Разнообразие, дискриминация и дипломатия: политическая экономика международной торговли в многоэтнических обществах

The high significance of results in Asia comparing with lower magnitude of results in America, Africa and Europe indicate that ethnic politics may take place in the regions in the middle of their development. For ethnic politics to have a sizeable effect the pie should be already significantly large to provoke competition and there should be institutions of large-scale ethnic mobilisation such as ethnic parties, however the effect will dampen with the further development of centralised political institutions. For this reason Asia which is between Europe and Africa on economic and political development has the highest degree of ethnic politics in foreign trade.

The only statistically significant effect in Africa is the effect of excluded groups with capacity to resist. The result is significant for import when the number of groups is used as an explanatory variable and significant for export when the size of population is used. Results for other variables such as animosity even though in the negative domain do not reach statistical significance. Furthermore, Africa is the only continent which experiences a significantly positive effect from the number of groups -- this may be caused by the importance of ethnic groups ties in the intra-African trade (i.e. Aker et al. (2014)). However, the absence of negative findings is very surprising as Africa is considered the main victim of ethnic politics.

Asia, in contrast, is the absolute pioneer of ethnic politics. There are robust and significant negative effects of both animosity between the governing and excluded group and group capacity. When the alternative measures such as the size of population are used the effect on export disappears or becomes not significant, however there is still a significant negative effect on import.

Europe also experiences significant negative effects of animosity between the governing and excluded groups. The coefficients for animosity are negative and significant for import and negative while not significant for export when the number of groups is used as an independent variable and negative and significant for both import and export when the population size is used as an independent variable. Surprisingly, Europe is the only region where the number of groups is significantly negatively correlated with trade.

Finally, in America (which accounts for both North and South Americas) the effect is similar to Europe. There are negative and statistically significant effects of animosity between the governing and excluded groups (while all coefficients are negative in the models that rely on the number of groups the effect is significant for export, while in the models that rely on the size of population the effect is significant for import). However, there is a peculiar positive effect of group capacity which is likely driven by a limited number of cases in the sample. It is noteworthy, that odd positive effect of group capacity in America in sum with negative effects in Asia and Africa makes the effect on the world level not significant.

Summing up, some adverse effect of ethnic exclusion is present in all continents, however looking closer reveals a great heterogeneity of effects. All continents bar Africa experience the negative effect of group animosity, while the effect of group capacity invisible on a world level still matters in Africa and Asia. Asia also has the biggest effect of ethnic politics overall as both animosity and group capacity on the continent have a statistically significant negative effect on both import and export. Low magnitude of effects in Africa is an important finding as it shows that at least in international trade Africa is less susceptible to ethnic politics than other regions like Asia and Europe. The differences between continents may be caused by factors varying from institutional development to the level of integration in the global trade system.

One of the hypothesis important to explore in the further research is that there is an inverted U-shape relationship between ethnic politics in trade and development. Ethnic politics has relatively lower effects in Africa, where the level of centralisation is low and the “pie” of trade is not so big. It becomes the most salient in Asia where trade is high and there is a possibility of capturing the centralised state. It again becomes lower in Europe where level of centralisation are already so high that the effect of ethnic politics declines. This theory is consistent with the long tradition in the literature showing the high vulnerability of states in the middle of political and economic development for political instability and civil wars including ethnic conflict (reviewed in Regan and Bell (2009)).

3.3 Structural factors

If the ideas regarding the adverse effect of ethnic politics on trade are correct, then there should be similar effects of primal causes of ethnic instability. One of the main ideas explored in the literature since the seminal work of Horowitz (1985) is how ethnic conflict is driven by the feeling of insecurity. When the governing group feels insecure and fears to loose power, it has more incentives to harm other groups, engage in sectarian politics and maximise group-specific and not the national benefit. In the terminology of Olson (1993), insecure groups are not stationary bandits and therefore do not engage in longterm national development.

The main channel discussed in the previous sections said that the governing group harms the trade of excluded groups due to the competition for power and sectarian politics. This channel will be supported by evidence if the insecure governing groups which have to engage in the competition for power are associated with the higher negative effect. In the two hypothesis discussed below group insecurity is estimated through two measures -- unranked ethnic systems and power-sharing regimes.

Horowitz (1985) conceptualised the distinction between ranked and unranked ethnic systems, when in the former there is a clear group hierarchy and the governing group feels secure in power while in the latter there is a high likelihood of power being contested. In the unranked ethnic systems each governing group faces a risk of being overthrown and has incentives to maximise short-term group benefit rather than long-term national development. This will result in inconsistent foreign trade policy as immediately after coming to power each group will change the terms of trade to benefit its exporters and importers and harm the development of other groups.

Consistent policy is often the feature of the highly unequal ranked systems where excluded groups never get power. The peculiar positive effects of ranked ethnic systems has already been proven in the conflict studies. For example, Vogt (2018) has shown that in the highly stratified societies there is an equilibrium of inequality, when there are less incentives and opportunities for sectarian politics.

The mechanism in trade is similar -- while the governing ethnic group in the unranked system is afraid of empowering competing groups through economic development or foreign trade promotion, the governing group in the ranked system has no such costs and is more likely to maximise the national interest. If this hypothesis is found to be correct, it will underscore the importance of ethnic diversity for the study of trade as it shows that not only formal institutions of representation matter, but structural relationships between ethnic groups.

Щ: trade is higher in the ranked ethnic systems

In the similar vein, there is a developed critique of power-sharing regimes which similarly points out that in political institutions based on ethnic cleavages there is a higher likelihood of sectarian politics. This line of thought is especially prominent in the literature on regional and ethnic parties which were shown to increase the likelihood of secessionism and ethnic conflict in Brancati (2006) and institutionalise the clientelistic sectarian system in Salloukh (2006). Horowitz (2014) has also criticised consociational power-sharing regimes by showing that they tend to degrade into dominance by one ethnicity with highly dissatisfied minority or become ossified and not sufficiently flexible.

Power-sharing regimes indicate both the insecurity of the governing ethnic group and the presence of political arrangement conducive for sectarian politics. The negative effect is likelier to be caused by stalemate in the political negotiations rather than active harm. It is important to study whether there is an effect as power-sharing is the widespread response to ethnic tensions and if it indeed has an adverse effect on trade this effect should be accounted for in the policy design.

H5: trade is lower in the presence of power-sharing arrangement

To check the hypothesis on the effect of ranked vs unranked ethnic societies I rely on the measure introduced by Vogt (2018). Ethnic Power Relations data allows to construct a count variable consecutive years of one-group dominance. This measure accounts for the length of one group rule coded as monopoly or dominant. It is important to use this measure and not just the dummy variable for dominance or monopoly as the goal is to capture the long-term hierarchy, however I also control for the dummy dominant which equals to 1 in all country-years when one group has either dominant or monopoly status and 0 otherwise.

Just using a dominant dummy in contrast to a time period of domination may capture a country with two powerful ethnic groups competing for power and dominating over each other for the short periods of time. This does not signify a presence of the ethnically ranked society and will have a detrimental effect on trade. The length of one group rule shows that the dominant group starts to develop long-term planning and is less afraid of loosing power to a competing ethnic group.

To measure the effect of power-sharing regimes I rely on variables taken directly from the Ethnic Power Relations data which distinguishes between ethnic group being a junior partner and a senior partner in the power-sharing regime. Importantly, power-sharing regimes are not only the peacemaking arrangements established after the conflicts but all the regimes where ethnicity is politically relevant and ethnic groups have access to legislative and executive power.

Examples of juniors partners are Latinos and African Americans in the United States, French speakers in Canada, Scots in the United Kingdom and Russians in Ukraine. Examples of senior partners are either groups sharing power with junior partners such as Whites in the US and English in the UK or groups engaged in full-scale power sharing regimes such as Flemings and Waloon in Belgium or Maronite Christians, Shia Muslims and Sunni Muslims in Lebanon.

To check the hypothesis about power sharing regimes, I rely on three measures. Pwr- share dummy simply takes 1 if there are junior or senior partners and not one-group dominance and zero otherwise. Junior and senior are count variables which account for the number of junior and senior partners. The logic is that the more partners there are in the power-sharing regime, the more difficult it will be to reach a consensus and the bigger will be the effect of sectarian politics.

The explanatory variables are added separately to the gravity model where the following characteristics for both countries are included: the number of groups, the log of GDP, the log of population, participation in GATT or WTO, level of democracy proxied by Polity IV index and the presence of conflict taken from EPR data. All the models are estimated using Poisson Pseudo Maximum Likelihood method by Silva and Tenreyro (2006). All models have time fixed effects, country-pair (one sided) fixed effects and continentxyear fixed effects. Standard errors are clustered three way by source country, destination country and year.

Table 4: “Structural factors”

Standard errors in parentheses and clustered three-way by exporter, importer and year Controls for population, GDP, Polity IV index and participation in GATT/WTO are included * p <0.05, ** p <0.01, *** p <0.001

The world-level findings support both hypothesis. There is a clear positive effect of ranked systems proxied through the number of consecutive year of one-group dominance. Even controlling for the dominance dummy does not change the result, therefore the main positive effect is caused by the entrenchment of one group dominance and not by dominance per se. Similarly, there is a clear negative effect from power-sharing agreements. While all coefficients are in the negative domain, the power sharing dummy effect and the number of senior partners reach statistical significance only for import, while the number of junior partners has a significant negative effect on both directions of trade flows.

The magnitude of effects is also high. Ten years of one group dominance result in the 4-5% increase in export and 3-4% increase in import completely offsetting all the negative consequences of discrimination found in the previous sections. The magnitude of power-sharing agreements is even bigger: one additional junior partner decreases export by 6.4% and import by 4.5%, while one additional senior partner decreases import by 15.2%. However, the high magnitude of effects for the number of senior partners should be interpreted accounting for the fact that a very limited number of countries has more than 1-2 senior partners in the coalition and the results may be driven by a small number of cases.

Overall these results are worrisome as they show that power sharing deals with a big number of partners tend to result either in stalemate or in sectarian politics, while one dominant group is efficient in boosting both import and export.

As the approach above showed significantly heterogeneity across continents, I repeat models 19, 21, 22 and 23 for each continent. The analysis is repeated for each continent such that each export (import) model shows the export (import) from (to) the specified continent to (from) the world. For example, again the Europe export (import) model shows the effect of not-represented groups in the export (import) from (to) 33 European countries to (from) the all 161 countries from the sample.

Splitting the sample shows that while all continents are under effect of the studied mechanisms, there is again a significant regional heterogeneity. First, Europe experiences almost no effect from ethnic insecurity -- there is no positive effect from ranked systems and the only negative effect of power-sharing regimes is the effect of junior partners which may indicate that the effect is primarily driven by stalemate in decision making rather than insecurity. In contrast, Africa and Asia are susceptible to all types of effects. This dynamics likely indicates that mechanisms of group insecurity are the feature of underinstitutionalised environments and stop being applicable with development of institutions that guarantee long-term security.

The second feature present in the results is that in all continents bar America export is more susceptible than import. This is different from the broad results for discriminated and separatist groups found in the previous section which were similar for export and import. In contrast power-sharing regimes and insecurity affect mostly export policies in most of the regions and especially in Africa. Probably, this is caused by the fact that export is more susceptible to different rent-seeking activities and exporters have more opportunities to lobby the government.

The effect of ranked systems is positive and significant for export in Africa and Asia and almost significant for export in America. The results for import are significant only in America. Europe is the only continent where ranked systems have no effect distinguishable from zero. This likely indicates that the main positive effect of ranked systems comes from eliminating group insecurity -- in Europe, where the governing group relies for security on institutions and not long-term domineering there is no positive effect, while in Africa, Americas and Asia where institutions cannot guarantee a convincing level of security, ranked ethnic systems start to play a role.

The presence of positive effect in 3 of 4 studied continents is surprising as ranked ethnic systems are rarely considered to have positive effects. Ranked systems presuppose longterm domination of one group over others and may be seen as ossified and discriminatory. However, it seems that ranked ethnic systems indeed manage to stem sectarian politics in foreign trade and crate the conditions for national foreign trade policy.

Power-sharing has a significant negative effect for all continents. In Africa powersharing agreements, the number of junior partners and the number of senior partners all have significant negative effects for export and no significant effect on import. Similarly in Asia all three variables have significant effect on export, while only the effect of the number of senior partners reaches statistical significance for import. In Europe there is the smallest negative effect of power-sharing agreements -- the only effect which reaches significance is the negative effect of junior partners on export. America is the distinct continent as it has no significant effect on export, while all three variables have significant negative effects on import.

Summing up, Asia and Africa are the most susceptible to mechanisms discussed in this section, while Europe experiences almost no effect. Export is also more susceptible than import. This is likely explained by different types of political institutions. In the regions where institutions guarantee security, ethnic dominance becomes less important for the decisions by the governing group. However, when there is an effect it mostly effects exporters where rent-seeking and lobbying is likely to be more widespread as the coalition of big exporters is more concentrated and have better access to the government than widely dispersed consumers of import.

4. Ethnic effects of Polity IV

The main empirical research of this paper is based on the Ethnic Power Relations data widely used in the studies of ethnic politics. However, it is important to check whether these findings are robust to using alternative data used not only in the literature on ethnic politics but in the wide array of works in International Relations, Political Science and Economics. Such approach would provide a robustness check for my ideas and make them of interest for a wider audience. Getting inside the Polity IV data allows to do exactly that by showing that factional characteristic of the index is negatively associated with trade. This in turn has important implications for the study of the effect of democracy on trade as factional is coded as a democratic characteristic and therefore biases the correlation between democracy and trade downward.

The cumulative Polity index is constructed by aggregating different measures such as constraints on chief executive, executive recruitment and political participation. Inside the variable categorising the competitiveness of participation, there is a category “factional” defined in Gurr and Marshall (2020) as: “Polities with parochial or ethnic-based political factions that regularly compete for political influence in order to promote particularist agendas and favour group members to the detriment of common, secular, or cross-cutting agendas.”

Factional political competition corresponds to the ethnic exclusion studied in this paper. The definition of factional almost exactly matches sectarian politics which was one of the main mechanisms discussed in the prior sections. Studying the effect of this measure is very important as it is accounted for as democratic in the Polity index (factional competition adds 1 point to the scale of democracy). If as my research predicts, factional politics have a negative effect on democracy, all the regressions of trade on democracy without accounting for factional competition produce results which are biased downward.

To check the effect of this measure, I construct a dummy factional which takes 1 if Polity IV codes the competitiveness of participation as “factional” and 0 otherwise. To check whether the effect of factional politics grows with ethnic diversity, I also create a variable groupsxfactional interacting the number of relevant groups from EPR data and factional dummy.

The same empirical procedure is conducted. The dummy and the interaction variable are added to the gravity model where the following characteristics for both countries are included: the number of groups, the log of GDP, the log of population, participation in GATT or WTO, level of democracy proxied by Polity IV index and the presence of conflict taken from the EPR data. All the models are estimated using Poisson Pseudo Maximum Likelihood method by Silva and Tenreyro (2006). All models have country-pair (one sided) fixed effects and continentxyear fixed effects. Standard errors are clustered three way by source country, destination country and year. Similarly to all previous models, these models cover 161 countries and years from 1948 to 2014.

Table 5: “The effect of factional politics”

Standard errors in parentheses and clustered three-way by exporter, importer and year Controls for the size of population and participation in GATT/WTO are included in all models * p <0.05, ** p <0.01, *** p <0.001

The results of the regression models show that the coefficient on factional is negative but not significant. Adding the interaction variable shows that the rise of the number of groups in the presence of factional politics has a significant negative effect of trade. This is especially important as by itself the number of groups has always had a positive coefficient on trade. However, the magnitude and sign of the effect of factional itself is still important compared to the Polity index which is positive but almost indistinguishable from zero. Moreover, controlling for the Polity index leads to no changes in the negative effect of factional politics.

As in the previous segments, I split the sample to see the heterogeneity on a regional level. The analysis is repeated for each continent such that each export (import) model shows the export (import) from (to) the specified continent to (from) the world. For example, again the Europe export (import) model shows the effect of not-represented groups in the export (import) from (to) 33 European countries to (from) the whole world. The coefficients on factional are reported from the model 27 (no interaction variable, but controlling for Polity IV), while the coefficients on interaction variable are reported from model 28.

Splitting the sample shows that all contents bar Africa experience a significant negative effect of factional politics. In Europe and Asia while there is no significant negative effect for factional per se, the interaction variable shows a significant negative relationship with

both import and export. In America both factional politics per se and interaction variable yields significant negative result bar the effect of interaction variable on import. Even in Africa, even though the negative effect of interaction variable does not reach statistical significance the effect is negative and almost reaches the 5% significance level.

There are peculiar positive effects of factional politics per se on European trade and African export, however these results are fully eliminated when interacting factional dummy with the number of groups. As a result, it is possible to say that in the ethnically diverse environment factional politics have a robust negative effect on both export and import. This finding based on the segment of Polity IV index supports all the findings made with the use of EPR data.

This finding is important as it contributes to the debate on trade of democracies (seminal papers are Mansfield et al. (2000) and Mansfield et al. (2002)). Polity index is one of the most widely used measures of democracy in these studies, however my findings show that factional political competition which has a negative effect on trade and is accounted for as democratic in the Polity index can bias the results downward. Factional competition is accounted for as democratic and not authoritarian in the Polity index (factional competition of political participation adds 1 to the democracy index, while repressed and suppressed are counted in the autocracy index). As a result factional politics is built in the measure of democracy and biases downward the results of research on the connection between trade and democracy.

To check whether subtracting country-years with prevalence of factional competition from the category of democracies indeed has an effect on results, I have conducted a series of models with two types of democracy dummy. The democracy variable takes 1 if Polity2 index takes from 6 to 10 which corresponds to the definition of democracy by Polity (as specified in Gurr and Marshall (2020) and Polity website). The second variable democracyalt is the same dummy subtracting all the country-years in which factional was equal to 1. As one of the main ideas in the literature is that democracies trade more with each other (i.e. Mansfield et al. (2000)), I have also created two democracy-pair variables: democracypair based on regular Polity definition of democracy and democra- cypairalt which again subtracts all the factional country-years.

There are almost no countries where there was persistent factional politics, therefore this procedure eliminates not countries but periods that are usually treated as democratic. For example, the procedure codes as undemocratic previously coded democratic periods such as Civil Rights era from 1967 to 1973 in the US; times of internal conflict in Latin America such as 1980 to 1991 in Peru or 1995 to 2014 in Columbia; tensions in Cyprus from 1960 to 1973; Turkey from 1997 to 2010; India from 1975 to 1994.

The model is the same Poisson Pseudo Maximum Likelihood estimation in which the log of GDP, the log of population, participation in GATT or WTO and the presence of conflict taken from the EPR data are included. The only difference is that there are no variables from EPR as the goal is to replicate the regular study of the effect of democracy on trade without any specific concerns for ethnicity. All models have country-pair (one sided) fixed effects and continentxyear fixed effects. Standard errors are clustered three way by source country, destination country and year. Similarly to all previous models, these models cover 161 countries and years from 1948 to 2014.

Table 6: “The effect of democracy¹

Standard errors in parentheses and clustered three-way by exporter, importer and year * p < 0.05, ** p < 0.01, *** p < 0.001

The results are striking and show the importance of paying attention to factional politics in trade research. While the regular democracy dummy is positive in all specifications it never reaches significance. However, subtracting all factional periods makes democracyalt positive and significant for export in model 35 and significant for both export and import in model 37. The negative signs on pair of democracies variables may seem confusing bearing in mind the previous findings in the literature, however as the coefficient on the trade of two democracies is democracy_o + democracy_d + democracypair the sum of these coefficients is always greater than the sum of coefficients of democratic and non-democratic country and these findings support both the idea that democracies trade more than autocracies and that democracies trade more with each other than with non-democratic countries.

The only type of models where using the alternative measure of democracies performs slightly worse than the regular measure is the pair variable without controlling for each country. While in models 38 and 39 both democracypair and democracypairalt have positive signs and do not reach significance, the former is bigger and more precise than the latter. This may be caused by the fact that factional politics mainly affects not trade of democracies with other democracies proven to be entrenched in a wide array of Preferential Trade Agreements (discussed in Mansfield et al. (2002)), but the overall trade of the country which also includes many non-democracies.

All in all, factional politics taken not from EPR but from Polity data also shows negative correlation with trade. Moreover this approach has highlighted inherent bias in the study of the connection between democracy and foreign trade -- periods of factional politics are coded as democratic and therefore bias the results downward. Subtracting periods of factionalism from democratic periods has resulted in the estimates of democratic effect becoming more precise and reaching statistical significance.

5. Discussion

The case studies and empirical models presented in the previous sections show that there is a significant effect of ethnic politics on international trade. This effect is robust and holds through different specifications. There are five points which are important to discuss regarding further research into this problem:

(1) trade is more susceptible to political conflict than broad economic development;

(2) trade suffers when groups harm each other both through active discrimination and while competing in power-sharing agreements;

(3) there are big differences in the degree of ethnic politics in different institutional and economic environments;

(4) alternative measures such as population size and the number of excluded groups lead to different results and capture different mechanisms;

(5) factional political competition which is a feature of democracies and anocracies has a negative effect on trade and should be accounted for in the studies of the effect of democracy on trade.

First, the empirical study showed that the effect of ethnicity on trade holds even after controlling for economic development -- therefore, trade is more susceptible to ethnic politics than economic development per se. This finding points out that foreign trade is less “natural” than economic activity and demands more coordination and political action. Understanding the specificity of foreign trade demands studying what exactly is the ethnic component in foreign trade negotiations and how foreign trade policy can be protected from sectarian interests.

Second, the main driver are sectarian politics. This study has found that animosity between the governing and excluded group has significant negative effect on a world level, while group's capacity to resist has significant negative effects in less developed regions such as Africa and Asia. Furthermore, the trade suffered under power-sharing regimes where there is inter-group competition and in contrast benefited from stable dominance of one group in ranked systems.

These results point out that trade is suffering from direct competition between ethnic groups for rents and imposition of trade policies motivated by sectarian and not national interests. The adverse consequences of both group discrimination and power sharing regimes with simultaneous benefits of countries dominated by one group should be taken into account by both scholars and policy-makers working on ethnic tensions and potential power-sharing agreements.

Third, splitting the sample by continents showed a significant heterogeneity of effects. My results indicate that the effect of ethnic politics is the most pronounced in Asia, where the countries are in the middle of their political and economic development. This is caused by the fact that the pie is already big enough to provoke competition and there are political institution which can be captured and controlled. In African states with low levels of trade and not developed political institutions, the effect of ethnic diversity on trade is lower. Either trade is still too undeveloped to be the object of contest or the competition is not happening though formal political power. In Europe with developed political institutions there are no negative effects from group insecurity and power-sharing regimes. This shows that in developed and centralised states, foreign policy is already very unified and is rarely the object of ethnic competition.

Fourth, the study points out the differences in measurements of ethnic politics. Measures such as ELF usually used in the studies of ethnic diversity take into account the population of ethnic groups, however limit the possibility of accounting for small but powerful groups. When the object of research is economic development, the control over resources or political power of the ethnic group is at least as important as its population size. Empirical research presented in this paper used both the number of excluded groups and the size of excluded population as the main explanatory variables to account for both the presence of ethnic group as an actor in a coordination game and the size of the ethnic group as a potential market.

Fifth, working with Polity IV data I have found that factional dummy responsible for factional political competition between parties based on parochial and ethnic-based identities has a negative effect on trade which becomes especially significant in the ethnically diverse environments. However, in the Polity index factional competition is accounted for as a democratic rather than authoritarian characteristic. Therefore, in the index of democracy often used to study interconnection between democracy and trade there is a by construction embedded characteristic which reduces trade. As this likely biases results downward, researchers in the ties between trade and democracy should account for this dynamics in their work.

Conclusion

In 1651 Thomas Hobbes wrote that the war of all against all continues on the international arena. For Hobbes each state already had a national interest expressed through the sovereign. However, unified foreign policy is not a feature, but a goal of the sovereign state. When there are multiple groups with different interests inside the state, the central government should first come up with foreign policy which accounts for all interests and then successfully communicate it to other states.

This paper illustrated the problem of disunity in foreign policy by showing that ethnic politics has an adverse effect on the foreign trade of the whole state. When groups are discriminated or engage in sectarian competition their trading potential is lost and the state trades less than similar but homogeneous counterparts. The difficulties in coordinating a unified foreign policy leads to the direct harm to the national interest.

These findings lead to direct policy implications -- specific provisions should be included in foreign trade policies and trading agreements to prevent discrimination and sectarian politics. This research also shows that this problem is the most pressing in the states in the middle of their development. Therefore, development institutions and organisations working with trade-promotion in Asian and South America should address the issue of ethnic politics.

On a theoretical level, the main contribution of this study is to broaden the discussion about ethnic diversity from domestic politics to international relations. Trade is the external public good which has to be guaranteed by the state. Moreover, it seems that foreign trade demands even more political action and coordination than internal economic development. As a result trade becomes susceptible to sectarian politics and ethnic competition. This will happen even more with the rise of interconnection as groups start to consider foreign policy not as the field of elite diplomacy but as their primary concern.

References

6. Ahmad, N. and S. Amin (2019). Does ethnic polarization stimulate or relegate trade and environmental performance? a global perspective. Environment, Development and Sustainability.

7. Aker, J. C., M. W. Klein, S. A. Oconnell, and M. Yang (2014). Borders, ethnicity and trade. Journal of Development Economics 107, 1-16.

8. Akerlof, G. A. and R. E. Kranton (2000). Economics and identity*. Quarterly Journal of Economics 115(3), 715-753.

9. Alesina, A., R. Baqir, and W. Easterly (1999, Jan). Public goods and ethnic divisions. The Quarterly Journal of Economics 114 (4), 1243-1284.

10. Alesina, A., A. Devleeschauwer, W. Easterly, S. Kurlat, and R. Wacziarg (2003). Frac- tionalization. Journal of Economic Growth 8 (2), 155-194.

11. Alesina, A. and E. L. Ferrara (2000). Participation in heterogeneous communities*. Quarterly Journal of Economics 115 (3), 847-904.

12. Aqeel, A. and M. Nishat (2004, Jan). The determinants of foreign direct investment in pakistan. The Pakistan Development Review 43(4II), 651-664.

13. Baldwin, R. and C. Magee (2000). Is trade policy for sale? congressional voting on recent trade bills. Public Choice 105, 79-101.

14. Bergstrand, J. H. (1985). The gravity equation in international trade: Some microeconomic foundations and empirical evidence. The Review of Economics and Statistics 67(3), 474.

15. Birnir, J. K. and D. M. Waguespack (2011). Ethnic inclusion and economic growth. Party Politics 17(2), 243-260.

16. Bluhm, R. and K. Thomsson (2020). Holding on? ethnic divisions, political institutions and the duration of economic declines. Journal of Development Economics 144, 102457.

17. Brancati, D. (2006). Decentralization: Fueling the fire or dampening the flames of ethnic conflict and secessionism? International Organization 60 (03).

18. Burgess, R., R. Jedwab, E. Miguel, A. Morjaria, and G. P. I. Miquel (2015). The value of democracy: Evidence from road building in kenya. American Economic Review 105 (6), 1817-1851.

19. Carballo, J., K. Handley, and N. Limao (2018). Economic and policy uncertainty: Export dynamics and the value of agreements. NBER Working Paper.

20. Deardorff, A. (1998). Determinants of Bilateral Trade: Does Gravity Work in a Neoclassic World?, pp. 7-32. The University of Chicago Press.

21. Dutt, P. and D. Mitra (2002). Endogenous trade policy through majority voting: an empirical investigation. Journal of International Economics 58(1), 107-133.

22. Easterly, W. (1999). Can institutions resolve ethnic conflict? Policy Research Working Papers.

23. Easterly, W. and R. Levine (1997, Jan). Africas growth tragedy: Policies and ethnic divisions. The Quarterly Journal of Economics 112 (4), 1203-1250.

24. Fearon, J. D. and D. D. Laitin (2000). Violence and the social construction of ethnic identity. International Organization 54 (4), 845-877.

25. Fouquin, M. and J. Hugot (2016). Two centuries of bilateral trade and gravity data: 1827-2014. CEPII Working Paper, N°2016-14.

26. Glick, R. and A. K. Rose (2002). Does a currency union affect trade? the time-series evidence. European Economic Review 46(6), 1125-1151.

27. Greif, A. (1993). Contract enforceability and economic institutions in early trade: The maghribi traders' coalition. The American economic review 83, 525-548.

28. Gurr, T. R. and M. G. Marshall (2020). Polity iv dataset user's manual,.

29. Horowitz, D. L. (1985). Ethnic groups in conflict. University of California Press.

30. Horowitz, D. L. (2014). Ethnic power sharing: Three big problems. Journal of Democracy 25(2), 5-20.

31. Humphreys, M. and J. M. Weinstein (2008). Who fights? the determinants of participation in civil war. American Journal of Political Science 52(2), 436-455.

32. Ilorah, R. (2009). Ethnic bias, favouritism and development in africa. Development Southern Africa 26 (5), 695-707.

33. Keating, M. and J. McGarry (2001). Minority nationalism and the changing international order. Oxford University Press.

34. Khalidi, R. J. and S. Taghdisi-Rad (2009). The economic dimensions of prolonged occupation: Continuity and change in israeli policy towards the palestinian economy. United Nations Conference on Trade and Development Special Report.

35. Kincaid, J. (2002). Fo...