The Political Economy of FDI 607 TABLE 6.Alternative democracy measure Variables Model 16 Model 17 Model 18 Model 19 LAGGED FDI 0.379** 0.309** 0.305** 0.308** (4.633) (3.977) (3.922) (3.962) MARKET SIZE -0.689 -0.614 -0.234 -0.608 (-1.505) (-1.330) (-0.493) (-1.278) DEVELOPMENT LEVEL 1.151* 0.947** 0.472 0.939* (2.543) (2.138) (1.022) (2.055) GROWTH 0.021** 0.023*** 0.023*** 0.023*** (2.685) (2.710) (2.778) (2.688) TRADE 0.002 0.007 0.008* 0.007 (0.317) (1.550) (1.662) (1.550) BUDGET DEFICIT -0.033** -0.024* -0.024* -0.024* (-3.017) (2.160) (-2.128) (-2.170) GOVERNMENT CONSUMPTION -0.061** -0.045** -0.043* -0.045** (-3.012) (-2.446) (-2.340) (-2.440) CAPITAL CONTROLS 0.061** (2.755) FDI INFLOWS CONTROLS -0.037 (-0.267) DICTATORSHIP -0.380*** -0.379** -0.369*** (-3.988) (-3.945) (-3.800) Time dummies Yes Yes Yes Yes Country dummies Yes Yes Yes Yes Observations 1823 1584 1568 1568 Countries 128 104 104 104 P 0.72 0.71 0.71 0.71 Note:All regressions are ordinary least squares (OLS)regressions using annual net FDI inflows as a percentage of GDP as the dependent variable. *p<.0L,*p<.05,*p<10. Table 6 presents the empirical results by replacing the Polity III measure of democracy with the ACLP measure of political regimes in the most conservative regressions-a lagged dependent variable with OLS panel-corrected standard er- rors with fixed effects.The empirical results are unaffected by this change in mea- sures of democracy.Democracies attract higher levels of FDI. The final set of empirical tests on the determinants of FDI examines the poten- tial selection effects of democracy on FDI inflows.Przeworski and colleagues find that very few poor democracies survive adverse economic conditions,leading to fewer observations of democratic governments in poor countries.54 Empirical tests that do not account for this dynamic may suffer from a potential selection bias,in our case biasing the results on the effects of democratic governance on FDI inflows. To control for these selection effects,I use a Heckman selection model.To es- timate the selection-corrected effects of democracy,in Table 7 I use the level of 64.Przeworski et al.2000.Table 6 presents the empirical results by replacing the Polity III measure of democracy with the ACLP measure of political regimes in the most conservative regressions—a lagged dependent variable with OLS panel-corrected standard er￾rors with fixed effects+ The empirical results are unaffected by this change in mea￾sures of democracy+ Democracies attract higher levels of FDI+ The final set of empirical tests on the determinants of FDI examines the poten￾tial selection effects of democracy on FDI inflows+ Przeworski and colleagues find that very few poor democracies survive adverse economic conditions, leading to fewer observations of democratic governments in poor countries+ 64 Empirical tests that do not account for this dynamic may suffer from a potential selection bias, in our case biasing the results on the effects of democratic governance on FDI inflows+ To control for these selection effects, I use a Heckman selection model+ To es￾timate the selection-corrected effects of democracy, in Table 7 I use the level of 64+ Przeworski et al+ 2000+ TABLE 6. Alternative democracy measure Variables Model 16 Model 17 Model 18 Model 19 lagged fdi 0+379*** 0+309*** 0+305*** 0+308*** ~4+633! ~3+977! ~3+922! ~3+962! market size 20+689 20+614 20+234 20+608 ~21+505! ~21+330! ~20+493! ~21+278! development level 1+151** 0+947** 0+472 0+939** ~2+543! ~2+138! ~1+022! ~2+055! growth 0+021*** 0+023*** 0+023*** 0+023*** ~2+685! ~2+710! ~2+778! ~2+688! trade 0+002 0+007 0+008* 0+007 ~0+317! ~1+550! ~1+662! ~1+550! budget deficit 20+033*** 20+024** 20+024** 20+024** ~23+017! ~2+160! ~22+128! ~22+170! government consumption 20+061*** 20+045** 20+043** 20+045** ~23+012! ~22+446! ~22+340! ~22+440! capital controls 0+061*** ~2+755! fdi inflows controls 20+037 ~20+267! dictatorship 20+380*** 20+379*** 20+369*** ~23+988! ~23+945! ~23+800! Time dummies Yes Yes Yes Yes Country dummies Yes Yes Yes Yes Observations 1823 1584 1568 1568 Countries 128 104 104 104 R2 0+72 0+71 0+71 0+71 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+ The Political Economy of FDI 607