions closely follow foreign policy preferences of key shareholders, governments will likely have more power to exploit their geo-strategic position and thus have fewer incentives to pursue prudent macroeconomic policies. For these reasons, we believe that future approaches to analyzing moral hazard in international financial markets need to account for these shifts in the global financial architecture. Notes Authors’ note: We are grateful to two anonymous reviewers, Dan Nexon and the editors at International Studies Quarterly, Eric Arias, Mark Copelovitch, Axel Dreher, Julia Gray, Mike Horowitz, Ed Mansfield, Tom Pepinsky, Julia Rittershausen, George Shambaugh, Alex Weisiger, Thomas Willett, and Matt Winters for extremely useful comments. We are also thankful for feedback from the participants to workshops at the University of Heidelberg, the University of Pennsylvania, Washington University in St. Louis, APSA, IPES, ISA, and MPSA. Finally, Theron Guzoto provided outstanding research assistance. Partial funding for this project from the Georgetown Competitive Grant initiative is gratefully acknowledged. All errors remain ours. Author Biographical Michaël Aklin is an assistant professor of political science at the University of Pittsburgh. He holds a courtesy appointment at the Graduate School of Public and International Affairs. Andreas Kern is an associate teaching professor in the McCourt School of Public Policy at Georgetown University. Footnotes 1 We offer quantitative and qualitative evidence from many sources, such as the so-called Wikileaks embassy cables, to validate our interpretation of the data. 2 See Section A3 for additional qualitative evidence. 3 Recently, the United States expanded its bond insurance program through which it fully backs the issuance of sovereign bonds of important allies in international financial markets. For example, Ukraine and Jordan were able to obtain more than |${\$}$|2.5 billion at historically low interest rates that would have otherwise not been available (Tepper 2014). 4 We are somewhat constrained by the paucity of large-N financial crises data that cover different types of crises. The main extant sources exist for the period until 2009 or 2010. All countries used in the analysis are listed in Table A4. The main sample does not include countries in years in which they are at war, but as Tables A39 and A40 show, the results are almost exactly the same when we include all countries and control for the presence of a violent conflict. 5 Figure A3 reports historical trends in troop deployment. We verify that country-years without any US troops are not driving the results in Table A27. 6 The results hold if we use a simple dummy variable that takes value 1 if there are any troops in the country (excluding security personnel) (Table A13). 7 We show that the results remain similar if we use troops per capita (Table A12) or the log of troops per capita (Table A11). 8 Table A22. 9 Table A24. 10 In Table A9, we explore the onset of crises instead of their occurrence in any given year, with similar results. For the onset of crises, we estimate the same models as below, except that we include a lagged dependent variable to reduce bias. 11 We replicate our main results using external sovereign debt crises as the sole dependent variable (Table A8). Since defaults are the type of crises that are most tied to the behavior of governments, we expect to obtain results in support of our hypothesis. This is indeed what we find. The significance level for the other types of crises tends to be more sensitive to specification, although it appears that the same results hold for banking crises and, to a lesser extent, currency and inflation crises (Table A10). Separately, we combine the Reinhart and Rogoff data with Laeven and Valencia (2013), with very similar results (Table A7). 12 Friedman (1997) argues that interest rates are not good indicators of monetary policy. However, we are unaware of a better measure that would be available for many countries over an extended period. 13 We removed countries with observations above the ninety-fifth percentile, since the spreads contain a few extreme outliers. Including them makes the results stronger. 14 We use random instead of country fixed effects in Table A45; the results are about the same. 15 Our results are robust to arbitrary spatial correlation using standard errors from Driscoll and Kraay (1998) in Table A44. 16 Our results remain the same without country fixed effects (Table A44). 17 See Section A11. 18 More precisely, since this is akin to a difference-in-difference model, the number of troops could actually decrease, but less so than in the rest of the world. 19 The sample is the same as in the previous results. See also Table A57 for the estimates controlling for wars. 20 The results using the entire sample are reported in Table A51 with similar results. 21 Our instrument does not offer much variation over time. We can obtain more temporal variation by looking at the intensity of the conflict. We measure it by the number of US casualties per year during the conflicts. We can then instrument the number of troops with the interaction of Echelon status with the number of casualties. Doing so, and including year fixed effects, does not fundamentally alter our findings (Table A55). We replicate the same analysis replacing the number of casualties with the number of troops in combat in Afghanistan and Iraq (Table A56). For sake of space, we leave the discussion of these models and the data we used in Section A12. We are grateful to one of our reviewers for this suggestion. References Alesina Alberto , Roubini Nouriel , Cohen Gerald D. . 1997 . Political Cycles and the Macroeconomy . Cambridge, MA : MIT Press . Alfaro Laura , Kalemli-Ozcan Sebnem , Volosovych Vadym . 2015 . “ Sovereigns, Upstream Capital Flows, and Global Imbalances .” Journal of the European Economic Association 12 ( 5 ): 1240 – 84 . Google Scholar Crossref Search ADS Ansell Ben . 2014 . “ The Political Economy of Ownership: Housing Markets and the Welfare State .” American Political Science Review 108 ( 2 ): 383 – 402 . Google Scholar Crossref Search ADS Antoniades Alexis , Calomiris Charles W. . 2015 . “ Mortgage Market Credit Conditions and U.S. Presidential Elections .” NBER Working Paper No. 24459, https://www.nber.org/papers/w24459 . Ba Heather . 2018 . “ The Systemic Causes of Financial Crises in the Long Nineteenth Century .” Business and Politics 20 ( 2 ): 208 – 38 . Google Scholar Crossref Search ADS Bailey Michael , Strezhnev Anton , Voeten Erik . 2015 . “ Estimating Dynamic State Preferences from United Nations Voting Data .” Journal of Conflict Resolution 61 ( 2 ): 430 – 56 . Google Scholar Crossref Search ADS Bartels Larry M. 2013 . “ Political Effects of the Great Recession .” Annals of the American Academy of Political and Social Science 650 ( 1 ): 47 – 75 . Google Scholar Crossref Search ADS Bauerle Danzman Sarah , Winecoff W. Kindred , Oatley Thomas . 2017 . “ All Crises Are Global: Capital Cycles in an Imbalanced International Political Economy .” International Studies Quarterly 61 ( 4 ): 907 – 23 . Google Scholar Crossref Search ADS Beck Nathaniel . 2015 . “ Estimating Grouped Data Models with a Binary Dependent Variable and Fixed Effects: What Are the Issues? ” Working Paper, https://arxiv.org/abs/1809.06505 . Berger Daniel , Easterly William , Nunn Nathan , Satyanath Shanker . 2013 . “ Commercial Imperialism? Political Influence and Trade During the Cold War .” American Economic Review 103 ( 2 ): 863 – 96 . Google Scholar Crossref Search ADS Biglaiser Glen , DeRouen Karl . 2007 . “ Following the Flag: Troop Deployment and Us Foreign Direct Investment .” International Studies Quarterly 51 ( 4 ): 835 – 54 . Google Scholar Crossref Search ADS Biglaiser Glen , DeRouen Karl . 2009 . “ The Interdependence of US Troop Deployments and Trade in the Developing World .” Foreign Policy Analysis 5 ( 3 ): 247 – 263 . Google Scholar Crossref Search ADS Brautigam Deborah . 2009 . The Dragon's Gift: The Real Story of China in Africa . Oxford : Oxford University Press . Brown Craig O. , Dinc I. Serdar . 2005 . “ The Politics of Bank Failures: Evidence from Emerging Markets .” Quarterly Journal of Economics 120 ( 4 ): 1413 – 44 . Google Scholar Crossref Search ADS Bueno de Mesquita Bruce , Smith Alastair , Siverson Randolph M. , Morrow James D. . 2003 . The Logic of Political Survival . Cambridge, MA : MIT Press . Caballero Ricardo J. , Farhi Emmanuel , Gourinchas Pierre-Olivier . 2006 . “ An Equilibrium Model of ‘Global Imbalances’ and Low Interest Rates .” NBER Working Paper 11996, https://www.nber.org/papers/w11996 . Caldera Louis , Reimer Dennis J. . 1999 . “ A Statement on the Posture of the United States Army Fiscal Year 2000 .” Statement presented to the Committees and Subcommittees of the United States Senate and the House of Representatives, Washington, DC . Calomiris Charles W. 2000 . “ When Will Economics Guide IMF and World Bank Reforms? ” Cato Journal 20 ( 1 ): 85 – 103 . Chaudoin Stephen , Milner Helen V. , Pang Xun . 2015 . “ International Systems and Domestic Politics: Linking Complex Interactions with Empirical Models in International Relations .” International Organization 69 ( 2 ): 275 – 309 . Google Scholar Crossref Search ADS Cheibub Jose A. , Gandhi Jennifer , Vreeland James . 2010 . “ Democracy and Dictatorship Revisited .” Public Choice 143 ( 1 ): 67 – 101 . Google Scholar Crossref Search ADS Cohen Craig , Chollet Derek . 2007 . “ When |${\$}$|10 Billion Is Not Enough: Rethinking U.S. Strategy Toward Pakistan .” Washington Quarterly 30 ( 2 ): 7 – 19 . Google Scholar Crossref Search ADS Conley Timothy G. , Hansen Christian B. , Rossi Peter E. . 2010 . “ Plausibly Exogenous .” Review of Economics and Statistics 94 ( 1 ): 260 – 72 . Google Scholar Crossref Search ADS Cooley Alexander . 2008 . Base Politics: Democratic Change and the U.S. Military Overseas . Ithaca, NY : Cornell University Press . Cooley Alexander , Nexon Daniel H. . 2013 . “ ’The Empire Will Compensate You’: The Structural Dynamics of the US Overseas Basing Network .” Perspectives on Politics 11 ( 4 ): 1034 – 50 . Google Scholar Crossref Search ADS Copelovitch Mark S. 2010 . “ Master or Servant? Common Agency and the Political Economy of IMF Lending .” International Studies Quarterly 54 ( 1 ): 49 – 77 . Google Scholar Crossref Search ADS Crespo-Tenorio Adriana , Jensen Nathan M. , Rosas Guillermo . 2013 . “ Political Liabilities Surviving Banking Crises .” Comparative Political Studies 47 ( 7 ): 1047 – 74 . Google Scholar Crossref Search ADS Cronin Richard . 1998 . “ Asian Financial Crisis: An Analysis of US Foreign Policy Interests and Options .” Congressional Research Service, Report No. 98-74 F, https://digital.library.unt.edu/ark:/67531/metacrs692/m1/1/high_res_d/98-74f_1998Apr23.pdf . Dam Lammertjan , Koetter Michael . 2012 . “ Bank Bailouts and Moral Hazard: Evidence from Germany .” Review of Financial Studies 25 ( 8 ): 2343 – 80 . Google Scholar Crossref Search ADS Davidson Adam . 2015 . “ Why Greece's Lenders Need to Suffer .” New York Times Magazine , July 28, https://www.nytimes.com/2015/08/02/magazine/why-greeces-lenders-need-to-suffer.html . Demirgüç-Kunt Asli , Detragiache Enrica . 2002 . “ Does Deposit Insurance Increase Banking System Stability? An Empirical Investigation .” Journal of Monetary Economics 49 ( 7 ): 1373 – 406 . Google Scholar Crossref Search ADS Department of State . 2008 . “ United States Department of State – Country Reports on Terrorism 2007 .” United States Department of State Publication Office of the Coordinator for Counterterrorism. https://www.state.gov/documents/organization/105904.pdf . Diamond Douglas W. , Dybvig Philip H. . 1983 . “ Bank Runs, Deposit Insurance, and Liquidity .” Journal of Political Economy 91 ( 3 ): 401 – 19 . Google Scholar Crossref Search ADS Dinç I. Serdar . 2005 . “ Politicians and Banks: Political Influences on Government-Owned Banks in Emerging Markets .” Journal of Financial Economics 77 ( 2 ): 453 – 79 . Google Scholar Crossref Search ADS Dixon Hugo . 2011 . “ Moral Hazard is Clue to Solving Euro Crisis .” Reuters , May 20, http://blogs.reuters.com/breakingviews/2011/05/20/moral-hazard-is-clue-to-solving-euro-crisis/ . Drazen Allan . 2000 . “ The Political Business Cycle After 25 Years .” In NBER Macroeconomics Annual 2000 , edited by Bernanke Ben , Rogoff Kenneth , 75 – 138 . Cambridge, MA : MIT Press . Dreher Axel . 2004 . “ Does the IMF Cause Moral Hazard? A Critical Review of the Evidence .” Exeter Working Paper, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=505782 . Dreher Axel , Vaubel Roland . 2004 . “ Do IMF and IBRD Cause Moral Hazard and Political Business Cycles? Evidence from Panel Data .” Open Economies Review 15 (1): 5 – 22 . Google Scholar Crossref Search ADS Dreher Axel , Sturm Jan-Egbert , Vreeland James Ramond . 2014 . “ Politics and IMF Conditionality .” Journal of Conflict Resolution 59 ( 1 ): 120 – 48 . Google Scholar Crossref Search ADS Drezner Daniel W. , McNamara Kathleen R. . 2013 . “ International Political Economy, Global Financial Orders and the 2008 Financial Crisis .” Perspectives on Politics 11 ( 1 ): 155 – 66 . Google Scholar Crossref Search ADS Driscoll John C. , Kraay Aart C. . 1998 . “ Consistent Covariance Matrix Estimation with Spatially Dependent Panel Data .” Review of Economics and Statistics 80 ( 4 ): 549 – 60 . Google Scholar Crossref Search ADS Eichengreen Barry . 1996 . “ Hegemonic Stability Theory and Economic Analysis: Reflections on Financial Instability and the Need for an International Lender of Last Resort .” CIDER Working Paper C96-080, https://eml.berkeley.edu/∼eichengr/research/c96-80.pdf . Farrell Henry , Newman Abraham . 2016 . “ The New Interdependence Approach: Theoretical Development and Empirical Demonstration .” Review of International Political Economy 23 ( 5 ): 713 – 36 . Google Scholar Crossref Search ADS Forbes Kristin J. , Warnock Francis E. . 2012 . “ Capital Flow Waves: Surges, Stops, Flight, and Retrenchment .” Journal of International Economics 88 ( 2 ): 235 – 51 . Google Scholar Crossref Search ADS Frankel Jeffrey A. 2005 . “ Contractionary Currency Crashes in Developing Countries .” NBER Working Paper 11510, https://www.nber.org/papers/w11508 . Friedman Milton . 1997 . “ Rx for Japan: Back to the Future .” Wall Street Journal , December 17, https://www.wsj.com/articles/SB882308822323941500 . Garner Thomas Nathaniel . 2014 . “ The Economy of Military Engagement: An Analysis of Positive Externalities of US Troop Deployments .” Dissertation, Georgetown University . Göktepe Gökçe , Satyanath Shanker . 2013 . “ The Economic Value of Military Connections in Turkey .” Public Choice 155 ( 3–4 ): 531 – 52 . Google Scholar Crossref Search ADS Gourevitch Peter . 2013 . “ Yet More Hard Times? ” In Politics in the New Hard Times: The Great Recession in Comparative Perspective , edited by Kahler Miles , Lake David A. , 253 – 74 . Ithaca, NY : Cornell University Press . Gourevitch Peter . 1978 . “ The Second Image Reversed: The International Sources of Domestic Politics .” International Organization 32 ( 4 ): 881 – 912 . Google Scholar Crossref Search ADS Gray Julia . 2009 . “ International Organization As a Seal of Approval: European Union Accession and Investor Risk .” American Journal of Political Science 53 ( 4 ): 931 – 49 . Google Scholar Crossref Search ADS Gray Julia . 2013 . The Company States Keep: International Economic Organizations and Investor Perceptions . New York : Cambridge University Press . Greene William . 2003 . “ The Behavior of the Maximum Likelihood Estimator of Limited Dependent Variable Models in the Presence of Fixed Effects .” Econometrics Journal 7 ( 1 ): 98 – 119 . Google Scholar Crossref Search ADS Guisinger Alexandra , Brune Nancy E. . 2017 . “ The Diffusion of Financial Liberalization in Developing Countries1 .” Foreign Policy Analysis 13 ( 1 ): 1 – 27 . Haggard Stephen . 2000 . The Political Economy of the Asian Financial Crisis . Washington, DC : Institute for International Economics . Hellwig Timothy . 2007 . “ Economic Openness, Policy Uncertainty, and the Dynamics of Government Support .” Electoral Studies 26 ( 4 ): 772 – 86 . Google Scholar Crossref Search ADS Hellwig Timothy , Coffey Eva . 2011 . “ Public Opinion, Party Messages, and Responsibility for the Financial Crisis in Britain .” Electoral Studies 30 ( 3 ): 417 – 26 . Google Scholar Crossref Search ADS Henke Marina E. 2018 . “ The Rotten Carrot: US-Turkish Bargaining Failure Over Iraq in 2003 and the Pitfalls of Social Embeddedness .” Security Studies 27 ( 1 ): 120 – 47 . Google Scholar Crossref Search ADS Hennessy Karen . 2016 . “ A Legacy of Lessons Learned: Landstuhl Regional Medical Center During Wartime, 2001–2014 .” Washington, DC : Department of Defense (DoD) . House of Representatives . 1998a . “ Hearing Before the Subcommittee on Asia and the Pacific and International Economic Policy and Trade of the Committee on International Relations .” House of Representatives, 105th Congress, Second Session, February 4, 1998 . House of Representatives . 1998b . “ Hearing Before the Subcommittee on Asia and the Pacific of the Committee on International Relations .” House of Representatives, 105th Congress, Second Session, May 7, 1998 . Imai Kosuke , Keele Luke , Tingley Dustin , Yamamoto Teppei . 2011 . “ Unpacking the Black Box of Causality: Learning About Causal Mechanisms from Experimental and Observational Studies .” American Political Science Review 105 ( 4 ): 765 – 89 . Google Scholar Crossref Search ADS Indriason Indrii H. 2014 . “ The Collapse: Economic Considerations in Vote Choice in Iceland .” Journal of Elections, Public Opinion, and Parties 24 ( 2 ): 134 – 59 . Google Scholar Crossref Search ADS Jeanne Olivier , Zettelmeyer Jeromin . 2005 . “ The Mussa Theorem (and Other Results on IMF-Induced Moral Hazard) .” IMF Staff Papers WP/04/192: p.1-26 . Johnson Chalmers . 2000 . Blowback: The Costs and Consequences of American Empire . New York : Holt . Kahler Miles . 1985 . “ Politics and International Debt: Explaining the Crisis .” International Organization 39 ( 3 ): 357 – 82 . Google Scholar Crossref Search ADS Kamin Steven B. 2004 . “ Identifying the Role of Moral Hazard in International Financial Markets .” International Finance 7 ( 1 ): 25 – 59 . Google Scholar Crossref Search ADS Kane Tim . 2004 . “ Global US Troop Deployment, 1950–2003 .” Center for Data Analysis, Report 04–11, http://s3.amazonaws.com/thf_media/2004/pdf/cda04-11.pdf . Kane Tim . 2012 . “ Development and US Troop Deployments .” Foreign Policy Analysis 3 ( 3 ): 255 – 73 . Google Scholar Crossref Search ADS Karcher Sebastian , Steinberg David A. . 2013 . “ Assessing the Causes of Capital Account Liberalization: How Measurement Matters .” International Studies Quarterly 57 ( 1 ): 128 – 37 . Google Scholar Crossref Search ADS Keefer Philip . 2007 . “ Elections, Special Interests, and Financial Crisis .” International Organization 61 ( 3 ): 607 – 41 . Google Scholar Crossref Search ADS Khwaja Asim Ijaz , Mian Atif R. . 2005 . “ Do Lenders Favor Politically Connected Firms? Rent Provision in an Emerging Financial Market .” Quarterly Journal of Economics 120 ( 4 ): 1371 – 411 . Google Scholar Crossref Search ADS Kindleberger Charles P. 1973 . The World in Depression: 1929–1939 . London : Allen Lane . Laeven Luc , Levine Ross . 2009 . “ Bank Governance, Regulation, and Risk Taking .” Journal of Financial Economics 93 ( 2 ): 259 – 75 . Google Scholar Crossref Search ADS Laeven Luc , Valencia Fabián . 2013 . “ Systemic Banking Crises Database .” IMF Economic Review 61 ( 2 ): 225 – 70 . Google Scholar Crossref Search ADS Lake David A. 2007 . “ Escape from the State of Nature: Authority and Hierarchy in World Politics .” International Security 32 ( 1 ): 47 – 79 . Google Scholar Crossref Search ADS Lake David A. . 2009 . Hierarchy in International Relations . Ithaca, NY : Cornell University Press . Lake David A. . 2013 . “ Legitimating Power: The Domestic Politics of U.S. International Hierarchy .” International Security 38 ( 2 ): 74 – 111 . Google Scholar Crossref Search ADS Lane Timothy , Phillips Steven . 2000 . “ Does IMF Financing Result in Moral Hazard? ” IMF Working Paper 168, https://www.imf.org/external/pubs/ft/wp/2000/wp00168.pdf . Lane Timothy , Phillips Steven . 2001 . “ IMF Financing and Moral Hazard .” Finance and Development 38 ( 2 ). Lauricella Tom . 2010 . “ Greek Deal Stirs Moral Hazard Concerns .” Wall Street Journal , February 10, https://www.wsj.com/articles/SB10001424052748704140104575057631234463158 . Machain Carla Martinez , Morgan T. Clifton . 2013 . “ The Effect of Us Troop Deployment on Host States’ Foreign Policy .” Armed Forces & Society 39 ( 1 ): 102 – 23 . Google Scholar Crossref Search ADS MacIntyre Andrew . 2001 . “ Institutions and Investors: The Politics of the Economic Crisis in Southeast Asia .” International Organization 55 ( 1 ): 81 – 122 . Google Scholar Crossref Search ADS Magalhaes Pedro C. 2014 . “ Introduction – Financial Crisis, Austerity, and Electoral Politics .” Journal of Elections, Public Opinion, and Parties 24 ( 2 ): 125 – 33 . Google Scholar Crossref Search ADS Marsh Michael , Mikhaylov Slava . 2014 . “ A Conservative Revolution: The Electoral Response to Economic Crisis in Ireland .” Journal of Elections, Public Opinion, and Parties 24 ( 2 ): 160 – 79 . Google Scholar Crossref Search ADS Mattern Janice Bially , Zarakol Ayşe . 2016 . “ Hierarchies in World Politics .” International Organization 70 ( 3 ): 623 – 54 . Google Scholar Crossref Search ADS Matthews James K. , Holt Cora J. . 1991 . “ So Many, So Much, So Far, So Fast – United States Transportation Command and Strategic Deployment for Operation Desert Shield/Desert Storm .” United States Transportation Command. http://www.jcs.mil/Portals/36/Documents/History/Monographs/Transcom.pdf . McCarty Nolan , Poole Keith T. , Rosenthal Howard . 2013 . Political Bubbles: Financial Crises and the Failure of American Democracy . Princeton, NJ : Princeton University Press . McDowell Daniel . 2012 . “ The US as ‘Sovereign International Last-Resort Lender’: The Fed's Currency Swap Programme During the Great Panic of 2007–09 .” New Political Economy 17 ( 2 ): 157 – 78 . Google Scholar Crossref Search ADS McDowell Daniel 2017 . “ Need for Speed: The Lending Responsiveness of the IMF .” Review of International Organizations 12 ( 1 ): 39 – 73 . Google Scholar Crossref Search ADS McManus Roseanne W , Yarhi-Milo Keren . 2017 . “ The Logic of ‘Offstage’ Signaling: Domestic Politics, Regime Type, and Major Power-Protégé Relations .” International Organization 71 ( 4 ): 701 – 33 . Google Scholar Crossref Search ADS Meernik James , Krueger Eric L. , Poe Steven C. . 1998 . “ Testing Models of Us Foreign Policy: Foreign Aid During and After the Cold War .” Journal of Politics 60 ( 1 ): 63 – 85 . Google Scholar Crossref Search ADS Mian Atif , Sufi Amir , Trebbi Francesco . 2010 . “ The Political Economy of the US Mortgage Default Crisis .” American Economic Review 100 ( 5 ): 1967 – 98 . Google Scholar Crossref Search ADS Monteiro Nuno P. , 2012 . “ Unrest Assured: Why Unipolarity is Not Peaceful .” International Security 36 ( 3 ): 9 – 40 . Google Scholar Crossref Search ADS Mosley Layna . 2003 . Global Capital and National Governments . New York : Cambridge University Press . Nichol Jim . 2003 . “ Tajikistan: Recent Developments and US Interests .” Norrlof Carla . 2010 . America's Global Advantage: US Hegemony and International Cooperation . Cambridge : Cambridge University Press . Norrlof Carla , Wohlforth William C . 2016 . “ Is US Grand Strategy Self-Defeating? Deep Engagement, Military Spending, and Sovereign Debt .” Conflict Management and Peace Science . https://doi.org/10.1177/0738894216674953 . Noy Ilan . 2008 . “ Sovereign Default Risk, the IMF, and Creditor Moral Hazard .” Journal of International Financial Markets, Institutions, and Money 18 ( 1 ): 64 – 78 . Google Scholar Crossref Search ADS Oatley Thomas . 2015 . A Political Economy of American Hegemony . New York : Cambridge University Press . Oatley Thomas , Yackee Jason . 2004 . “ American Interests and IMF Lending .” International Politics 41 ( 3 ): 415 – 29 . Google Scholar Crossref Search ADS Oatley Thomas , Winecoff W. Kindred , Pennock Andrew , Danzman Sarah Bauerle . 2013 . “ The Political Economy of Global Finance: A Network Model .” Perspectives on Politics 11 ( 1 ): 133 – 53 . Google Scholar Crossref Search ADS Pepinsky Thomas B. 2012 . “ The Global Economic Crisis and the Politics of Non-Transitions .” Government and Opposition 47 ( 2 ): 135 – 61 . Google Scholar Crossref Search ADS Pond Amy . 2018 . “ Financial Liberalization: Stable Autocracies and Constrained Democracies .” Comparative Political Studies 51 ( 1 ): 105 – 35 . Google Scholar Crossref Search ADS Rajan Raghuram G. 2010 . Fault Lines: How Hidden Fractures Still Threaten the World Economy . Princeton, NJ : Princeton University Press . Reinhart Carmen M. , Rogoff Kenneth S. . 2009 . This Time Is Different: Eight Centuries of Financial Folly . Princeton, NJ : Princeton University Press . Rey Hélène . 2016 . “ International Channels of Transmission of Monetary Policy and the Mundellian Trilemma .” IMF Economic Review 64 ( 1 ): 6 – 35 . Google Scholar Crossref Search ADS Reynaud Julien , Vauday Julien . 2008 . “ IMF Lending and Geopolitics .” European Central Bank Working Paper Series 0965. http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1292331 . Rickard Stephanie J. , Caraway Teri L. . 2018 . “ International Demands for Austerity: Examining the Impact of the IMF on the Public Sector .” Review of International Organizations 1–23, https://doi.org/10.1007/s11558-017-9295-y . Passari Evgenia , Rey Hélène . 2015 . “ Financial Flows and the International Monetary System .” Economic Journal 125 ( 584 ): 675 – 98 . Google Scholar Crossref Search ADS Satyanath Shanker . 2005 . Globalization, Politics, and Financial Turmoil: Asia's Banking Crisis . New York : Cambridge University Press . Schneider Christina J. , Tobin Jennifer L. . 2013 . “ Bilateral Financial Bailouts .” Working Paper, https://quote.ucsd.edu/cjschneider/files/2017/10/InternationalBailouts_04032017.pdf . Schneider Gerald , Troeger Vera E. . 2006 . “ War and the World Economy: Stock Market Reactions to International Conflicts .” Journal of Conflict Resolution 50 ( 5 ): 623 – 45 . Google Scholar Crossref Search ADS Schularick Moritz , Taylor Alan M. . 2012 . “ Credit Booms Gone Bust: Monetary Policy, Leverage Cycles, and Financial Crises, 1870–2008 .” American Economic Review 102 ( 2 ): 1029 – 61 . Google Scholar Crossref Search ADS Smale Alison , Yardley Jim . 2015 . “ Language of Greek Crisis Shifts From Financial Jargon to Humiliation .” New York Times , March 12, https://www.nytimes.com/2015/03/13/world/language-of-greek-crisis-shifts-from-financial-jargon-to-humiliation.html . Snidal Duncan . 1985 . “ The Limits of Hegemonic Stability Theory .” International Organization 39 ( 4 ): 579 – 614 . Google Scholar Crossref Search ADS Sopko John F. 2014 . “ Afghanistan's Banking Sector: The Central Bank's Capacity to Regulate Commercial Banks Remains Weak .” SIGAR 14-16 Audit Report. https://www.sigar.mil/pdf/audits/sigar%2014-16-ar.pdf. Staiger Douglas , Stock James H. . 1997 . “ Instrumental Variables Regression with Weak Instruments .” Econometrica 65 ( 3 ): 557 – 86 . Google Scholar Crossref Search ADS Stasavage David . 2011 . States of Credit . Princeton, NJ : Princeton University Press . Steinberg David A , Koesel Karrie , Thompson Nicolas . 2013 . “ Political Regimes and Currency Crises .” http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2299685 . Tepper Steven G. 2014 . “ A Helping Hand .” International Financial Law Review 33 ( 3 ): 60 – 64 . Thacker Strom C. 1999 . “ The High Politics of IMF Lending .” World Politics 52 ( 1 ): 38 – 75 . Google Scholar Crossref Search ADS Tomz Michael . 2007 . Reputation and International Cooperation: Sovereign Debt Across Three Centuries . Princeton, NJ : Princeton University Press . Vaubel Roland . 1983 . “ The Moral Hazard of IMF Lending .” The World Economy 6 ( 3 ): 291 – 304 . Google Scholar Crossref Search ADS Voeten Erik . 2000 . “ Clashes in the Assembly .” International Organization 54 ( 2 ): 185 – 215 . Google Scholar Crossref Search ADS Walter Stefanie . 2013 . Financial Crises and the Politics of Macroeconomic Adjustments . New York : Cambridge University Press . Wikileaks . 2003 . “ Istanbul Markets ‘Wait and See’ on Government Policy and Iraq: Betting on US Aid .” Wikileaks . Cable Reference 03ISTANBUL113_a. https://wikileaks.org/plusd/cables/03ISTANBUL113_a.html . Wikileaks . 2011a . “ GOP Still Lacks Plan to Tackle Economic Issues .” WikiLeaks. Cable Reference 08ISLAMABAD3059_a. Accessed October 12, 2018. https://wikileaks.org/plusd/cables/08ISLAMABAD3059_a.html . Wikileaks . 2011b . “ DCM Meets with Stata Bank Governor Akhtar .” WikiLeaks . Cable Reference 08ISLAMABAD3532. Accessed October 12, 2018. https://wikileaks.org/plusd/cables/08ISLAMABAD3532_a.html . Wikileaks . 2011c . “ Tranching and Conditioning of U.S. Assistance to Turkey .” Wikileaks . Cable Reference 03ANKARA2243_a. Accessed October 12, 2018. https://wikileaks.org/plusd/cables/03ANKARA2243_a.html . Wikileaks . 2011d . “ Wolfowitz and Grossman Press Turks for Support on Iraq .” WikiLeaks . Cable Reference 02ANKARA9058_a. Accessed October 12, 2018. https://wikileaks.org/plusd/cables/02ANKARA9058_a.html . Wikileaks . 2011e . “ Scenesetter for SACEUR General Jones's January 24th Visit to Ankara .” WikiLeaks . Cable Reference 03ANKARA548_a. Accessed October 12, 2018. https://wikileaks.org/plusd/cables/03ANKARA548_a.html . Wilf Meredith . 2016 . “ Credibility and Distributional Effects of International Banking Regulations: Evidence from Us Bank Stock Returns .” International Organization 70 ( 4 ): 763 – 96 . Google Scholar Crossref Search ADS Woods Ngaire . 2008 . “ Whose Aid? Whose Influence? China, Emerging Donors, and the Silent Revolution in Development Assistance .” International Affairs 84 ( 6 ): 1205 – 21 . Google Scholar Crossref Search ADS © The Author(s) (2019). Published by Oxford University Press on behalf of the International Studies Association. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - Moral Hazard and Financial Crises: Evidence from American Troop Deployments
JF - International Studies Quarterly
DO - 10.1093/isq/sqy047
DA - 2019-03-01
UR - https://www.deepdyve.com/lp/oxford-university-press/moral-hazard-and-financial-crises-evidence-from-american-troop-FhzQP02vhq
SP - 15
VL - 63
IS - 1
DP - DeepDyve
ER -