Research

“Political Mergers as Coalition Formation:  Estimates from Japanese Municipal Mergers” [pdf]

Political coalition formation games can describe the formation and dissolution of nations, as well as the creation of coalition governments, the establishment of political parties, and other similar phenomena.  These games have been studied from a theoretical perspective, but the models have not been used extensively in empirical work.  This paper presents a method of estimating political coalition formation models with many-player coalitions, and then applies this method to the recent heisei municipal amalgamations in Japan to estimate structural coefficients that describe the behaviour of municipalities.  The method enables counterfactual analysis, which in the Japanese case shows that the national government could increase welfare via a counter-intuitive policy involving transfers to richer municipalities conditional on their participation in a merger.

“Why Do We Have the Countries We Have?  Selection in the International State System”

The countries present today are only part of a larger set of potential countries. Since many modern states originated as colonies, colonial data can be used to examine correlates of independence. Empirical results support existing theories regarding efficiencies of scale and costs to heterogeneity, and in addition show that more economically successful colonies were more likely to become independent. The estimated selection model implies that analysing only currently independent countries can introduce substantial selection bias. This is shown empirically by replicating the Frankel and Romer [1999] trade-instrument regressions on selected and unselected samples. Selection bias may also explain why, despite commonly held opinions about efficiencies of scale, small countries appear to have higher GDP both in terms of growth and levels.

“Endogeneity of Linguistic Fragmentation and Implications”

Ethnic fragmentation is often treated as constant and exogenous.  In many countries, however, linguistic fragmentation has decreased a great deal over the past two centuries, and frequently used measures of ethnic fragmentation rely heavily on linguistic differences to distinguish ethnic groups.  Previous qualitative research suggests that linguistic homogenization is correlated with administrative centralization.  This hypothesis is tested empirically using the population of the largest city in each country in 1900 as a proxy for centralization, along with the population of the country as a whole and its surface area.  Using these proxies produces a statistically significant relationship between centralization and linguistic homogenization.  Furthermore, when these variables are included in regressions predicting economic growth the coefficient on fragmentation is halved and becomes statistically insignificant.  Similar results are obtained when the relationship between fragmentation and corruption is examined.