Table A1: Countries covered in the analysis (total = 116) Albania Gambia Nepal Algeria Georgia Netherlands Argentina Germany, Federal Republic of New Zealand Armenia Ghana Nicaragua Australia Greece Niger Austria Guatemala Nigeria Azerbaijan Guinea-Bissau Norway Bangladesh Guyana Panama Belarus Haiti Papua New Guinea Belgium Honduras Paraguay Benin Hungary Peru Bolivia India Philippines Botswana Indonesia Poland Brazil Iran Portugal Bulgaria Ireland Romania Burkina Faso Israel Russia Burundi Italy Senegal Cambodia Jamaica Sierra Leone Canada Japan Slovakia Central African Republic Kenya Slovenia Chile Korea, South South Africa Colombia Kyrgyzstan Spain Comoros Latvia Sri Lanka Congo, Democratic Republic of Lebanon Sudan Congo, Republic of Lesotho Sweden Costa Rica Liberia Switzerland Cote D Ivoire Lithuania Tanzania Croatia Macedonia Thailand Cyprus Madagascar Trinidad and Tobago Czech Republic Malawi Turkey Denmark Malaysia Uganda Djibouti Mali Ukraine Dominican Republic Mauritania UK Ecuador Mauritius USA El Salvador Mexico Uruguay Estonia Moldova Venezuela Fiji Mongolia Zambia Finland Mozambique Zimbabwe France Namibia Note: This table lists the countries covered in regressions (2)-(5) of Table 2A.
Table A2: List of trade agreements covered, 1957-2007 (total = 217) ACM* EC-FYROM Kyrgyz-Moldova AFTA EC-Hungary Kyrgyz-Russia AMU* EC-Iceland Kyrgyz-Ukraine Albania-Moldova EC-Israel Kyrgyz-Uzbekistan Albania-Bosnia & Herzeg. EC-Jordan LAIA* Albania-FYROM EC-Latvia Laos-Thailand* Albania- Serbia & Mont. EC-Lebanon MERCOSUR Arab Free Trade Area* EC-Lithuania MRU* Armenia-Kazakhstan EC-Malta MSG* Armenia-Moldova EC-Mexico Macedonia-Bosnia & Herzeg. Armenia-Russia EC-Morocco Mexico-Nicaragua Armenia- Turkmenistan EC-Norway Mexico-Costa Rica Armenia- Ukraine EC-Poland Mexico-Israel BAFTA EC-Romania Moldova-Croatia Bangkok* EC-Slovak Moldova-FYROM Bangkok-China* EC-Slovenia Moldova-Serbia & Mont. Bulgaria-Albania EC-South Africa Moldova- Bosnia & Herzeg. Bulgaria-Estonia EC-Switzerland Morocco-Turkey Bulgaria-Israel EC-Syria NAFTA Bulgaria-Latvia EC-Turkey New Zealand-Singapore Bulgaria-Lithuania EC-Tunisia PATCRA Bulgaria-Macedonia ECCAS* PTN* Bulgaria-Turkey ECO* Pan-Arab Free Trade CACM ECOWAS* Panama-El Salvador CACM-Chile EFTA Panama-Singapore CACM-Costa Rica* EFTA-Chile Panama-Taiwan CAFTA-Dominican Rep. EFTA-South Korea Poland-Faroe Islands CAN EFTA-Tunisia Poland-Israel CARICOM EFTA-Bulgaria Poland-Latvia CARICOM-Colombia* EFTA-Croatia Poland-Lithuania CARICOM-Costa Rica EFTA-Czech Poland-Turkey CEFTA EFTA-Estonia Rep. of Korea-Singapore CEPGL* EFTA-Finland Romania-Moldova CER EFTA-Hungary SACU CIS EFTA-Israel SADC COMESA* EFTA-Jordan SAPTA* Canada-Chile EFTA-Latvia SPARTECA* Canada-Costa Rica EFTA-Lithuania Singapore-Australia Canada-Israel EFTA-Macedonia Slovak-Estonia Chile-China EFTA-Mexico Slovak-Israel Chile-Costa Rica EFTA-Morocco Slovak-Latvia Chile-El Salvador EFTA-Poland Slovak-Lithuania Chile-Korea EFTA-Romania Slovak-Turkey Chile-Mexico EFTA-Singapore Slovenia-Bosnia & Herzeg. China-HK EFTA-Slovak Slovenia-Croatia China-Macao EFTA-Slovenia Slovenia-Estonia Croatia-Macedonia EFTA-Turkey Slovenia-Israel Croatia-Serbia & Mont. El Salvador-Mexico Slovenia-Latvia Croatia- Bosnia & Herzeg. Estonia-Faroe Islands Slovenia-Lithuania Croatia-Albania Estonia-Turkey Slovenia-Macedonia Czech-Estonia Estonia-Ukraine TRIPARTITE*
Czech-Israel GCC* Thailand-Australia Czech-Latvia GSTP* Thailand-New Zealand Czech-Lithuania Georgia-Armenia Trans-Pacific SEP Czech-Slovak Georgia-Azerbaijan Turkey-Tunisia Czech-Turkey Georgia-Kazakhstan Turkey-Bosnia & Herzeg. Dominican Rep.-Costa Rica Georgia-Russia Turkey-Croatia Dominican Rep.-El Salvador Georgia-Turkmenistan Turkey-Israel Dominican Rep.-Guatemala Georgia-Ukraine Turkey-Latvia Dominican Rep.-Honduras Guatemala-Mexico Turkey-Lithuania EAC* Honduras-Mexico Turkey-Macedonia EAEC Hungary-Estonia Turkey-Slovenia EC Hungary-Israel US-Canada EC-Albania Hungary-Latvia US-Chile EC-Algeria Hungary-Lithuania US-Israel EC-Andorra Hungary-Turkey US-Jordan EC-Bulgaria India-Sri Lanka* US-Singapore EC-Chile Japan-Malaysia United States-Australia EC-Croatia Japan-Mexico United States-Bahrain EC-Cyprus* Japan-Singapore United States-Morocco EC-Czech Jordan-Singapore WAEMU* EC-Egypt Kyrgyz-Armenia EC-Estonia Kyrgyz-Kazakhstan Data Sources: (1) WTO: http://www.wto.org/english/tratop_e/region_e/region_e.htm. (2) WTO Archive, WTO, Geneva, Switzerland. (3) Frankel, Stein and Wei (1997). (4) Schiff and Winters (2003). (5) Foreign Trade Information System: http://www.sice.oas.org/agreements_e.asp. Notes: Trade agreements followed by a * are partial-scope agreements (PTAs) and all others are free trade agreements (FTAs), as classified according to the rules explained in the main text, which are based on whether an agreement was notified to the GATT/WTO and, if so, whether it was notified under the Enabling Clause or under Article XXIV. Accession agreements to existing agreements are not displayed separately. Service agreements are not counted separately. An agreement that has entered into force before July 1 of a year is considered as effective in that year, and considered effective in the following year otherwise.
Table A3.1: Cases of transitions to autocracy (Enddemo=1) based on Polity data (total = 61) Country Name Year Country Name Year Country Name Year Albania 1996 Fiji 1987 Peru 1962 Argentina 1976 Fiji 2006 Peru 1968 Armenia 1996 Gambia 1994 Peru 1992 Azerbaijan 1993 Ghana 1972 Philippines 1972 Bangladesh 1974 Ghana 1981 Sierra Leone 1967 Bangladesh 2007 Greece 1967 Sierra Leone 1971 Belarus 1995 Guatemala 1974 Sierra Leone 1997 Benin 1963 Guinea-Bissau 1998 Sudan 1970 Brazil 1964 Guinea-Bissau 2003 Sudan 1989 Burkina Faso 1980 Guyana 1978 Thailand 1971 Cambodia 1997 Haiti 2000 Thailand 1976 Central African Republic 2003 Iran 2004 Thailand 1991 Chile 1973 Kenya 1966 Thailand 2006 Comoros 1995 Korea, South 1961 Turkey 1971 Comoros 1999 Korea, South 1972 Turkey 1980 Congo, Republic of 1963 Lesotho 1970 Uganda 1985 Congo, Republic of 1997 Lesotho 1998 Uruguay 1972 Cote D Ivoire 2002 Nepal 2002 Zambia 1968 Dominican Republic 1963 Niger 1996 Zimbabwe 1987 Ecuador 1961 Nigeria 1966 Ecuador 1970 Nigeria 1984 Notes: This table lists the cases of enddemo=1 covered by regressions (2)-(5) in Table 2A. Enddemo=1 when a democracy fails, defined as a change from a strictly positive polity2 score to a non-positive polity2 score. Table A3.2: Cases of transitions to autocracies based on Cheibub et al. s (2010) data (total = 44) Country Name Year Country Name Year Country Name Year Argentina 1962 Ghana 1972 Peru 1968 Argentina 1966 Ghana 1981 Peru 1990 Argentina 1976 Greece 1967 Philippines 1965 Bangladesh 2007 Guatemala 1963 Sierra Leone 1967 Bolivia 1980 Guatemala 1982 Sierra Leone 1997 Brazil 1964 Guinea-Bissau 2003 Sri Lanka 1977 Burundi 1996 Honduras 1963 Sudan 1969 Central African Republic 2003 Honduras 1972 Sudan 1989 Chile 1973 Korea, South 1961 Thailand 1976 Comoros 1995 Mauritania 2008 Thailand 1991 Congo, Republic of 1963 Nepal 2002 Thailand 2006 Congo, Republic of 1997 Niger 1996 Turkey 1980 Ecuador 1963 Nigeria 1966 Uganda 1985 Ecuador 2000 Nigeria 1983 Uruguay 1973 Fiji 2000 Peru 1962 Notes: This table lists the cases of enddemo=1 covered by regressions (2)-(5) in Table 2B. Enddemo=1 when a democracy fails, as defined by Cheibub at al. (2010).
Table A4: Robustness checks, enddemo regressions using import shares in 1960 (1) (2) (3) (4) (5) logit logit xtlogit Weibull PH Cox FTA_impsh60 i,t-1-7.216*** -3.510*** -3.587* -3.616** -3.318** (2.128) (1.356) (2.088) (1.542) (1.353) log(gdp/capita) I,t-1-0.485*** -0.520** -0.707*** -0.504*** (0.162) (0.225) (0.167) (0.165) DOM i,t-1-10.552** -10.405-17.058*** -70.786 (4.997) (8.466) (4.628) (0.000) FOR i,t-1-4.895*** -5.056*** -5.924*** -5.082*** (1.800) (1.659) (1.851) (1.794) war it 1.195 1.199* 1.092 0.861 (0.733) (0.653) (0.677) (0.569) war i,t-1-0.411-0.428-0.777-0.383 (0.715) (0.696) (0.718) (0.644) socialist i -0.062-0.021-0.302-0.173 (0.707) (0.866) (0.773) (0.751) Africa i 0.679 0.727 0.217 0.536 (0.477) (0.500) (0.410) (0.422) Middleeast i 0.231 0.287 0.957 0.422 (1.206) (1.157) (0.878) (1.052) Spain_colony i -0.459-0.457-0.441-0.439 (0.487) (0.477) (0.475) (0.454) UK_colony i -1.151** -1.198** -1.132** -1.155*** (0.497) (0.501) (0.450) (0.426) (M/GDP) i,t-1-0.349-0.389-0.287-0.221 (1.203) (0.974) (1.164) (1.073) duration it 0.545** 0.553 (0.226) (0.389) duration it^2-0.008** -0.009* (0.003) (0.005) country RE Yes rho 0.031 Test rho=0 (p-value) [0.390] Pseudo R2 0.062 0.217 0.111 Log Lik -276.6-204.4-204.4-83.69-197.4 Observations 2,687 2,523 2,523 2,523 2,523 Notes: Robust standard errors in parentheses, clustered by country in logit regressions. In this PH Cox regression, the standard deviation of DOM i,t-1 could not be estimated and was reported as a zero. *** p<0.01, ** p<0.05, * p<0.1
Table A5: Robustness checks, enddemo regressions with additional control variables (1) (2) (3) (4) (5 (6) logit Weibull PH Cox logit Weibull PH Cox FTA_impsh i,t-1-3.001** -3.503*** -2.968** -2.818** -3.240** -2.756** (1.303) (1.349) (1.285) (1.322) (1.373) (1.296) log(gdp/capita) I,t-1-0.539*** -0.716*** -0.539*** -0.498*** -0.676*** -0.504*** (0.156) (0.155) (0.151) (0.151) (0.156) (0.148) DOM i,t-1-11.497** -17.630*** -70.034-11.492** -17.782*** -69.820*** (5.490) (4.585) (0.000) (5.648) (4.755) (7.691) FOR i,t-1-4.036*** -5.463*** -4.417*** -3.788** -5.140*** -4.084*** (1.547) (1.613) (1.526) (1.564) (1.738) (1.542) war it 1.302* 1.281* 0.996* 1.292* 1.352* 0.983 (0.724) (0.668) (0.559) (0.762) (0.782) (0.611) war i,t-1-0.321-0.702-0.330-0.373-0.830-0.373 (0.709) (0.699) (0.617) (0.727) (0.790) (0.630) socialist i 0.142 0.235 0.077 0.063 0.090 0.028 (0.604) (0.623) (0.576) (0.586) (0.608) (0.577) Africa i 0.486 0.161 0.376 0.405 0.080 0.288 (0.407) (0.374) (0.372) (0.397) (0.348) (0.344) Middleeast i 0.196 0.891 0.327 0.247 0.959 0.418 (1.082) (0.794) (0.928) (1.159) (0.834) (1.020) Spain_colony i -0.592-0.545-0.579-0.453-0.440-0.417 (0.546) (0.506) (0.474) (0.505) (0.477) (0.452) UK_colony i -1.228** -1.263*** -1.237*** -1.204** -1.302*** -1.201*** (0.496) (0.434) (0.416) (0.506) (0.492) (0.435) (M/GDP) i,t-1-0.074 0.053-0.009-0.485-0.377-0.370 (0.989) (0.947) (0.898) (1.119) (1.072) (0.983) duration it 0.610** 0.589** (0.256) (0.262) duration it^2-0.009** -0.009** (0.004) (0.004) WTO_pre1990 i -0.326-0.236-0.348 (0.348) (0.322) (0.315) WTO_from1990 i -1.128-0.950-1.034 (0.732) (0.655) (0.657) IOnum i,t-1-0.015** -0.013-0.014** (0.007) (0.008) (0.007) former military leader it 0.074-0.029 0.127 (0.444) (0.435) (0.384) Pseudo R2 0.217 0.117 0.216 0.117 Log Lik -230.5-95.86-233.8-230.8-95.94-233.9 Observations 2,827 2,827 2,827 2,827 2,827 2,827 Notes: WTO_pre1990=1 from the year a country joined the GATT as a formal member, provided that this happened before 1990, and zero otherwise. WTO_from1990=1 from the year a country joined the GATT/WTO as a formal member, provided that this happened after 1990, and zero otherwise. IO_num measures the number of international organizations a country belongs to in a given year during 1948-2005 based on IGO database v2.3 (only full membership is considered). For years before 1965, data are available only at 5-year intervals. We assume that IOnum within two years before or after 1950 is the same as the IOnum in 1950. The same rule applied to years around 1955, 1960 and before 1965. We assign zeros to countries with missing IOnum data. Former military leader is unity if a country s chief executive is a former military leader and zero otherwise, based on data from Gandhi and Przeworski (2006). Robust standard errors in parentheses, clustered by country in logit regressions. In this PH Cox regression, the standard deviation of DOM i,t-1 could not be estimated and was reported as a zero. *** p<0.01, ** p<0.05, * p<0.1.
Table A6: Robustness checks, enddemo regressions with restricted samples Drop Socialist Countries Drop Some EU & MERCOSUR (1) (2) (3) (4) (5) (6) logit Weibull PH Cox logit Weibull PH Cox FTA_impsh i,t-1-4.829** -5.185*** -4.557** -2.929** -3.375** -2.837** (1.949) (1.993) (1.923) (1.400) (1.380) (1.309) log(gdp/capita) I,t-1-0.486*** -0.697*** -0.503*** -0.551*** -0.739*** -0.562*** (0.158) (0.159) (0.156) (0.158) (0.153) (0.148) DOM i,t-1-11.976** -16.853*** -68.352-12.018** -17.794*** -71.963 (5.480) (4.428) (0.000) (5.730) (4.641) (0.000) FOR i,t-1-4.563*** -5.851*** -4.757*** -4.653*** -5.930*** -4.964*** (1.591) (1.641) (1.546) (1.590) (1.610) (1.541) war it 1.186 1.144 0.878 1.284* 1.257* 0.973* (0.754) (0.713) (0.576) (0.678) (0.672) (0.544) war i,t-1-0.430-0.799-0.400-0.290-0.677-0.293 (0.734) (0.754) (0.650) (0.709) (0.700) (0.601) socialist i 0.449 0.093 0.305 0.246 0.266 0.192 (0.421) (0.370) (0.370) (0.599) (0.621) (0.583) Africa i 0.151 0.872 0.325 0.396 0.050 0.273 (1.165) (0.894) (1.039) (0.403) (0.359) (0.360) Middleeast i -0.470-0.417-0.438 0.242 0.975 0.431 (0.480) (0.473) (0.450) (1.179) (0.894) (1.046) Spain_colony i -1.163** -1.197*** -1.154*** -0.488-0.449-0.453 (0.474) (0.433) (0.400) (0.489) (0.471) (0.450) UK_colony i -0.351-0.138-0.205-1.206*** -1.230*** -1.221*** (1.226) (1.116) (1.045) (0.433) (0.426) (0.405) (M/GDP) i,t-1 0.612** -0.059 0.078 0.072 (0.255) (1.008) (0.982) (0.917) duration it -0.009** 0.617** (0.004) (0.262) duration it^2-0.010** -0.009** (0.005) (0.004) Pseudo R2 0.217 0.104 0.207 0.106 Log Lik -210.2-85.64-206.2-232.0-97.09-235.2 Observations 2,518 2,518 2,518 2,749 2,749 2,749 Notes: Regressions (1)-(3) drop all countries with socialist legal origin. Regressions (4)-(6) drop the following EU & MERCOSUR members: (a) Czech Rep., Estonia, Cyprus, Latvia, Lithuania, Hungary, Malta, Poland, Slovenia and Slovakia from 2004 onward; (b) Bulgaria and Romania from 2007 onward; (c) Argentina, Brazil, Paraguay, and Uruguay from 1998 onward. Robust standard errors in parentheses, clustered by country in logit regressions. In this PH Cox regression, the standard deviation of DOM i,t-1 could not be estimated and was reported as a zero. *** p<0.01, ** p<0.05, * p<0.1
Table A7: Robustness check, enddemo regressions with estimated weight on social welfare (1/b i ) Enddemo based on Polity tta based on Cheibub et al. (2010) (1) (2) (3) (4) (5) (6) logit Weibull PH Cox logit Weibull PH Cox FTA_impsh i,t-1-5.335*** -5.104** -4.780** -6.522*** -8.055*** -6.309** (1.995) (2.341) (2.023) (1.855) (2.531) (2.767) log(gdp/capita) I,t-1-0.651** -0.861*** -0.728*** -0.289-0.493-0.509 (0.283) (0.284) (0.254) (0.370) (0.337) (0.348) DOM i,t-1-7.765-9.453** -67.994-3.298** -2.607* -2.185* (6.367) (4.153) (0.000) (1.286) (1.533) (1.310) FOR i,t-1-7.962*** -7.330*** -8.028*** -3.280-4.982* -4.304 (2.120) (2.116) (2.037) (2.016) (2.734) (2.743) war it 1.659 1.505* 0.753 1.147 1.368 1.394 (1.230) (0.910) (0.490) (1.133) (1.002) (1.000) war i,t-1-0.898-1.077-0.063-0.854-0.467-0.339 (1.108) (0.899) (0.542) (1.034) (0.936) (0.973) socialist i -15.850*** -43.709-13.091*** -40.737*** (1.103) (0.000) (1.329) (1.152) Africa i 17.977*** 17.546*** 26.376*** -13.629*** -41.240*** (1.654) (1.561) (0.714) (1.512) (1.530) Spain_colony i -0.849-0.490-0.608 0.186-0.268-0.370 (0.693) (0.574) (0.576) (0.497) (0.552) (0.620) UK_colony i -18.527*** -18.895*** -27.067-2.356** -3.222*** -2.841*** (1.353) (1.040) (0.000) (1.180) (1.094) (0.991) (M/GDP) i,t-1 1.534-2.126 1.350 0.726 2.613 3.379 (3.905) (4.922) (3.314) (5.008) (4.004) (4.614) duration it 0.530* 0.211*** (0.313) (0.062) duration it^2-0.008** -0.003*** (0.004) (0.001) 1/b i -0.017-0.006-0.001-0.037-0.095-0.047 (0.038) (0.057) (0.043) (0.063) (0.104) (0.084) Pseudo R2 0.312 0.301 0.219 0.253 Log Lik -84.86-33.11-55.75-94.78-26.84-42.26 Observations 1,514 1,582 1,582 1,400 1,426 1,426 Notes: 1/b is the weight on social welfare relative to special interests estimated by Gawande et al. (2010). The middleeast dummy is dropped from these regressions because the estimated weight on social welfare (1/b) is not available for any of the Middle East countries in our sample. Robust standard errors in parentheses, clustered by country in logit regressions. In this PH Cox regression, the standard deviation of DOM i,t-1 could not be estimated and was reported as a zero. *** p<0.01, ** p<0.05, * p<0.1.
Table A8: Robustness checks, enddemo regressions with resource abundance variables (1) (2) (3) (4) (5) (6) logit Weibull PH Cox logit Weibull PH Cox FTA_impsh i,t-1-3.308** -3.985** -3.233** -5.129*** -5.583*** -4.834*** (1.628) (1.721) (1.615) (1.848) (1.967) (1.756) log(gdp/capita) I,t-1-0.543*** -0.702*** -0.541*** -0.378** -0.627*** -0.412** (0.158) (0.161) (0.154) (0.181) (0.194) (0.178) DOM i,t-1-12.871** -16.075*** -67.651*** -8.314** -15.353*** -69.789 (5.947) (4.395) (2.149) (4.119) (4.010) (0.000) FOR i,t-1-5.265*** -6.284*** -5.329*** -6.732*** -7.951*** -6.686*** (1.606) (1.666) (1.535) (2.093) (1.989) (1.663) war it 1.285* 1.234* 0.934* 1.374** 1.347** 1.001** (0.728) (0.666) (0.536) (0.682) (0.600) (0.457) war i,t-1-0.237-0.629-0.198-0.679-0.938-0.479 (0.686) (0.678) (0.587) (0.635) (0.584) (0.502) socialist i 0.844 1.090 0.839-0.271-0.308-0.407 (0.918) (0.944) (0.885) (1.404) (1.381) (1.334) Africa i -0.031-0.240-0.123 0.925 0.447 0.678 (0.428) (0.436) (0.418) (0.670) (0.518) (0.514) Middleeast i 0.539 1.282** 0.641 0.418 1.291 0.568 (1.008) (0.613) (0.845) (1.573) (1.229) (1.343) Spain_colony i -0.957* -0.809-0.885* -0.622-0.523-0.558 (0.489) (0.544) (0.494) (0.538) (0.524) (0.522) UK_colony i -1.295*** -1.295*** -1.273*** -1.861*** -1.792*** -1.721*** (0.478) (0.421) (0.395) (0.711) (0.590) (0.525) (M/GDP) i,t-1-0.222-0.023-0.053-0.324-0.294-0.116 (1.139) (1.039) (0.974) (1.911) (1.870) (1.508) duration it 0.656** 0.504** (0.272) (0.198) duration it^2-0.010** -0.008*** (0.004) (0.003) resource_rich i 0.854** 0.606 0.760** (0.401) (0.440) (0.380) OM_expsh it 0.015** 0.016** 0.013* (0.007) (0.008) (0.007) Pseudo R2 0.222 0.106 0.244 0.178 Log Lik -214.9-88.61-214.4-145.1-60.31-122.9 Observations 2,525 2,525 2,525 2,450 2,450 2,450 Notes: Resource_rich is a dummy from Albornoz et al. (2012) that is 1 if the country s main exports are based on natural resources. They construct the variable using data on country exports from the CIA World Factbook 2010. OM_expsh measures the share of ores & metals exports among total merchandise exports, based on data from the World Development Indicator 2010 of the Word Bank. Robust standard errors in parentheses, clustered by country in logit. In this PH Cox regression, the standard deviation of DOM i,t-1 could not be estimated and was reported as a zero. *** p<0.01, ** p<0.05, * p<0.1.
Table A9: Robustness checks, structural ΔFTA and ΔPTA regression using import shares in 1960 enddemo ΔFTA_impsh60 ΔPTA_impsh60 (1) (2) (3) (4) (5) hazard it 0.151** 0.187*** -0.018-0.008 (0.064) (0.064) (0.035) (0.042) FTA_impsh60 i,t-1-3.521*** -0.097*** -0.111*** (1.354) (0.024) (0.026) PTA_impsh60 i,t-1-0.980-0.093*** -0.095*** (2.436) (0.021) (0.022) log(gdp/capita) I,t-1-0.474*** 0.013 0.011-0.008-0.008 (0.160) (0.011) (0.011) (0.009) (0.009) DOM i,t-1-10.365** -0.006-0.001 (4.923) (0.021) (0.008) FOR i,t-1-4.697*** 0.264*** -0.025 (1.717) (0.087) (0.032) war it 1.175-0.007-0.009** 0.005* 0.004 (0.729) (0.006) (0.006) (0.003) (0.003) war i,t-1-0.408 0.001 0.001-0.001-0.001 (0.707) (0.005) (0.005) (0.002) (0.002) socialist it -0.083 (0.692) Africa it 0.664 (0.463) Middleeast 0.372 (1.322) UK_colony -1.097** (0.472) Spain_colony -0.445 (0.500) (M/GDP) i,t-1-0.382-0.030** -0.023* 0.016 0.016 (1.201) (0.016) (0.014) (0.016) (0.016) duration it 0.536** 0.000 0.000 (0.224) (0.001) (0.000) duration it^2-0.008** 0.000-0.000 (0.003) (0.000) (0.000) remoteness it -0.074-0.693* -0.286-0.215 (0.432) (0.414) (0.232) (0.289) L.contagion_lgdpldist it 1.735** 0.943-0.664-0.580 (0.729) (0.752) (0.412) (0.407) Year Dummies Yes Yes Yes Yes Country FE Yes Yes Yes Yes Within R-squared 0.087 0.096 0.068 0.069 Observations 2,523 2,522 2,522 2,522 2,522 Notes: Hazard is the predicted hazard based on the enddemo logit regression in the first column after setting FTA_impsh60 = PTA_impsh60 = 0. Standard errors, in parentheses, are clustered by country in the logit regression in column (1). Standard errors in the second stage (columns (2)-(5)) are corrected by bootstrapping (500 replications). *** p<0.01, ** p<0.05, * p<0.1
Table A10: Robustness checks, structural ΔFTA regression with additional controls (1) (2) (3) ΔFTA_impsh ΔFTA_impsh ΔFTA_impsh hazard it 0.207** 0.186** 0.183** (0.084) (0.077) (0.078) FTA_impsh i,t-1-0.132*** -0.132*** -0.135*** (0.018) (0.017) (0.018) log(gdp/capita) i,t-1 0.004 0.004 0.005 (0.008) (0.008) (0.008) war it -0.012** -0.011* -0.009 (0.008) (0.008) (0.007) war i,t-1 0.001 0.001 0.003 (0.006) (0.006) (0.006) (M/GDP) i,t-1-0.000 0.000-0.001 (0.019) (0.018) (0.020) remoteness it -0.345-0.338-0.365 (0.489) (0.459) (0.487) contagion_lgdpldist i,t-1 1.102 1.074 1.211* (0.659) (0.695) (0.685) reg_change i,t-1-0.014*** (0.005) var(polity)_10yr it -0.000 (0.000) polcomp i,t-1 0.003** (0.001) Year dummies Yes Yes Yes Country FE Yes Yes Yes Within R-squared 0.119 0.118 0.122 Observations 2,826 2,826 2,761 Notes: Hazard is the predicted hazard based on the enddemo logit regression in the first column in Table 6. reg_change is a dummy indicating if a country s polity score changes sign from non-positive to strictly positive, lagged by one year. var(polity)_10yr is the variance of polity scores during the last 10 years. polcomp measures the degree of general political competition in a country, as defined in Polity IV, lagged by one year. Standard errors are corrected by bootstrapping (500 replications). *** p<0.01, ** p<0.05, * p<0.1.
Table A11: Structural ΔFTA regressions with variables defined as 5-year lags/differences D5.FTA_impsh D5.FTA_impsh_5ma (1) (2) (3) (4) hazard i,t-5 0.711*** 0.658** 0.922** 0.930** (0.213) (0.263) (0.429) (0.446) FTA_impsh i,t-5-0.506*** -0.547*** -0.367*** -0.421*** (0.042) (0.040) (0.047) (0.045) log(gdp/capita) i,t-5 0.049 0.025 0.015-0.007 (0.032) (0.035) (0.036) (0.035) DOM i,t-5-0.066-0.101 (0.060) (0.080) FOR i,t-5 0.923*** 0.984*** (0.304) (0.305) war it 0.009 0.013 0.005 0.011 (0.019) (0.022) (0.017) (0.020) war i,t-5-0.042*** -0.041*** -0.051*** -0.051*** (0.010) (0.012) (0.014) (0.014) (M/GDP) i,t-5-0.053-0.035-0.080-0.035 (0.054) (0.058) (0.078) (0.082) remoteness i,t-5-1.375-3.087** -2.084-4.288** (1.464) (1.538) (1.742) (1.659) contagion_lgdpldist i,t-5 1.254-1.765 0.655-3.143 (2.061) (1.807) (2.142) (2.026) duration it 0.000 0.001 (0.002) (0.002) duration it^2-0.000-0.000 (0.000) (0.000) Year dummies Yes Yes Yes Yes Country FE Yes Yes Yes Yes R-squared 0.344 0.368 0.260 0.306 Observations 2,428 2,170 1,872 1,812 Notes: Hazard is the predicted hazard based on the enddemo logit regression in the first column in Table 6. In columns (1) and (2), the dependent variable is D5.FTA_impsh, the 5-year difference in FTA_impsh; all explanatory variables except the current war dummy and duration dependence terms are lagged by 5-years. Columns (3) and (4) are similar to columns (1) and (2) except that we calculate first the 5-year moving averages for all variables except for the current war dummy and duration dependence terms, before taking their 5-year lags or 5-year differences. Standard errors are corrected by bootstrapping (500 replications). *** p<0.01, ** p<0.05, * p<0.1.
Table A12: Dynamic panel estimation of the structure regressions, two-step system GMM (1) (2) (3) (4) FTA/CU FTA/CU PTA PTA hazard it 0.351*** 0.239* 0.033 0.003 (0.130) (0.128) (0.044) (0.070) hazard it^2-0.870* 0.142-0.337-0.184 (0.449) (0.705) (0.209) (0.155) FTA_impsh i,t-1 0.940*** 0.969*** (0.014) (0.012) PTA_impsh i,t-1 0.968*** 0.962*** (0.004) (0.005) log(gdp/capita) I,t-1 0.009** 0.009*** -0.002*** -0.001 (0.003) (0.003) (0.001) (0.002) DOM i,t-1 0.087*** -0.017 (0.030) (0.020) FOR i,t-1-0.012-0.028 (0.099) (0.051) war it -0.025*** -0.036*** 0.010*** 0.010 (0.007) (0.012) (0.003) (0.008) war i,t-1-0.013-0.004-0.001 0.000 (0.013) (0.007) (0.001) (0.003) (M/GDP) i,t-1 0.057** 0.067*** -0.027*** -0.032*** (0.024) (0.025) (0.004) (0.007) duration it -0.002*** 0.001 (0.001) (0.001) duration it^2 0.000*** -0.000 (0.000) (0.000) remote it -0.084-0.086 0.013*** 0.011 (0.087) (0.115) (0.002) (0.009) contagion_lgdpldist i,t-1 1.156** 0.325 0.000 0.000 (0.494) (0.526) (0.000) (0.000) AR(1) test (p-value) 0.000 0.000 0.000 0.037 AR(2) test (p-value) 0.285 0.559 0.815 0.821 Observations 2,826 2,826 2,826 2,826 Notes: The dependent variable is FTA_impsh or PTA_impsh. Year dummies are always included. Both the lagged FTA/PTA variables and the estimated hazard terms are treated as endogenous. We use four lags of the level variables as instruments in the differenced equations and the contemporaneous first-differences as instruments in the level equations. The p-value of the Hansen over-identification tests is always close to 1. Robust standard errors in parentheses. *** p<0.01, ** p<0.05, * p<0.1.
Figure A1: Number of free trade agreements in force & cumulative number of new democracies and of new autocracies (1948-2007), based on Polity data Notes: See notes in Appendix 2 for the data sources of Free Trade Agreements. New democracy is defined as a change from a non-positive to positive polity2 score. New autocracy is defined as a change from a positive to non-positive polity2 score. Figure A2: Histogram of the predicted hazard Density 0 10 20 30 40 0.1.2.3.4 Predicted Hazard Note: The predicted hazard is based on the estimates in column (1) of Table 3.