606 International Organization TABLE 5.Robustness tests Variables Model 13 Model 14 Model 15 LAGGED FDI 0.354** 0.367** 0.369*** (4.738) (5.119) (5.166) MARKET SIZE -0.382 0.119 (-0.630) Dropped (0.574) DEVELOPMENT LEVEL 0.509 0.219 (0.810) (0.981) Dropped GROWTH 0.024*** 0.025*** 0.026*** (2.819) (3.048) (3.147) TRADE 0.006 0.005 0.005 (1.166) (1.235) (1228) BUDGET DEFICIT -0.024* -0.023* -0.023** (-2.009) (-2.196) -2.142) GOVERNMENT CONSUMPTION -0.039** -0.037** -0.037** -2.243) -2.288) (-2.263) DEMOCRACY 0.020** 0.019** 0.019** (2.307) (2.503) (2.447) Time dummies Yes Yes Yes Country dummies Yes Yes Yes OECD included No Yes Yes Observations 122 1630 1630 Countries 93 114 114 R2 0.71 0.72 0.72 Note:All regressions are ordinary least squares (OLS)regressions using annual net FDI inflows as a percentage of GDP as the dependent variable. *p<01,*p<.05.*p<10. In models (14)and (15)in Table 5,I examine the robustness of the democracy relationship by dropping one of the two variables.Again,the empirical results on democracy remain unchanged. Another potential criticism is that these results may be driven by a particular measure of democracy.Although the Polity measure of political regimes remains the standard measure of democracy employed in most empirical studies,it is es- sentially a subjective measure.Unfortunately,all measures of democracy have some degree of subjectivity. Theoretically,the strongest measure of political regimes comes from Alvarez, Cheibub,Limongi,and Przeworski (ACLP).52 This variable codes democracies as a 0 and authoritarian regimes as a 1.This measure is in many ways a less subjec- tive measure than the Polity III variable since it uses a stricter,more minimalist definition of democracy,and is based solely on observables.63 Although this mea- sures is highly correlated with the Polity variable in my sample (-0.87),it is at a minimum an important robustness test of the link between democracy and FDI. 62.Alvarez et al.1996. 63.See Ibid.;and Przeworski et al.2000 for a more detailed discussion of the variable.In models ~14! and ~15! in Table 5, I examine the robustness of the democracy relationship by dropping one of the two variables+ Again, the empirical results on democracy remain unchanged+ Another potential criticism is that these results may be driven by a particular measure of democracy+ Although the Polity measure of political regimes remains the standard measure of democracy employed in most empirical studies, it is es￾sentially a subjective measure+ Unfortunately, all measures of democracy have some degree of subjectivity+ Theoretically, the strongest measure of political regimes comes from Alvarez, Cheibub, Limongi, and Przeworski ~ACLP!+ 62 This variable codes democracies as a 0 and authoritarian regimes as a 1+ This measure is in many ways a less subjec￾tive measure than the Polity III variable since it uses a stricter, more minimalist definition of democracy, and is based solely on observables+ 63 Although this mea￾sures is highly correlated with the Polity variable in my sample ~20+87!, it is at a minimum an important robustness test of the link between democracy and FDI+ 62+ Alvarez et al+ 1996+ 63+ See Ibid+; and Przeworski et al+ 2000 for a more detailed discussion of the variable+ TABLE 5. Robustness tests Variables Model 13 Model 14 Model 15 lagged fdi 0+354*** 0+367*** 0+369*** ~4+738! ~5+119! ~5+166! market size 20+382 0+119 ~20+630! Dropped ~0+574! development level 0+509 0+219 ~0+810! ~0+981! Dropped growth 0+024*** 0+025*** 0+026*** ~2+819! ~3+048! ~3+147! trade 0+006 0+005 0+005 ~1+166! ~1+235! ~1+228! budget deficit 20+024** 20+023** 20+023** ~22+009! ~22+196! ~22+142! government consumption 20+039** 20+037** 20+037** ~22+243! ~22+288! ~22+263! democracy 0+020** 0+019** 0+019** ~2+307! ~2+503! ~2+447! Time dummies Yes Yes Yes Country dummies Yes Yes Yes OECD included No Yes Yes Observations 1223 1630 1630 Countries 93 114 114 R2 0+71 0+72 0+72 Note: All regressions are ordinary least squares ~OLS! regressions using annual net FDI inflows as a percentage of GDP as the dependent variable+ ***p , +01, **p , +05, *p , +10+ 606 International Organization