Policy and Place in International Economic Coercion

Introduction

Economic statecraft has been a subject of scholarly debate for millennia, from Athens' trade embargo on Megara some 2,400 years ago to the United States' ongoing sanctions against Cuba. Despite this history of discussion, there is little consensus about how and when sanctions should be used, and whether or not they are effective. Indeed, the very task of defining the term "sanctions" is no small challenge. Many scholars and policymakers understand modern sanctions as a means of coercion entirely of its own (Kaempfer and Lowenberg 1992; Drezner 1999), while others see sanctions as an evolved form of big-stick diplomacy (Doxey 2000; Elliot 1995). Still others find sanctions as an effective alternative to war (Eland 1995; Kanovalov et. al. 1995). For the sake of brevity and simplicity, this study will define sanctions as a government or coalition of governments' intentional disruption of normal economic relations with a target state.

Understanding the circumstances under which sanctions are successful is essential for the development of future policies. Additionally, the scholarly implications of answering such a question would shed light on the role of policymakers within international political and economic structures. This study attempts to identify how sanction outcomes vary according to the policy goal of the sender nation and the region in which the sanctions are implemented.

Historical Background

The prevalence of sanctions has drastically increased in the past century (See Appendix 1)¹. The formation of the League of Nations represented a shift in international relations towards viewing economic statecraft as a more peaceful venue for coercion and diplomacy. Most post-World-War economic sanctions regimes have been characterized by the U.S. leadership of an alliance of developed nations to pursue objectives that advance the foreign policy goals of the U.S. (Hufbauer et al. 2007). The rise of U.S. influence has coincided with a sharp rise in the use of economic sanctions. In the first forty-five years of the U.N. Security Council's existence, economic sanctions were only imposed twice – first against South Africa in 1977 and against Rhodesia in 1996. During the 1990s, however, the Security Council enacted sanctions a total of sixteen times (Cortwright and Lopez 1995).

Since their first use, sanctions episodes have fallen into one of six categories: promoting democracy and human rights, forcing compliance with sender government's nuclear policy, influencing the outcomes of internal disputes of the target country, demanding the transfer of persons across international borders, influencing the control over territory, and diminishing the target government's level of power in the international arena (Hufbauer et al. 2007). In the late 1970s, the United States began a pattern of implementing sanctions against countries that it claimed were violating human rights. These cases initially involved much of Latin America, including the 1977 sanctions against Paraguay, Guatemala, Argentina, Nicaragua, El Salvador, and Brazil. As the 20th century came to a close the use of sanctions then spread to the rest of the world (Hufbauer et al. 1990).

Sanctions for nuclear policy compliance have served predominantly for the enforcement of the 1970 Treaty on the Non-Proliferation of Nuclear Weapons (NPT). These sanctions episodes have seen some degrees of success in nonproliferation in South Korea (1975) and Taiwan (1976), but were largely unsuccessful in India (1974, 1978, 1998), Pakistan (1974, 1978, 1998), Argentina (1978), and Brazil (1978) (Ibid.). With regards to influencing internal disputes of target countries, sanctions played at least some role in overthrowing Rafael Trujillo in the Dominican Republic (1961), Joao Goulart in Brazil (1964), and Salvador Allende in Chile (1971). Sanctions also played a large role in ending Apartheid in South Africa (Kaempfer and Lowenberg 1992; Hufbauer et al. 2007). Sanctions have also been implemented on occasion for the extradition of criminals, release of prisoners, or repatriation of persons of interest to the sender country. This category of sanction is exemplified in 1992-2003 U.N. sanctions against Libya for the extradition of the Pan-Am attack suspects and U.S. sanctions against Iran in 1979 for the release of American hostages. Sanctions implemented to influence control over territory include measures against expropriation of land, for withdrawal of target country from disputed territory, for assimilation of target countries bordering the sender country, and for the promotion of independence of certain territories from target nations (Hufbauer et al. 1990).

The last category, diminishing the target government's level of power in the international arena, often serves the sole purpose of weakening the target state, without any particular policy change in mind. This category is characterized by ongoing U.S. sanctions against North Korea to impair their military power and by the 1918 U.K. embargo on Russia to destabilize the Bolsheviks (Ibid.). These "impairment" sanctions are more comparable to direct military intervention because they lack specific policy objectives and are a part of a zero-sum struggle for power in the international arena. they are therefore consistent with the sanctions-as-war-alternative hypothesis.

These six categories are not exhaustive. There are a number of cases that are too specific to be placed in any single category. An example of this very specific sender policy goal is U.S. measures against Japan in the early 1990s for Japan's trade in endangered sea turtles. This particular sanctions episode has little in common with other sanctions episodes, and had no precedent. We can treat such cases as outliers – interesting, but not useful in identifying trends.

Literature Review

This study is an analysis of two variables that help predict compliance or noncompliance in sanctioned states. These two variables – policy goal of sender country (policy) and region in which sanctions are implemented (place) – are widely neglected in existing sanctions research.

The rise in the use of sanctions in the past decades has fostered a growth in both volume and diversity of scholarly work. One of the most pressing questions that these scholars address is "Do sanctions work?" International relations and economics scholars tend to argue that economic sanctions are ineffective methods for changing target country behavior. The most definitive analysis of the effectiveness of sanctions, conducted by Gary Hufbauer et al., concluded that sanctions lead to compliance only 34% of the time (2007). Edward Pound and Jihan El-Tahri agreed that "this most favored tool of diplomats is such a deeply flawed instrument" (1994). Margaret Doxey, one of the founders of the study of economic statecraft, acknowledges the poor success rate of sanctions and argues that threatening or using military action is quicker and more effective than sanctions (2000).

Most scholars concur that sanctions generally have little positive effect. However, they often identify circumstances under which sanctions can be effective, and they tend to take two different approaches to identifying such circumstances.

The first approach identifies domestic groups as the primary force behind sanctions enactment. Kaempfer and Lowenberg argue that "a theory explaining sanctions... should be based on individual political agents, including politicians, voters, bureaucrats, and interests groups" (1992). These two scholars advocate for the public choice approach to economic statecraft analysis, an approach that contends that interest groups within states seek redistributional benefits for themselves and their members. They claim that sanctioning governments seek to display their resolve to change the target government's behavior and also to minimize the domestic economic consequences of implementing those sanctions. Sanctions are enacted on behalf of domestic interest groups against interest groups in the target country as symbolic measures.

Similarly, Hufbauer et al. (2007) identify "demonstration of resolve" as a common reason for implementing economic sanctions. Heads of state, and the U.S. president in particular, often feel the need to send a message both to domestic and international audiences about their disapproval of certain target country behaviors. When the U.S. government implements sanctions for this reason, they often do so because the cost of not taking any action is greater than the cost of implementing sanctions (Hufbauer et al. 2007) That is, policymakers in these cases believe that not enacting sanctions to block a target country behavior that domestic and international audience condemn could hurt elected officials' re-election hopes at home and would harm U.S. reputation internationally. Thus establishing a precedent for what kinds of objectionable behavior a misbehaving government can get away with. The Comprehensive Anti-Apartheid Act (1986), for example, was drafted in an atmosphere of large anti-apartheid sentiments within the American public and eventually passed over the veto of President Reagan, who was in his second term and therefore unconcerned with re-election.

The second approach to sanctions analysis treats economic statecraft as a form of direct government-to-government coercion, often as an alternative or precursor to military engagement. Daniel Drezner (1999) maintains that the primary factor determining the outcome of sanctions is anticipation of future conflict: "the strategies of the two countries depend on the opportunity costs of a stalemate outcome and the expectations of future conflict. As the target country's opportunity costs of deadlock rise, sanctions are less likely and less useful." Hufbauer et al. (2007) echo this approach in one of their central arguments, arguing that sanctions are successful when the economic and political costs to the target of noncompliance is greater than the political and security costs of compliance.

Within these two approaches there are three factors that scholars must grapple with in order to draw meaningful conclusions about economic statecraft. They are: (1) the speed and duration of the sanctions; (2) the differing effects on the target leader, target elite groups, and target citizens; (3) the level of multilateral compliance.

Scholars make one of two assertions about the speed with which sanctions should be implemented. One argument holds that sanction implementation should be rapid and comprehensive; the other holds that implementation should be protracted and gradual. Kimberly Elliot (1995) maintains that the impact of sanctions is maximized when they are implemented "quickly and decisively." This decisiveness she says, "allows the target no time to adjust and enhances the political credibility of the sender, signaling its commitment to the sanctioning." This side of the argument is joined by Cortwright and Lopez (1995), who argue, "comprehensive, rigorously enforced sanctions are more likely to be successful than limited, unenforced measures."

On the other hand, Cortwright and Lopez (1995) also point out that the U.N. is more likely to maintain its credibility as a force for peace "by providing a graduated and escalatory response to conflict.. in a softer, nonmilitary mode." In addition, some scholars contend that there is a psychological effect to slower, more protracted sanctions. Ivan Eland (1995) posits, "the fear of the unknown contained within the threat of future measures may have greater psychological effects on the target than the actual sanctions that turn out to have less drastic economic effects than anticipated." Eland also argues that quickly implemented sanctions do not allow for threats of future economic coercion, nor do they allow political leaders in the sender state to ease sanctions without losing face internationally and domestically.

A second area of contention is what sector of the target state sanctions should affect. Margaret Doxey (2000) highlights the humanitarian issues that accompany sanctions, arguing that most sanctions episodes harm the population of the target country, often placing stress on a part of society that often has little control over policymakers' compliance or noncompliance. Therefore, Doxey and other scholars of this school of thought advocate for more targeted measures, such as freezes on the funds of government leaders, expulsion from international sporting events, and travel restrictions. According to Kaempfer and Lowenberg (1992), however, "investment and financial sanctions are also likely to have relatively weak effects on the target country, because of the size and efficiency of world capital markets." For this reason, Kaempfer and Lowenberg advocate for measures that go after the target nation's ruling elite in order to avoid a "rally-around-the-flag" effect, which occurs when political leaders in a target country are able to re-direct the population's frustration into a call to action against the sender country.

The degree to which multilateral compliance is necessary constitutes a third subject for debate. Some scholars, such as Ivan Eland (1995), contend that multilateralism is crucial for sanction success: "failure to get adequate multilateral cooperation can send a message of weakness to the target, rather than the signal of resolve intended by the sanctioning nations." In some cases however, multilateral sanctions can be counterproductive because of the temptation to take advantage of low competition for target countries' export markets.

These factors, while useful in understanding the outcomes of individual sanctions cases, neglect two of the most important elements of any sanctions episode – policy and place. Most game-theoretic sanctions models begin with a binary decision of the sender party: to sanction or not to sanction. To phrase that decision as a question from the sender's perspective – would sanctioning or not sanctioning best advance my foreign policy? As outlined above, analysis of this initial decision tends to focus on the role of interest groups, the level of future conflict that the sender foresees with the target, the magnitude of sanctions, the parties within the target that will be sanctioned, the level of international support, etc. These analyses are incomplete because they take for granted two seemingly simple decisions that precede all other variables – what is the "foreign policy" that the sender government wants to promote, and where do they want to promote it? That is to say, the decision of whether or not a sender should use sanctions to promote their foreign policy can be made only after the sender decides what that policy is and where it should be promoted. This study will show that these factors do play important roles in the outcomes of sanctions episodes.

Some scholars have taken sender demands into consideration in their models, but only in terms of the magnitude of political capital demanded. In his game-theoretic analysis of sanctions success, Drezner (1999) identifies three questions that must be answered to determine the outcome of the sanctions "game." They are: (1) when will the sender prefer "deadlock" to "status quo"?; (2) under what demands will the target prefer "acquiescence" over "deadlock"?; and (3) how will the sender respond to the target's decision? Drezner's answer to the second question (under what demands will the target prefer "acquiescence" over "deadlock"?) could provide a framework to better explain the policy aspect of sanctions episodes.

Drezner bases his answer on an "optimum possible concession," and concludes, "the target's concessions will increase when the difference between Target's and Sender's opportunity costs of deadlock increases." This conclusion, although at first glance rather banal, is useful in understanding the role that political capital plays in economic coercion. Where Drezner and other scholars come short, however, is their failure to identify which types of concessions carry the highest political cost. Understanding which of a sender's demands is more politically costly to a target government is critical to understanding what determines sanction success. This study will seek to determine which of these sender policies is more likely to be met.

Also critical to comprehending sanctions success is an understanding of the regional factors at play. Scholarly investigation of sanctions regimes has restricted itself to isolated sender-target interactions, and for that reason, has overlooked the regional implications of sanctions. This means that scholars have discovered very little about the effect that sanctions regimes have on countries that neighbor the target country or that have strong economic relationships with it. The "demonstration of resolve" that serves as the reasoning behind many senders' implementation of sanctions implies that third, fourth, and nth parties have important impacts on sender-target relations, but the effect within those third, fourth, and nth parties remains neglected. Indeed, some scholars have all but given up on determining these effects. As Kimberly Elliot (1995) puts it, "it is difficult, if not impossible, to know whether a sanction successfully deterred a third party from taking objectionable actions in most cases." Similarly, Hufbauer et al. (2007) begin their analysis with this disclaimer: "our analysis concentrates on the response of the immediate targets. We do not underestimate the exemplary power of forceful action, but it is inherently difficult to know when a good thrashing of one wrongdoer deters bystanders from committing similar misdeeds." Ioana Petrescu (2010) has shown that sanctions can have a deterrent effect on a target government's future actions, but the farther-reaching regional impacts are left unaddressed. This study takes a first step at addressing this seemingly "impossible" question of regional impacts.

Data Analysis

The statistical analyses conducted for this study have shown that the degree to which sanctions episodes are successful varies for different sender policy objectives and for different regions of the world. In other words, policy and place are significant variables in sanctions success. To evaluate sanctions success, this study utilized the 16-point scale developed by Hufbauer et al. This score consists of a "policy result" score of range 1-4 multiplied by a "sanctions contribution score" of 1-4. Hufbauer et al. characterize any sanctions episode with a score of 9 or higher as a successful sanctions episode. This scale is outlined in Economic Sanctions Reconsidered (2007):

Policy Result

1. Failed outcome: illustrated by the Soviet attempt to destabilize Yugoslavia's Marshal Tito in the period 1948-55 (Case 48-4) and by the Indian nuclear tests in the face of threatened U.S. sanctions (Case 98-1).
2. Unclear but possibly positive outcome: illustrated by the U.S. and Saudi efforts to get Jordan to reduce its support for Iraq during the run-up to the first Gulf War.
3. Positive outcome, meaning the sender's goals were partly reached: illustrated by U.S. efforts to promote a return to democracy in several Latin American countries in the 1970s and by U.N. efforts to halt ethnic cleansing and civil conflict in the Balkans (Case 91-1).
4. Successful outcome, in the sense that the sender's goals were largely or entirely realized: illustrated by the joint efforts of the United Kingdom and the United States to overthrow Idi Amin in Uganda in the late 1970s (Case 72-1), by the end of the apartheid era in South Africa (Case 85-1), and by the removal of General Raul Cédras from power in Haiti (Case 91-5).

Sanctions Contribution

1. Negative contribution: illustrated by the U.S./U.N. campaign against the Cédras regime in Haiti (Case 91-5), which triggered an outflow of desperate migrants trying to escape the impacts of sanctions and find work in the United States.
2. Little or no contribution: illustrated by the Soviet withdrawal of assistance from China in the 1960s (Case 60-2) and by U.S. suspension of economic and military aid to Thailand in the wake of the 1991 coup (Case 91-3).
3. Substantial contribution: illustrated by the withdrawal of Dutch economic aid to Suriname between 1982 and 1988 (Case 82-2) and by U.S.-U.N. sanctions against South Africa over apartheid (Case 851).
4. Decisive contribution: illustrated by the U.S. success in deterring a coup in Guatemala in 1993 (Case 93-2).

This study adopted Hufbauer et al.'s data (n=224), which includes nearly a century of sanctions episodes and detailed information relating to each case. This data is the most comprehensive dataset available, and Hufbauer et al.'s 16-point scale is very useful when conducting statistical tests. Other datasets and scoring systems do exist, but none are as exhaustive as this.

To begin the analysis of sender policy effect on sanctions success, six different sender policy objectives were identified and coded. The categories are s follows: (1) promoting democracy and human rights (n = 54), (2) forcing compliance with sender government's nuclear policy (n = 22), (3) influencing the outcomes of internal disputes of the target country (n = 26), (4) demanding the transfer of persons across international borders (n = 8), (5) influencing the control over territory (n = 26), and (6) diminishing the target government's level of power in the international arena (n = 39). In an ideal world, these samples would be larger in order to increase the likelihood of finding a significant difference. However, even with the relatively small sizes of these samples, we will see that there still exists a distinguishable trend in levels of sanction success according to sender policy.

As a side note, I argue that the category of "diminishing the target government's level of power in the international arena," is a fundamentally different form of statecraft (more similar to "economic warfare" than to "economic coercion") because, in these episodes, no policy demands are made of the target state. I included this category of sanctions in the dataset because not doing so would unjustly decrease the degrees of freedom upon which the subsequent analyses of variance are based, thus producing inaccurate results. Aside from its treatment in the statistical analysis, however, this category of sanctions will not be deeply discussed. An in-depth discussion of "economic warfare" as its own form of statecraft is a topic for future research.

Of the 224 cases made available by Hufbauer et al., 175 pertained to one of the six categories outlined in the previous paragraph. The remaining 49 cases did not belong in any single category, and so they were not included in any statistical tests for difference in success score between policy categories. In some cases, the sender party made multiple demands of the target government, so for this reason sanctions episodes were assigned to a category based on the primary sender objective. For example, although the U.N. cited human rights concerns as one of the justifications for their 1993 sanctions against Angola, their primary policy goal was to bring an end to Angola's civil war (Hufbauer et al. 2007). So, this sanctions episode was coded as (3) influencing the outcomes of internal disputes of the target country, rather than (1) promoting democracy and human rights.

In order to begin identifying regional variations in sanctions success, I coded for seven different world regions: (1) the Middle East and North Africa (n = 38), (2) Western Europe (n = 23), (3) Africa South of the Sahara (n = 32), (4) Latin America (n = 44), (5) Pacific Asia (n = 38), (6) South Asia (n = 14), and (7) Eastern Europe, the Balkans, and Ex-Soviet states (n = 32) (see Appendix). It is important to note that a "North American" region consisting of the U.S. and Canada is not included because such a region would have too small a sample size to draw any meaningful statistical conclusions.

The division of the globe into seven regions is, of course, debatable. It is not, however, arbitrary. These regional categories are based on the shared history and close diplomatic and economic ties of each region's members. Not only do countries within these regions share these characteristics, but also there is a noticeable "clustering" of sanctions episodes to each region, which is visible in the map provided (see Appendix 2.2) that shows sanctions episodes by region and by level of success.

I ran a series of analyses of variance to compare mean success scores among different policy categories and different regional categories. The dependent variable in these tests is successscore, derived from the 16-point scale developed by Hufbauer et al. The independent variables are policy_code, coded from 1-6 as described above, and target_code, coded from 1-7 as described above. A preliminary visual comparison of mean successscore and 95% confidence intervals pertaining to those means reveals that there appears to be notable differences in mean successscore for different values of the two variables:

In the above confidence interval comparisons, sanctions for human rights and democracy (policy_code [1]) and for nuclear policy (policy_code [2]) appear to be less successful than sanctions for civil disputes (policy_code [3]) transferring persons (policy_code [4]) or control over territory (policy_code [5]). Likewise, it appears that sanctions in the Middle East and North Africa (target_code [1]), in Western Europe (target_code [2]), in Africa south of the Sahara (target_code [3]), or in Latin America (target_code [4]) are more successful than sanctions in Pacific Asia (target code [5]), in South Asia (target_code [6]), or in Eastern Europe, the Balkans, and Ex-Soviet states (target_code [7]).

A one-way analysis of variance for each variable confirms that there is at least one statistically significant difference in successscore between the six sender policy objectives (F=4.271, df=5,169) and between the seven regional categories (F=2.160, df=7,216). This one-way ANOVA, however, does not take into account the effect that policy_code might have on target_code, or vice-versa. That is, one-way analysis of variance does not consider the fact that one region may be less successful than other regions simply because sanctions with less successful sender objectives are implemented in that region more frequently in other regions. Similarly, one-way analysis of variance does not take into account the fact that one sender policy objective may be less successful than other sender policy objectives simply because that particular sender demand was made in a region that is more resistant to economic coercion than other regions:

A univariate, two-way analysis of variance, which takes into account the effect that policy_code may have on target_code and vice-versa, confirms that policy_code still contains at least one significant difference (F= 4.857, df=5,169) when controlling for the effects of target_code, and target_code still contains at least one significant difference (F=3.033, df=7,216) when controlling for the effects of policy_code (see Appendix 3 for ANOVA output).

Post-hoc analyses reveal which specific policies and regions differ from one another. Tukey's Honest Significant Difference Test² provides pairwise comparisons between all policy objectives and between all regional categories. The following table presents the significant pairwise comparisons for the post-hoc testing on policy_code:

* Significant at the 95% confidence level

According to these results, sanctions for the promotion of democracy and human rights are significantly less effective than sanctions that seek to influence domestic disputes of the target country, to force the transfer of persons, or to influence the control of territory. Likewise, sanctions that promote the nuclear policy of the sender government are significantly less successful than sanctions for the transfer of persons. It is worth noting that successscore for nuclear policy is also approaching significant difference relative to successscore for both internal disputes (p-value=.114) and control of territory (p-value=.126). The entire post-hoc output for policy_code is available in the Appendix (Appendix 4).

A Tukey's HSD post-hoc analysis of target_code yields one significant difference – between Latin America and South Asia (p-value=.051). The entire post-hoc output for target_code is available in the Appendix (Appendix 5). Approaching significance are the differences between Western Europe and South Asia (p-value=.123), between Africa south of the Sahara and South Asia (p-value=.129), and between Latin America and Pacific Asia (p-value=.101).

These statistical analyses suggest that sanctions success varies based on the sender's desired policy outcome and on the region in which sanctions are implemented. Basing the conclusions of this analysis solely on the statistical results would not allow for such results to be extrapolated to the entire population of sanctions episodes. So, given the imperfect nature of the dataset and the myriad methods by which such data could be interpreted, it is important to take into account both statistical significance and observable patterns. That is to say, although statistical significance is not present in every single post-hoc pairwise comparison for policy_code and target_code, there is an observable pattern revealing that sanctions for human rights/ democracy, and nuclear policy are generally less effective than other categories of sanctions, and that sanctions implemented in Pacific Asia, South Asia, and Eastern Europe/Balkans/Ex-Soviet states are generally less effective than sanctions implemented in the Middle East/North Africa, Western Europe, Africa south of the Sahara, and Latin America.

Discussion

Policy and power Maximizing

Why are sanctions for human rights/democracy or nuclear policy less successful than other types of sanctions? The answer appears to lie in the way governing bodies balance the anarchic structure of the international system with competing domestic interests. Different concessions from one governing body to another incur domestic and international political costs. Members of these governing bodies seek to minimize these costs and maximize their gains from any transaction involving political capital. Having established a statistically significant difference between sanctions success for human rights/democracy or nuclear policy and sanctions success for target domestic conflict, transfer of persons, and control over territory. It logically follows that governments targeted by sanctions prefer bringing an end to domestic conflict and forfeiting control over single individuals or small territories over ending human rights abuses, initiating a democratization process, or giving up nuclear power. A model of international relations that takes into account this balance between international and domestic pressures is necessary for a proper interpretation of these results.

There has been much debate in international relations theory surrounding the role of domestic pressures in the formation of foreign policy. Many scholars who take the Realist approach to international relations all but discount the influence of domestic pressures in world politics, treating states as unitary actors that seek to maximize their power in a zero-sum international structure (Waltz 1959; Jervis 1978; Walt 1985; Mearsheimer 2001). This Realist approach is, ironically, not very realistic when considering the fact that states are led by coalitions of individuals, each of which seeks to make themselves better off in the short-run in all circumstances. States themselves are not individuals. States do not act or think. Most importantly, they have no memory and cannot make predictions about the future. These tasks are left entirely up to individuals. When deciding how to respond to any form of coercion on behalf of a foreign power, each individual member of the target state's governing body who is involved in the formulation of foreign policy will seek to improve, in absolute terms, their individual circumstance. The aggregate result of these individual decisions is what ultimately forms foreign policy.

Many scholars of international relations have recognized the central role of domestic politics in foreign policy outcomes. Stephen Krasner (1977) points out that the U.S. government, in its post-WWII efforts to promote free trade and liberal international economics, "confronted [both] external and internal constraints – resistance from other international actors and from particular domestic interest groups. This dual challenge of external and internal resistance must be faced by any state leader." Robert Putnam (1988) goes into more detail about the nature of the interplay between international (Level I) and domestic (Level II) demands in the formation of foreign policy:

"Central executives have a special role in mediating domestic and international pressures precisely because they are exposed to both spheres... The politics of many international negotiations can usefully be conceived as a two-level game. At the national level, domestic groups pursue their interests by pressuring the government to adopt favorable policies, and politicians seek power by constructing coalitions among those groups. At the international level, national governments seek to maximize their own ability to satisfy domestic pressures, while minimizing the adverse consequences of foreign developments. Neither of the two games can be ignored by central decision-makers, so long as their countries remain interdependent, yet sovereign."

Putnam also recognizes the weight that individual decisions have in the process of diplomacy and the development of foreign policy:

"An experienced negotiator familiar with the respective domestic tables should be able to maximize the cost-effectiveness (to him and his constituents) of the concessions that he must make to ensure ratification abroad, as well as the cost-effectiveness of his own demands and threats, by targeting his initiatives with an eye to their Level II incidence, both at home and abroad."

When a head of state is confronted with sanctions, they have a binary decision to make: concede to the sender's demands (end human rights abuses, sign a nuclear treaty, change domestic political structures, release prisoners, withdraw control over a territory) or suffer the economic and political ramifications of sanctions, be they in the form of a trade embargo, freeze of assets, etc. The statistical analysis conducted in this study has shown that heads of states' responses to sanctions vary according to the sender's demands. When heads of state in target countries do not comply with sender demands, it is because each individual within the governing coalitions of the targeted state are better off with the status quo (noncompliance) than with the alternative (compliance).

Heads of state seek to guarantee their own survival and maximize their own level of power, as defined by Morgenthau (1948) in both domestic and international spheres. States that are run by coalitions of such individuals is among the most fundamental properties of international relations. When heads of state comply with sender demands, it is because the projected result of compliance will leave them better off than the projected results of noncompliance. This variation is a direct result of attempts of the governing individuals in the target state to make themselves better off.

The results of this study's statistical analysis suggests that heads of state believe that they are better off by not conceding to human rights/democracy sanctions or nuclear policy sanctions. The results similarly suggest that heads of state believe that conceding to sanctions for domestic disputes, transfer of persons, and control over territory makes themselves better off than not conceding to such sanctions. When confronted with such sanctions, many heads of state, in an effort to convince the sender government that human rights abuses will soon cease, often release a number of political prisoners as their immediate response. Other abuses, however, continue as before. Heads of state take such action because releasing prisoners poses no immediate threat to their survival or individual level of power domestically and internationally.

The following table outlines the strategic process of heads of state when faced with these decisions. This table does not aim to establish universally applicable rules for sanctions outcomes, rather it hopes to lay out a set of generalized, observable patterns that shed light on the role that individual power maximizing plays in the formation of foreign policy.

Implications

So far, this paper has identified individual power maximizing as a core determinant of international relations. Using this model, it becomes clear why there is significant regional variation in sanctions success. Each region of the world contains of a network of governing bodies that engage in regional interstate relations. Each of these governing bodies is comprised of power-maximizing individuals. Thus, each region is home to a community of governing individuals, each of which seeks to maximize their own power. Treated as such, regions become suitable units for analysis using threshold models of collective behavior.

Sociologists have long used threshold models to explain collective action. Some scholars have applied such models to systems of international coercion. Kaempfer and Lowenberg (1992) have used threshold models in their work on economic sanctions, proposing, "Sanctions can serve an important function in stimulating the process of attaining a critical mass of support within a collectivity." Based on the results of this study's statistical analyses of regional variation in sanctions success, it is reasonable to conclude that power-maximizing heads of state within a community of regional governing individuals tend to make decisions about foreign policy based on the decisions of their peers. In other words, when heads of state are deciding how to respond to sanctions, they are likely to make that decision according to how heads of state in neighboring countries are have previously responded (or are responding) to sanctions.

Threshold models for collective action, when applied to this situation, reveal that heads of state, when they perceive that a critical threshold value of other heads of state have responded to sanctions in a particular way, will take the same course of action as those other heads of state. That critical threshold value varies from individual to individual, but it may also be the case that each region has a distinct average critical threshold value. Granovetter (1978) explains how context helps determine threshold values:

"Thresholds are situation-specific. An individual's riot threshold is not a number that he carries with him from one riot to another but rather results from the configuration of costs and benefits, to him, of different behaviors in one particular riot situation. Inevitably, some situations will engage an actor more ideologically than others; one will seem more dangerous, one more exciting."

The existence of context-based variation in critical threshold values could imply that each region of the world has its own range of threshold values for the type of collective action activated in response to economic sanctions. This context-based variation would produce the regional variation in sanction success seen in this study's statistical analyses. Further research is needed to establish what, exactly, determines a region's average critical threshold value. Different standard deviations for threshold critical values within regional samples could also produce this variation in sanction success, and it is even possible that different standard deviations account for more regional variability in sanction success than a difference in means does.

Conclusions

This study has sought to identify and analyze two neglected variables in the study of economic coercion: policy, defined as the desired policy outcome of the sender government, and place, defined as the region within which sanctions are implemented. This study has shown, using multivariate analysis of variance, that there is a statistically significant difference in sanctions success for different values of policy and place. With regards to sender policy objectives, the results of these tests show that heads of state prefer to change political structures, give up control over persons of interest, or withdraw control over disputed territory rather than democratize, end human rights abuses, or change nuclear policy. With regards to regional variation, these tests show that sanctions in the Middle East and North Africa, in Western Europe, in Africa south of the Sahara, and in Latin America are generally more successful than sanctions in Pacific Asia, in South Asia, and in Eastern Europe, the Balkans, and Ex-Soviet states. This study then moves to interpret these results using the power-maximizing individual as the primary unit of analysis, concluding that individual power-maximizing behaviors account for the variability in sanctions success for both policy and place.

Using the power-maximizing individual as the primary unit of analysis, this study suggests that individuals within a governing coalition believe that not conceding to sanctions for human rights/democracy or nuclear policy makes them better off than conceding to such sanctions because concession relinquishes control over the domestic sphere in the case of human rights/ democracy sanctions. In the case of sanctions for nuclear policy, concession relinquishes a share of control of the international sphere. Heads of state seem to believe that concession in either of these situations diminishes the level of power that those individuals possess. Likewise, governing individuals believe that conceding to sanctions for domestic disputes within the target country, for the transfer of persons of interest or for the control over territory makes them better off than not conceding to such sanctions, because not conceding would invite too much international pressure. In this case, heads of state seem to believe that not conceding diminishes the level of power that they possess.

This study also addresses regional variation in sanctions success using the power-maximizing-individuals model. Threshold models of collective behavior, when applied to instances of international coercion, reveal that heads of state, when they perceive that a critical threshold value of other heads of state have responded to sanctions in a particular way, will take the same course of action as other heads of state in the same region. Numerous factors could explain how regions came to possess different threshold levels for collective action, and this is a topic for future research.

Author

Cullen Moran is a student of International Relations. School of International Service, American University.

References

Cortright, D., & Lopez, G. A. (2000). The Sanctions Decade: Assessing U.N. Strategies in the 1990s. Boulder: Lynne Rienner Publishers, Inc.

Cortwright, D., & Lopez, G. A. (1995). "Economic Sanctions in Contemporary Global Relations". In D. Cortwright, & G. A. Lopez, Panacea or Peacebuilding in a Post-Cold War World? (pp. 1-16). Boulder: Westview Press.

Doxey, M. (2000). United Nations Sanctions: Lessons of Experience. Diplomacy and Statecraft, 11(1), 1-18.

Drezner, D. W. (1999). The Sanctions Paradox: Economic Statecraft and International Relations. Cambridge: Cambridge University Press.

Eland, I. (1995). Economic Sanctions as Tools of Foreign Policy. In D. Cortright, & G. A. Lopez, Economic Sanctions: Panacea or Peacebuilding in a Post-Cold War World? (pp. 29-42). Boulder: Westview Press.

Elliot, K. A. (1995). Factors Affecting the Success of Sanctions. In D. Cortright, & G. A. Lopez, Economic Sanctions: Panacea or Peacebuilding in a Post-Cold War World? (pp. 51-59). Boulder: Westview Press.

Granovetter, Mark. (1978). "Threshold Models of Collective Behavior." The American Journal of Sociology 86 (6), p. 1420-1443.

Hufbauer, G., Schott, J., Elliot, K. A., & Oegg, B. (2007). Economic Sanctions Reconsidered (3rd Edition ed.). Washington, DC: Peterson Institute for International Economics.

Hufbauer, G., Schott, J., Elliot, K. A., & Oegg, B. (1990). Economic Sanctions Reconsidered: Supplemental Case Histories. Washington, DC: Peterson Institute for International Economics.

Jervis, Robert. (1978). "Cooperation Under the Security Dilemma." World Politics 30 (2), p. 167- 214.

Keohane, Robert O. (1988). "International Institutions: Two Approaches." 32 (4), p. 379-396.

Konavalov, A., Oznobistchev, S., & Evstafiev, D. (1995). A Review of Economic Sanctions: A Russion Perspective. In D. Cortright, & G. Lopez, Economic Sanctions: Panacea or Peacebuilding in a Post-Cold War World? (pp. 43-48). Boulder: Westview Press.

Krasner, Stephen (1977). "U.S. Commercial and Monetary Policy: Unraveling the Paradox of Eternal Strength and Internal Weakness." International Organization. 31 (4), p. 635- 671.

Lowenberg, W. H. and Kaempfer, L. (1992). International Economic Sanctions. Boulder: Westview Press.

Mearsheimer, John. (2001). The Tragedy of Great Power Politics. London: Norton.

Morgenthau, Hans. (1960). Politics Among Nations: the Struggle for Power and Peace. New York: Alfred A. Knopf.

Petrescu, Ioana M. (2010). "Rethinking Economic Sanctions Success: Sanctions as Deterrents." University of Maryland.

Pound, E., & El-Tahri, J. (1994, October 23). "Sanctions: The Pluses and Minuses". U.S. News and World Report, p. Web.

Putnam, Robert D. (1988). "Diplomacy and Domestic Politics: The Logic of Two-Level Games." 42 (3), p. 427-460.

Thucydides. (2013). The History of the Peloponnesian War. (J. Mynott, Ed., & J. Mynott, Trans.) New York, New York: Cambridge University Press.

Walt, Stephen M. (1985). "Alliance Formation and the Balance of World Power." International Security 9 (4), p. 3-43.

Waltz, Kenneth. (2001). Man, The State, and War: A Theoretical Analysis. New York: Columbia University Press.

Endnotes

1. The appendix data referenced in this study is archived online at www.american.edu/clocksandclouds.
2. This particular post-hoc test is more useful than the other options because it adjusts the a-level in accordance with the number of comparisons being made. When making multiple comparisons, the probability of committing a Type-I error increases as the number of comparisons increases. Tukey's HSD lowers the a-level in order to take this effect into account. Tukey's HSD is thus more conservative than Fisher's Least Squared Difference, but it is not quite as conservative a test as the Bonferonni Correction.