Good Bye Chiang Kai-shek? The Long-Lasting Effects of Education under the Authoritarian Regime in Taiwan

Good Bye Chiang Kai-shek? The Long-Lasting Effects of Education under the Authoritarian Regime in Taiwan Yu Bai University of Bologna

Introduction Existing literature suggests that people s attitudes and ideology can be formed through a variety of channels: family (Bisin and Verdier, 2001), media and propaganda (Della Vigna and Kaplan, 2011), peer to peer (Sacerdote, 2001), etc. Education is among the most evident ones There are full of examples of dictators were using schooling in particular to control the outlook of children and young adults (Voigtländer and Voth, 2015)

Introduction Many countries ended the rule of authoritarian regimes between 1974 and 1990: the impacts of authoritarian regime may still exist Education: The long-term effects of socialist education on labor market outcomes and the likelihood of obtaining a college degree: Fuchs-Schündeln & Masella (2016) The short-run effects of curriculum on attitudes towards government under the authoritarian regime in China: Cantoni et al. (2017) This paper investigates whether the exposure to an authoritarian education during childhood and adolescence contributes to shaping the individual s political preference, voting participation, and national identity through the educational channel

Why Taiwan Historical background: The island of Taiwan was mainly inhabited by Taiwanese aborigines until the 17th century; During the Sino-Japanese War, the Qing ceded Taiwan to Japan in 1895; A half century later, the Republic of China (ROC) took control of Taiwan Chiang Kai-shek was the highest leader of the Kuomintang party (KMT) The contradictions led to an anti-kmt campaign around the whole island on February 28, 1947 After 1945: the explicit goal of the KMT-led authoritarian educational system was to impose government ideology on students: requiring unconditional support for the KMT; promoting patriotic education, worshiping leadership as well as anti-communism; going counteroffensive against the mainland China, etc.

Why Taiwan Historical background: The lifting of martial law was proclaimed by former Taiwanese leader Chiang Chingkuo on July 14, 1987, followed by the liberalization and democratization of Taiwan From an educational point of view, people were starting to express their dissatisfaction with the political indoctrination contents in the textbooks Overwhelming criticism about the political indoctrination in schools (Zhang, 2005). After September 1987: the schools sharply decreased the amount of patriotic activities and the Ministry of Education published revised textbooks. Most of the political indoctrination has been deleted We examine the effects of an authoritarian education on political outcomes by exploiting the sharp decrease of ideological content in school after 1987 in Taiwan

Why Taiwan

Why Taiwan

Why Taiwan

Roadmap Background Data and Methodology Results Baseline results To rule out alternative interpretations of the results Conclusions

Background: the political regime transition (Polity IV)

Background: the political regime transition 1945-1975: Chiang Kai-shek was a leader of the KMT, started to control Taiwan and perform martial law from 1949 to 1987 1975-1987: Chiang Chingkuo took over control of the government in 1975; The martial law was still valid 1987-1992: Martial law was lifted. Li Teng-hui became the first native-born Taiwanese president; People regain political rights and were free to establish political parties

Background: education Political indoctrination in the textbooks The explicit goal of the authoritarian education was to impose KMT ideology on students Distinguishing the standard teaching contents and political indoctrination in the textbooks: Lin (2000), Lau (2000), Wang (1996) and Tsai (2007)

Background: frequency of ideological contents frequencies 40 frequency of ideological contents 35 30 25 20 15 10 old textbook new textbook 5 0 Patrio.c Sen.ment Authoritarian Structure Leader Worships An.- mainland Sen.ment Authoritarian Value types of content

Data Data: Taiwan Social Change Survey, a cross-sectional survey, aimed at analyzing the social changes in Taiwan. Nine waves Sample: birth cohort (1972.1-1980.12)

Outcome variables Political preference Preference for democracy: if you think democracy is preferable to any other kind of government (1=Yes, 0=No; obs: 1382) Party affiliation: do you support for the KMT? (1=Yes, 0=No; obs: 2682) Political participation: did you turn out to vote in the most recent Taiwanese leader election? (1=Yes, 0=No; obs: 2663) National identity: which nationality do the you identify yourself as: Taiwanese or Chinese (2= Taiwanese, 1=both Taiwanese and Chinese; 0=Chinese) or (1=fully Taiwanese, 0=fully or partially Chinese; obs: 2243) Unification: if support for unification with China (2=Unification, 1=Keeping the status quo, 0=Independence) or (1=Unification, 0=Independence; obs: 2242)

Explanatory variables Treat ib : we define an individual i born in year b before September 1 as treatment group The idea: Our identification strategy exploits cut-off birth dates for school enrollment that lead to variations in the length of exposure to the authoritarian education within the same birth cohort (Fuchs-Schundeln & Masella, 2016) The difference between respondents in the treatment and control groups is that for any given birth cohort still in school when martial law was lifted off in July, 1987, respondents in the treatment group were one year more exposed to the authoritarian education system than respondents in the control group

Explanatory variables Structure of Education: Nine-year compulsory education was established in 1968: six years of primary school, three years of junior high school Cut-off birth date (1st September) for school enrollment (Spohr, 2003; Tsai, 2009; Chou et al., 2009): In Taiwan, children must be 6 years old on September 1 to enter the first grade in that year, and must be 12 years old on September 1 to enter the seventh grade

Explanatory variables

Summary Statistics: control variables All Control Treated Unconditional (Sep.-Dec.) (Jan.-Aug.) Diff. p-value Age 32.32 32.24 32.46 0.220 0.285 (5.218) (5.244) (5.143) Male 0.503 0.500 0.504 0.004 0.808 (0.500) (0.500) (0.500) Father born in Taiwan 0.894 0.899 0.891-0.008 0.540 (0.308) (0.302) (0.311) Mother born in Taiwan 0.945 0.950 0.943-0.007 0.432 (0.227) (0.218) (0.232) Father edu. (years) 8.957 9.019 8.922-0.097 0.545 (4.033) (3.978) (4.065) Mother edu. (years) 7.675 7.715 7.652-0.063 0.679 (3.810) (3.785) (3.826) Religious 0.733 0.737 0.730-0.007 0.692 (0.443) (0.441) (0.444) Urban 0.799 0.804 0.796-0.008 0.597 (0.401) (0.397) (0.403) Married 0.457 0.456 0.457 0.001 0.991 (0.498) (0.498) (0.498) Treated 0.639 0 1 1 - (0.480) #obs. 2733 996 1737 1 Columns 4 and 5 report raw (unconditional) differences in means across cohort, and the p-value for a t-test of differences in means. 2 Father edu. and Mother edu. are measured by parental years of schooling. Other control variables are dummies

Summary Statistics: outcomes variables All Control Treated Unconditional (Sep.-Dec.) (Jan.-Aug.) Diff. p-value Preference for democracy 0.571 0.610 0.551-0.059 0.048 (0.500) (0.500) (0.499) Turn-out 0.783 0.809 0.768-0.040 0.015 (0.412) (0.394) (0.422) Support for KMT 0.320 0.301 0.332 0.031 0.073 (0.460) (0.448) (0.466) National identity (0-2) 1.610 1.643 1.592-0.051 0.039 (0.532) (0.521) (0.538) Support for re-unification (0-2) 0.898 0.899 0.897 0.002 0.966 (0.621) (0.639) (0.611) 1 Columns 4 and 5 report raw (unconditional) differences in means across cohort, and the p-value for a t-test of differences in means. 2 Democracy (#obs: 1382) appears only in wave 2000, 2004, 2005, 2010 and 2015; Reunification (#obs: 2242) and Taiwanese identity (#obs: 2243) dummies show only in wave 2000, 2004, 2005, 2010, 2013, 2014, 2015; Turnout (#obs: 2682) and Support KMT (#obs: 2663) are dummies that exist in every wave

Identification Strategy: educational channel Model: Y ibpw = α + βtreat ib + η b + δ p + φ w + γx i + ɛ ibpw Y ibpw is the outcomes of individual i of birth cohort b from administrative area p, surveyed in wave w η b birth cohort fixed effects δ p administrative area fixed effects φ w wave fixed effects X i individual-level characteristics: age, gender, parental status (born in Taiwan dummy, years of schooling), religious status, residence, marital status

Baseline results Probit Model Preference for democracy Turn-out Support KMT (1) (2) (3) (4) (5) (6) Treated -0.053** -0.052** -0.050** -0.045** 0.036** 0.030* (0.016) (0.016) (0.020) (0.016) (0.017) (0.017) Birth cohort & region FE Y Y Y Y Y Y Wave FE N Y N Y N Y N 1382 1382 2682 2682 2663 2663 Figures without brackets are marginal effects, namely df/dx. Simples controls include age, gender, parental information (born in Taiwan dummy, years of schooling), religious status, residence, marital status. Significant at * 10%, ** 5%, *** 1%.

Baseline results National identity Ordered Probit Model Support for unification Chinese Both Taiwanese Unification Status quo Independence Treated 0.009** 0.038** -0.048** 0.010 0.003-0.013 (0.004) (0.015) (0.009) (0.012) (0.004) (0.015) Birth cohort & region FE Y Y Controls & Wave FE Y Y N 2243 2009 Pseudo R2 0.2021 0.1146 Figures without brackets are marginal effects, namely df/dx. Simples controls include age, gender, parental information (born in Taiwan dummy, years of schooling), religious status, residence, marital status. Significant at * 10%, ** 5%, *** 1%.

Results for subsample (respondents born between May and December) Pref. for democracy Turn-out Support KMT 1972-1980 1972-1980 1972-1980 Treated -0.058* -0.037* 0.045** (0.030) (0.020) (0.019) Birth trends (birthmonth-9) Y Y Y Birth cohort & region FE Y Y Y Controls & Wave FE Y Y Y N 1104 1901 2386 Chinese Unification 1972-1980 1972-1980 Treated 0.010** 0.003 (0.004) (0.012) Birth trends (birthmonth-9) Y Y Birth cohort & region FE Y Y Controls & Wave FE Y Y N 2066 1712 Figures without brackets are marginal effects, namely df/dx. Simples controls include age, gender, parental information (born in Taiwan dummy, years of schooling), religious status, residence, marital status. Significant at * 10%, ** 5%, *** 1%.

Results for placebo cohort groups Pref. for democracy Turn-out Support KMT 1960-1968 1960-1968 1960-1968 Treated 0.032 0.005 0.004 (0.026) (0.012) (0.018) Birth cohort & region FE Y Y Y Controls & Wave FE Y Y Y N 1382 2649 2471 Chinese Unification 1960-1968 1960-1968 Treated 0.005 0.004 (0.014) (0.027) Birth cohort & region FE Y Y Controls & Wave FE Y Y N 2139 1592 Figures without brackets are marginal effects, namely df/dx. Simples controls include age, gender, parental information (born in Taiwan dummy, years of schooling), religious status, residence, marital status. Significant at * 10%, ** 5%, *** 1%.

Results by different cut-offs Dep.Var. Pref. for democracy Turn-out March June November March June November Treated -0.001-0.046-0.012-0.013-0.012-0.019 (0.003) (0.033) (0.047) (0.019) (0.034) (0.029) Birth cohort & region FE Y Y Y Y Y Y Controls & Wave FE Y Y Y Y Y Y N 1382 1382 1382 2682 2682 2682 Dep.Var. Support KMT Chinese March June November March June November Treated 0.014 0.013 0.014 0.004 0.003 0.004 (0.020) (0.035) (0.029) (0.009) (0.009) (0.006) Birth cohort & region FE Y Y Y Y Y Y Controls & Wave FE Y Y Y Y Y Y N 2663 2663 2663 2243 2243 2243 Figures without brackets are marginal effects, namely df/dx. Simples controls include age, gender, parental information (born in Taiwan dummy, years of schooling), religious status, residence, marital status. Significant at * 10%, ** 5%, *** 1%.

Results by gender Preference for democracy Turn-out Support KMT Male Female Male Female Male Female Treated -0.038** -0.014-0.042** -0.027 0.049** 0.037 (0.037) (0.036) (0.016) (0.021) (0.022) (0.030) Birth cohort & region FE Y Y Y Y Y Y Controls & Wave FE Y Y Y Y Y Y N 705 677 1363 1319 1343 1320 Figures without brackets are marginal effects, namely df/dx. Simples controls include age, gender, parental information (born in Taiwan dummy, years of schooling), religious status, residence, marital status. Significant at * 10%, ** 5%, *** 1%. Male National identity Female Chinese Both Taiwanese Taiwanese Both Chinese Treated 0.009** 0.061** -0.060** -0.007 0.006 0.001 (0.029) (0.027) (0.026) (0.028) (0.023) (0.004) Birth cohort & region FE Y Y Controls & Wave FE Y Y N 1145 1098 Figures without brackets are marginal effects, namely df/dx. Simples controls include age, gender, parental information (born in Taiwan dummy, years of schooling), religious status, residence, marital status. Significant at * 10%, ** 5%, *** 1%.

Robustness Checks Table: Confounding factors: education and job profiles Years of Schooling Log Income Job Type Officials Professionists Workers Treated -0.197** 0.002 0.001 0.017-0.017 (0.087) (0.021) (0.015) (0.019) (0.019) Birth cohort FE Y Y Y Y Y Region FE Y Y Y Y Y Controls & Wave FE Y Y Y Y Y N 2780 2490 896 490 1024 Job type is based on ISCO-08 occupations code, which has 9 major groups. Using these 9 binary dummies directly does not alter the conclusion.figures without brackets are marginal effects, namely df/dx. Simples controls include age, gender, parental information (born in Taiwan dummy, years of schooling), religious status, residence, marital status. Significant at * 10%, ** 5%, *** 1%.

Robustness Checks Table: Results by different diplomas High school diploma Three-year college diploma Four-year college diploma above High school diploma Treated 0.020-0.015-0.020** -0.012 (0.017) (0.013) (0.008) (0.013) Birth cohort FE Y Y Y Y Region FE Y Y Y Y Controls & Wave FE Y Y Y Y N 926 723 789 2366 Figures without brackets are marginal effects, namely df/dx. Simples controls include age, gender, parental information (born in Taiwan dummy, years of schooling), religious status, residence, marital status. Significant at * 10%, ** 5%, *** 1%.

Robustness Checks Table: Confounding factors: education and job profiles Preference for democracy Turn-out Support KMT Chinese Treated -0.053* -0.042** 0.031* 0.010** (0.029) (0.019) (0.018) (0.004) Controls & Wave FE Y Y Y Y Additional controls: schooling,earnings,occupation Y Y Y Y Birth cohort & region FE Y Y Y Y N 1170 2066 2599 2210 Figures without brackets are marginal effects, namely df/dx. Simples controls include schooling, earnings, occupation age, gender, parental information (born in Taiwan dummy, years of schooling), religious status, residence, marital status. Significant at * 10%, ** 5%, *** 1%.

Confounding factors: education and job profiles

Confounding factors: education and job profiles

Confounding factors: education and job profiles

Robustness Checks Table: Subsample of individuals with a high school diploma Pref. for democracy Turn-out Support KMT Chinese 1972-1980 1972-1980 1972-1980 1972-1980 Treated -0.053* -0.037** 0.045* 0.010** (0.029) (0.020) (0.020) (0.004) Birth cohort & region FE Y Y Y Y Controls & Wave FE Y Y Y Y N 1104 1901 2386 2066 Figures without brackets are marginal effects, namely df/dx. Simples controls include schooling, income, occupation, age, gender, parental information (born in Taiwan dummy, years of schooling), religious status, residence, marital status. Significant at * 10%, ** 5%, *** 1%.

Robustness Checks Table: Birth cohort region FE Preference for democracy Turn-out Support KMT Chinese Unification Treated -0.045** -0.039* 0.040** 0.009* 0.029 (0.020) (0.021) (0.020) (0.005) (0.028) Controls & Wave FE Y Y Y Y Y Birth cohort & region FE Y Y Y Y Y Birth cohort region FE Y Y Y Y Y N 1382 2682 2663 2021 2009 Figures without brackets are marginal effects, namely df/dx. Simples controls include age, gender, parental information (born in Taiwan dummy, years of schooling), religious status, residence, marital status. Significant at * 10%, ** 5%, *** 1%.

Confounding factors: education and job profiles

Choose different cutoffs rather than 1987 Preference for democracy Turn-out Support KMT Chinese (1988) (1989) (1988) (1989) (1988) (1989) (1988) (1989) Treated -0.046* -0.033-0.044** -0.042** 0.030* 0.028* 0.007** 0.005* (0.028) (0.030) (0.028) (0.016) (0.017) (0.016) (0.003) (0.003) Birth cohort & region FE Y Y Y Y Y Y Y Y N 1390 1323 2130 2082 2667 2661 2275 2261 Figures without brackets are marginal effects, namely df/dx. Simples controls include age, gender, parental information (born in Taiwan dummy, years of schooling), religious status, residence, marital status. Significant at * 10%, ** 5%, *** 1%.

Conclusions This paper sheds light on the long-run impacts of political regime by exploiting evidence from the variation of the ideological content in schools We find that one additional year of exposure to authoritarian education has significant effects on political outcomes Our results persist after ruling out alternative interpretations and are robust to the sensitivity tests and different groups definitions

Thank you for your kind attention!