American Political Science Review Vol.104.No.3 TABLE 5.Different Samples Portfolio FDI Rich→Rich Rich→Poor Rich→Rich Rich→Poor Log(migrant stock from d in s) 0.406* 0.231* 0.392* 0.245* (0.123) (0.0788) (0.165) (0.101) Log(product of GDPs) -0.0162 0.195* 0.123* 0.2414 (0.0476) (0.0970) (0.0726) (0.0770) Log(distance) -0.126 -1.309* -0.989* -1.453* (0.288) (0.268) (0.337 (0.464) Common border -0.250 -1.279 0.207 2.393 (0.357) (1.210) (0.718) (0.812) Official common language -0.292 0.189 -0.332 2.756 (0.287) (0.430) (0.698) (0.824) Correlation of growth rates -0.484 0.296 -0.00821 0.111 (0.449) (0.253) (0.530) (0.315) Common exchange rate peg 0.298 0.434 -0.0290 0.0577 (0.364) (0.649) (0.755) (0.747) Dual taxation treaty 1.040* 0.510 0.706 0.455 (0.415) (0.376) (0.553) (0.411) Preferential trade agreement 1.404* 1.368* 0.794 0.466 (0.683) (0.599) (1.054) (0.533) Common legal heritage 0.582* -0.160 0.932* 0.196 (0.325) (0.263) (0.496) (0.431) Common dominant religion 0.208 0.679 0.186 0.132 (0.195) (0.421) (0.531) (0.361) Genetic distance 0.000379 0.000623 0.0000191 -0.0000864 (0.000370) (0.000547) (0.000825) (0.000632) Log(bilateral telephone volume) 0.0712 0.0132 0.0620 -0.104 (0.144) (0.0852) (0.360) (0.140) Constant 23.54 -127.4* -96.81 -159.9* (41.11) (67.17) (64.62) (53.22) Observations 796 2.597 805 1,536 Adjusted R2 0.830 0.722 0.583 0.620 FDI,foreign direct investment;GDP,gross domestic product. Robust standard erors in parentheses. Dependent variable:log(portfolio investment)in columns 1 and 2;log(FDI)in columns 3 and 4. Robust standard errors clustered by both origin and destination country. All models include both origin and destination dummy variables. p<.10;*wp<.05. countries that are likely to have greater transactions already documented that migrant laborers remit a substan- costs,higher barriers to entry,and larger information tial amount of capital (Leuth and Ruiz-Arranz 2006;Ratha asymmetries. and Shaw 2007).Our findings suggest that migrant-driven investment is yet another way in which diaspora communities influence developments in their home countries.That these CONCLUSION AND DIRECTIONS two flows of capital likely have different effects on inequality FOR FUTURE RESEARCH and poverty provides an interesting avenue for future work connecting diasporas to development. Access to international capital markets is a perennial prob- There is still quite a bit we do not know.Is migrant- lem confronted by all countries.Students of international led investment countercyclical?When a destination expe- political economy have invested considerable time in try- riences a shock-a natural disaster or a financial crisis-is ing to understand the theoretical and empirical connections migrant-led investment more stable than traditional invest- across countries and markets.One general conclusion from ment channels?28 What about the effects of migrant networks these efforts is that information asymmetries represent a on other channels ofinvestment such as FDI and cross-border large cost to cross-border economic transactions.We have mergers and acquisitions behavior?Connecting these pro- demonstrated that migrant networks-connections between cesses will help us better understand institutional and nonin- coethnics across countries-play an important role in decreas- stitutional determinants of global investment. ing asymmetries and,consequently,promote both portfolio and FDI. 28 The countercyclicality of remittances also has important implica- How do these findings square with the extant literature tions for our understanding of exchange rate regime choice (Singer on immigration and on capital flows?Recent studies have 2010). 595American Political Science Review Vol. 104, No. 3 TABLE 5. Different Samples Portfolio FDI Rich → Rich Rich → Poor Rich → Rich Rich → Poor Log(migrant stock from d in s) 0.406∗∗ 0.231∗∗ 0.392∗∗ 0.245∗∗ (0.123) (0.0788) (0.165) (0.101) Log(product of GDPs) −0.0162 0.195∗∗ 0.123∗ 0.241∗∗ (0.0476) (0.0970) (0.0726) (0.0770) Log(distance) −0.126 −1.309∗∗ −0.989∗∗ −1.453∗∗ (0.288) (0.268) (0.337) (0.464) Common border −0.250 −1.279 0.207 2.393∗∗ (0.357) (1.210) (0.718) (0.812) Official common language −0.292 0.189 −0.332 2.756∗∗ (0.287) (0.430) (0.698) (0.824) Correlation of growth rates −0.484 0.296 −0.00821 0.111 (0.449) (0.253) (0.530) (0.315) Common exchange rate peg 0.298 0.434 −0.0290 0.0577 (0.364) (0.649) (0.755) (0.747) Dual taxation treaty 1.040∗∗ 0.510 0.706 0.455 (0.415) (0.376) (0.553) (0.411) Preferential trade agreement 1.404∗∗ 1.368∗∗ 0.794 0.466 (0.683) (0.599) (1.054) (0.533) Common legal heritage 0.582∗ −0.160 0.932∗ 0.196 (0.325) (0.263) (0.496) (0.431) Common dominant religion 0.208 0.679 0.186 0.132 (0.195) (0.421) (0.531) (0.361) Genetic distance 0.000379 0.000623 0.0000191 −0.0000864 (0.000370) (0.000547) (0.000825) (0.000632) Log(bilateral telephone volume) 0.0712 0.0132 0.0620 −0.104 (0.144) (0.0852) (0.360) (0.140) Constant 23.54 −127.4∗ −96.81 −159.9∗∗ (41.11) (67.17) (64.62) (53.22) Observations 796 2,597 805 1,536 Adjusted R2 0.830 0.722 0.583 0.620 FDI, foreign direct investment; GDP, gross domestic product. Robust standard errors in parentheses. Dependent variable: log(portfolio investment) in columns 1 and 2; log(FDI) in columns 3 and 4. Robust standard errors clustered by both origin and destination country. All models include both origin and destination dummy variables. ∗ p < .10; ∗∗ p < .05. countries that are likely to have greater transactions costs, higher barriers to entry, and larger information asymmetries. CONCLUSION AND DIRECTIONS FOR FUTURE RESEARCH Access to international capital markets is a perennial problem confronted by all countries. Students of international political economy have invested considerable time in trying to understand the theoretical and empirical connections across countries and markets. One general conclusion from these efforts is that information asymmetries represent a large cost to cross-border economic transactions. We have demonstrated that migrant networks—connections between coethnics across countries—play an important role in decreasing asymmetries and, consequently, promote both portfolio and FDI. How do these findings square with the extant literature on immigration and on capital flows? Recent studies have already documented that migrant laborers remit a substantial amount of capital (Leuth and Ruiz-Arranz 2006; Ratha and Shaw 2007). Our findings suggest that migrant-driven investment is yet another way in which diaspora communities influence developments in their home countries. That these two flows of capital likely have different effects on inequality and poverty provides an interesting avenue for future work connecting diasporas to development. There is still quite a bit we do not know. Is migrantled investment countercyclical? When a destination experiences a shock—a natural disaster or a financial crisis—is migrant-led investment more stable than traditional investment channels?28 What about the effects of migrant networks on other channels of investment such as FDI and cross-border mergers and acquisitions behavior? Connecting these processes will help us better understand institutional and noninstitutional determinants of global investment. 28 The countercyclicality of remittances also has important implications for our understanding of exchange rate regime choice (Singer 2010). 595