Disrupted Economic Relationships: Disasters, Sanctions, Dissolutions

Introduction

Tibor Besedeš and Volker Nitsch

For a long time, cross-border interactions have been characterized by a strengthening of economic relationships. After the end of World War II, barriers to trade were gradually removed, and there were also massive improvements in the communication and transportation infrastructure. As a result, for most of the postwar period, international trade and investment have grown faster than output, a process that has frequently been labeled globalization.

In recent years, however, economic relationships have become more fragile. Established relationships appear to be increasingly threatened by various types of shocks. In view of the increased interdependencies between economic actors (as exemplified by the establishment of global value chains), these disruptions seem to be particularly costly and may require appropriate policy responses.

The first type of major shock with the potential to lastingly disrupt cross-border interactions comprises disasters, both natural and man-made. Natural disasters that have recently affected international economic relationships include the 2004 Indian Ocean tsunami; the 2010 eruption of the Eyjafjallajökull volcano, which grounded the flights of 10 million passengers; and the 2011 Great Tōhoku earthquake, which led to the catastrophe in Fukushima. Man-made disasters include conflicts such as the Russia-Ukraine crisis, war, and terrorism.

Another reason for the sudden disruption of economic relationships is the imposition of sanctions, which may include trade sanctions, financial sanctions, and/or asset freezes. Sanctions may be targeted against countries or against individuals or entities. According to the European Commission, sanctions are an essential tool of European Union foreign policy; most prominently, they have recently been imposed on Russia, with ongoing controversy about their impacts on both the target and the senders.

Finally, economic relationships may seriously suffer from the dissolution of existing institutional arrangements. Examples of disintegration include tendencies toward secession in several European countries (such as the United Kingdom, Spain, and Belgium) or the exit of countries from international arrangements (such as the United Kingdom’s decision to leave the European Union). Among other campaign promises, US president Donald Trump was voted into office on a promise to walk away from both negotiated and implemented trade agreements such as the Transatlantic Trade and Investment Partnership (TTIP) and the North American Free Trade Agreement (NAFTA).

In such an environment, this is a particularly timely volume. It offers a number of studies examining causes and consequences of disruptions in economic relationships. Given the breadth of issues and dimensions involved, and in view of an active body of literature on some issues, various selected aspects are addressed. Chapters 1 and 2 focus on conflicts and examine their causes theoretically. Then, chapters 3–5 empirically analyze the various effects of economic sanctions, largely motivated by the reemergence of discussions on the role of economic sanctions as a diplomatic instrument coinciding with the restrictive measures imposed on Russia in the wake of its annexation of Crimea. Next, chapter 6 examines the related issue of the consequences of a (consumer) boycott on international trade, while chapter 7 discusses the effects of natural disasters on trade. Finally, chapters 8–10 deal with disruptions, examining causes of the collapse of the Soviet Union and the duration of trade relationships. Overall, we believe that this book provides a thought-provoking (if necessarily incomplete) overview of methods and research questions on disrupted economic relationships.

On Conflict

A major impediment to cross-border economic interaction is political conflict. Consequently, a sizable amount of literature, not only in economics, examines a wide range of issues related to conflicts, from identifying their determinants to quantifying their effects.

Enrico Spolaore and Romain Wacziarg contribute to this literature by providing an interesting new perspective on the association between cultural heterogeneity and conflict. In particular, they argue that the impact of ethnic and cultural dissimilarity on conflict varies according to the types of goods over which groups are fighting. For public goods, which are characterized by nonrivalry in consumption and are commonly shared by all within a jurisdiction, heterogeneous populations are likely to experience more conflict. For private (rival) goods, in contrast, it is populations that are more similar, sharing closer preferences, that are more likely to fight with each other. Building on their earlier work, Spolaore and Wacziarg provide convincing theoretical and empirical evidence in support of their argument.

Arthur Silve and Thierry Verdier examine another interesting feature of conflicts. Based on the observation that civil wars (and, more generally, weak state institutions) often tend to cluster in time and space, they provide a theoretical framework to analyze the diffusion of conflicts. In their model, the possibility of conflict spillovers from abroad induces a regional complementarity between governments choosing their own national political regimes. In view of the resulting possibility of multiple regional political regime equilibria, an important policy lesson of their analysis, therefore, is the importance of regional institutions in coordinating national policies in order to avoid regional clusters of civil conflicts.

Sanctions, Boycotts, and Disasters

Conflicts come in different forms and intensities. Therefore, when analyzing their effects, it is essential that this heterogeneity be taken into account. In this respect, the implementation of sanctions is a particularly interesting form of cross-border dispute. While sanctions are less intense than military conflict, they imply an action and therefore go (far) beyond other diplomatic measures in order to achieve foreign policy goals.

Julian Hinz adds to the growing literature that aims to empirically assess the effects of sanctions on international trade. Applying a structural gravity framework, three recent episodes of sanctions are examined: those against Iran, Russia, and Myanmar (Burma). Overall, Hinz estimates that during these sanction regimes, trade was reduced by more than US$50 billion in 2014, or about 0.4 percent of world trade, with the bulk of this lost trade carried by a few countries.

Tristan Kohl and Chiel Klein Reesink are concerned with a related issue. While they are also interested in the effects of sanctions on trade, they differentiate between the threat of sanctions and their actual imposition. Analyzing a large sample of almost 1,500 sanction cases, they find that, in contrast to impositions, threats do not have a significant (negative) impact on international trade, despite their often extensive media coverage, which causes considerable uncertainty among economic agents.

Daniel Ahn and Rodney Ludema apply a more direct approach to assess the economic impact of sanctions. The business performance of companies targeted by sanctions is examined by analyzing sanctions against Russia. According to Ahn and Ludema’s estimates, a sanctioned company loses, on average, about one-third of its operating revenue, over one-half of its asset value, and about one-third of its employees after being targeted compared to nonsanctioned companies. Consequently, targeted sanctions seem to have a powerful impact on the targets themselves. For the overall economy, however, the effects are found to be small.

Kilian Heilmann examines another form of trade disruption, consumer boycotts. Using a natural experiment, the boycott of Danish brands in Muslim countries after the publication of the so-called Muhammad cartoons in a Danish newspaper in 2005, he examines the effects of this measure on trade in services. Heilmann shows that service trade was significantly disrupted after the boycott was announced, although the effect turned out to be short lived and was mainly concentrated in the recreational and travel service sectors.

Chenmei Li and Peter van Bergeijk offer a new, perhaps provocative, perspective on the effects of natural disasters on trade. Contrary to findings that disasters may reduce trade, it is argued that natural disasters are associated with a positive shift in the real annual growth rates of imports and exports. Plausible explanations for these findings include the need for reconstruction and the replacement of domestic production destroyed by the disaster as well as greater export orientation because of a reduction in domestic demand.

Disruptions and Duration

Sudden disruptions of existing cross-border interactions have to be analyzed in context. Recent empirical findings suggest that international trade relationships are often extremely short lived. If trade is highly flexible, however, the economic costs of disruptions may generally be limited.

Maria Persson and Wolfgang Hess examine the literature on trade duration in detail. Their survey not only illustrates that short trade durations are a very robust empirical finding, but they also explore a large set of explanatory variables that have been found to affect the duration of trade. Interestingly, Persson and Hess also show that explanatory factors that have previously been identified as relevant are not enough to explain the variation in trade survival over time.

Melise Jaud, Madina Kukenova, and Martin Strieborny focus on a specific determinant of trade duration, a country’s level of financial development. Analyzing product-level exports from ten developing countries in the Middle East and North Africa (MENA) region and sub-Saharan Africa, they argue that a significant portion of financial costs related to exports of agricultural goods emerges as a result of required compliance with sanitary and phytosanitary standards. They find that the long-term export survival of products with high export-related financial needs indeed benefits from financial development, as a well-developed financial system helps export firms establish a long-term presence in foreign markets.

Marvin Suesse examines a particularly strong form of disintegration, the dissolution of a country, using the end of the Soviet Union as a case study. After critically reviewing the hypothesis that oil played a decisive role in the collapse of the Soviet economy, alternative explanations are offered. In particular, it is argued that policy measures and territorial disintegration carry substantially greater explanatory power than oil for understanding the Soviet collapse.

Overall, the studies in this book cover a broad range of (selected) issues. By applying a variety of methods and approaches, both theoretically and empirically, it is hoped that they will provide suggestions and ideas for further research.

1

The Political Economy of Heterogeneity and Conflict

Enrico Spolaore and Romain Wacziarg

1.1    Introduction

Conflicts within and between nations are paramount sources of economic and social disruption. International wars—such as the two world wars in the twentieth century—account for some of the largest losses of lives and physical capital in human history. Conflicts that are more localized can also produce large economic and human costs, especially when they persist over time. As noted by Blattman and Miguel (2010), since 1960, over 50 percent of nations in the world have experienced internal armed conflict, and in 20 percent of them, conflict has lasted for at least 10 years, often causing extensive fatalities and displacement of entire communities. For example, the current conflict in Syria, which started in 2011, has already resulted in over 290,000 victims (International Institute for Strategic Studies 2017) and millions of refugees, with global political, social, and economic repercussions. Other areas directly affected by conflict between armed groups in recent years include Afghanistan, Burma (Myanmar), Iraq, Israel/Palestine, South Sudan, Ukraine, and many others. At the same time, terrorism and other forms of political violence have played a disruptive role in the economic and political lives of numerous societies all over the world and have affected the political debate in Europe, the United States, and elsewhere.

As the world continues to experience warfare and tensions, observers have wondered to what extent these conflicts may be linked to the heterogeneity of cultural, linguistic, and religious traits. Does diversity of cultural and ethnic origins go hand in hand with more conflict between different groups?

Social scientists have often held polarized views on this question. At one extreme is the optimistic view that historically heterogeneous populations, by interacting and cooperating with each other, can converge on common norms and values and achieve peaceful and sustainable integration. For instance, versions of this view inspired the functionalist approach to European integration (Haas 1958, 1964) as well as broader theories about communication and political cooperation across communities (Deutsch 1964). Indeed, after World War II, Europeans managed to create common institutions through peaceful integration of a growing and more dissimilar set of populations. However, the recent wave of crises and disruptions in Europe—including Britain’s vote in June 2016 to exit the European Union (Brexit) and the surge of anti-EU political movements in several countries—has challenged such optimistic assumptions, raising questions about costs and instability associated with political and cultural heterogeneity.¹

At the other extreme is the pessimistic view that ethnic and cultural dissimilarities prevent cooperation and bring about conflicts and wars. This primordialist view has a long intellectual pedigree (see, for instance, Sumner 1906) but has received renewed attention in recent decades, especially since the collapse of the Soviet Union. A well-known example of this position is Huntington’s (1993, 1996) Clash of Civilizations hypothesis, stressing religious and cultural cleavages as major sources of violent conflict since the Cold War.

While wars and conflicts are traditionally studied by historians and political scientists, there exists a growing body of literature that attempts to understand these important phenomena using the theoretical and empirical tools of contemporary political economy (for overviews, see Garfinkel and Skaperdas 2007; Blattman and Miguel 2010). In particular, recent empirical studies have focused on the relation between measures of ethnic and cultural diversity and civil conflict (e.g., Montalvo and Reynal-Querol 2005; Esteban, Mayoral, and Ray 2012; Desmet, Ortuño-Ortín, and Wacziarg 2012; Arbatli, Ashraf, and Galor 2015), while in our own work we have explored the relation between historical relatedness and international conflict (Spolaore and Wacziarg 2016a).

Motivated by findings from this empirical literature, in this chapter we present a conceptual framework that provides insights on the relation between conflict and cultural heterogeneity. Our central point is that the impact of heterogeneity should depend on whether groups are fighting over control of public goods or rival goods. Heterogeneous preferences and traits negatively affect the provision of public goods, which are nonrival in consumption and must be shared by all within a jurisdiction, whether one likes them or not. In contrast, diversity across individuals and groups comes with benefits when considering interactions about rival goods, because a diversity of preferences and cultures should be associated with lower levels of antagonism over a specific private good. In such cases, it is similarity of preferences that should bring about more conflict.

This chapter’s main idea can be illustrated with a simple example. Consider two people in a room with two sandwiches: a chicken sandwich and a ham sandwich (rival goods). People who share preferences that are more similar are more likely to want the same kind of sandwich, and possibly to fight over it, while people with preferences that are more diverse are more likely to be happy with different sandwiches. In contrast, suppose that there is a television set in the room, which both individuals must share (public good). Each can watch television without reducing the other person’s utility from watching, but they may disagree over which channel to watch and fight over the remote control. In this case, people with preferences that are more similar are less likely to fight, because they can agree on the same show.

If this distinction is relevant for understanding actual conflict, we should observe more conflict over public goods among groups that are more dissimilar but more conflict over rival goods among groups that are closer to each other in preferences, values, and cultures. In order to bring these hypotheses to the data, we must be able to measure the heterogeneity and distance between different groups and populations. Such measurements are complex and conceptually tricky, but, as already mentioned, there is now a large and growing volume of empirical literature that has made substantial progress on these issues. In our own work on this topic, we have taken a genealogical approach to heterogeneity across different populations (for a recent discussion, see, for example, Spolaore and Wacziarg 2016b). This approach is based on the idea that all human populations are related to each other but that some share common ancestors who are more recent than others, with direct implications for the extent to which they are more similar in several relevant characteristics and preferences. As different populations have split from each other over a long time, they have gradually diverged regarding sets of traits that are transmitted from one generation to another, including language, values, and norms. As a result, on average, populations with a more recent common history have had less time to diverge regarding such intergenerationally transmitted traits and tend to be more similar. Hence, we can use measures of long-term relatedness between populations to test whether heterogeneity across different groups is associated with more or less conflict among them.

Our empirical findings on heterogeneity and conflict, focused on international wars (Spolaore and Wacziarg 2016a), are consistent with the central hypothesis of this chapter. In particular, we found that, over the past two centuries, sovereign states inhabited by populations that are more closely related have been more likely to engage in violent conflict over rival goods, such as territories and natural resources (fertile soil in the nineteenth century, oil in the twentieth). On the other hand, evidence on civil conflict, such as the already mentioned contributions by Esteban, Mayoral, and Ray (2012), Desmet, Ortuño-Ortín, and Wacziarg (2012), and Arbatli, Ashraf, and Galor (2015), is broadly consistent with the hypothesis that ethnic and cultural diversity is associated with greater levels of internal violence when different groups fight over the control of public goods and policies.

The rest of this chapter is organized as follows. In section 1.2, we provide an analytical framework, capturing our main ideas about the links between intergenerational transmission of preferences, heterogeneity, and conflict over rival goods or public goods. In section 1.3, we discuss recent empirical studies of conflict that strongly support the implications of our analytical framework. Our conclusions are presented in section 1.4.

1.2    Heterogeneity and Conflict: An Analytical Framework

In this section, we present a theoretical framework linking the intergenerational transmission of preferences, genealogical distance, and the probability of conflict between populations. As far as we know, this is the first model that explicitly connects such variables within a unified formal setting. First, we model the transmission of preferences over time with variation across populations in order to explain why differences in values, norms, and preferences are linked to the degree of genealogical relatedness between populations. We show that populations that are more closely related—that is, those at a smaller genealogical distance—tend to have preferences that are more similar. Second, we model conflict over rival goods and show that conflict is more likely to arise when different populations care about the same rival goods and resources. Third, we show how the effect of relatedness on conflict changes if the dispute is about control of nonrival goods (public goods). Finally, we present a generalization of the framework, which includes conflict over rival goods and conflict over nonrival goods as special cases.

1.2.1    Intergenerational Transmission of Preferences

Our starting point is a simple model of the intergenerational transmission of preferences over the very long run. Consider three periods: o for origin, p for prehistory, and h for history. In period o, there exists only one population: population 0. In period p, the original population splits into two populations: population 1 and population 2. In period h, each of the two populations splits again into two separate populations: population 1 into population 1.1 and population 1.2, and population 2 into population 2.1 and population 2.2, as displayed in figure 1.1. In this setting, the genealogical distance dg(i, j) between population i and population j can be simply measured by the number of periods since they were one population:

and

Figure 1.1

Population tree.

These numbers have an intuitive interpretation: populations 1.1 and 1.2 are sibling populations, sharing a common parent ancestor (population 1), while populations 2.1 and 2.2 are also sibling populations, sharing a different common parent ancestor (population 2). In contrast, populations 1.1 and 2.1, for example, are cousin populations sharing a common grandparent ancestor (population 0).

For simplicity, preferences are summarized by two types (A and B). At time o, the ancestral population 0 is either of type A or of type B. For analytical convenience and without loss of generality, we assume that population 0 is of type A with probability 1/2 and of type B with probability 1/2.² Populations inherit preferences from their ancestors with variation—a population i’ descending from a population i will have preferences of the same type as their parent population i with probability µ and of the other type with probability 1 − µ.

We capture the fact that populations inherit preferences from their ancestors by assuming µ > 1/2 and capture the fact that there is variation (inheritance is not perfect) by assuming µ < 1.³ Thus, on average, populations at a smaller genealogical distance from each other will tend to be more similar in preferences. For instance, the probability that two sibling populations (say, 1.1 and 1.2) have identical types is

while the probability that two cousin populations (say, 1.1 and 2.1) have identical types is

It can be easily shown that⁴

which implies the following proposition.

Proposition 1.1    The probability that two populations are of the same type is decreasing in genealogical distance.

This result plays a key role in our analysis of conflict that follows.

1.2.2    Conflict over Rival Goods

Consider two populations (i and j), each forming a sovereign state. For simplicity, we assume that each state is a unified agent, formed by one population with homogeneous preferences.⁵

Suppose that sovereign state i is in control of a valuable prize of type t, from which it obtains the following benefits bi,

where ti* denotes state i’s ideal type, and R > 0 is the size of the prize. If the prize is of type A, t = tA, and if it is of type B, t = tB. Without loss of generality, we assume that the prize is of type A with probability 1/2 and of type B with probability 1/2. State i’s ideal type is also equal to either tA or tB. We assume that the state benefits from controlling the prize even if it is not of its favored type; that is,

The prize can be interpreted as any valuable good that can be controlled by a sovereign state—natural resources, land, cities, trade routes, colonies, protectorates, and so on (we return to the interpretation of the model later when we discuss possible extensions). Sovereign state j also values the prize, and would gain benefits if it could control the prize. State j’s benefits bj from controlling the prize are

State j can try to obtain control over the prize by challenging state i—that is, state j can take two actions: challenge state i (C) or not challenge (NC) it. If state j chooses action NC, state i keeps full control over the prize and obtains a net utility equal to bi, while state j obtains net benefits equal to 0. If state j challenges state i for possession of the prize, state i can respond either with fight (F) or not fight (NF). If state i does not fight, state j obtains control of the prize and net benefits equal to bj, while state i obtains net benefits equal to 0.

If state i decides to fight in response to the challenge, a war takes place.⁶ When a war occurs—that is, when actions {C, F} are taken—the probability that state i wins, denoted by πi, is a function of the two states’ relative military capabilities (denoted respectively by Mi and Mj),

while the probability that state j wins the war is 1 − πi. This is an instance of a contest success function of the ratio type (Hirshleifer 1989). In general, the literature on the technology of conflict assumes that the probability of success is a function of either the ratio or the difference between military capabilities (for a general discussion, see Garfinkel and Skaperdas 2007).⁷ In our model, it is the ratio.

Ex ante, each state obtains an expected utility, respectively given by

and

where ci > 0 and cj > 0 denote the respective costs of going to war. The extensive form of the game is illustrated in figure 1.2.

Figure 1.2

Extensive-form game.

It is immediate to show the following.

Lemma 1.1    War is subgame perfect equilibrium if and only if min{Ui, Uj} ≥ 0. War is the unique subgame perfect equilibrium when min{Ui, Uj} > 0.⁸

Proof    When Ui > 0 and Uj = 0, two subgame perfect equilibria exist: {C, F} and {NC, F}. When Ui = 0 and Uj > 0, there are also two subgame perfect equilibria: {C, F} and {C, NF}. When Ui = Uj = 0, three equilibria may occur: {C, F}, {C, NF}, and {NC, F}. When min{Ui, Uj} < 0, the only subgame perfect equilibria are peaceful. If Ui < 0, the only subgame perfect equilibrium is {C, NF}. If Ui > 0 and Uj < 0, the only subgame perfect equilibrium is {NC, F}. Finally, when Ui = 0 and Uj < 0, there are two (peaceful) equilibria: {NC, F} and {C, NF}. QED.

We are now ready to investigate how similarity in preferences between the two states affects the probability of war. To simplify the analysis, we assume equal capabilities (Mi = Mj = M) and costs (ci = cj = c). Let P(i, j) denote the probability of a war between state i and state j.

A war never occurs (that is, P(i, j) = 0) if each state’s expected utility from going to war is negative even when the prize is of its preferred type. This happens at a very high cost of war:

In contrast, a war always occurs (that is, P(i, j) = 1) if each state’s expected utility from going to war is positive even when the resource is not of its favored type. This happens at a very low cost of war:

Therefore, we focus on the more interesting case where war may occur with probability between 0 and 1 (that is, 0 < P(i, j) < 1), which happens when the cost of war takes on an intermediate value:⁹

Under these assumptions, a war will occur if and only if the two states have the same preferred type, and that type is equal to the type of the prize under dispute—that is, ti* = tj* = t. If the two states always had identical preferences, the probability of a war would be 1/2. This would occur, for instance, if preferences were transmitted without variation across generations: µ = 1. In contrast, if the preferences of each state were independently distributed, with each state having a 50 percent chance of preferring type A to type B (and vice versa), the probability of war would be 1/4. This would occur, for instance, if preferences were transmitted purely randomly across generations: µ = 1/2.

In general, for 1/2 <