Political Decentralization and Legitimacy: Cross-Country Analysis of the Probable Influence

Political Decentralization and Legitimacy: Cross-Country Analysis of the Probable Influence Olga Gasparyan National Research University Higher School of Economics April 6, 2014

Key aspects of the research Topicality and the problematic of the research Key aspects of the research Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 2 / 19

Key aspects of the research Topicality and the problematic of the research Key aspects of the research Topicality of the research The "center-periphery"problem in the contemporary world (in the unitary as well as in the federal states). Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 2 / 19

Key aspects of the research Topicality and the problematic of the research Key aspects of the research Topicality of the research The "center-periphery"problem in the contemporary world (in the unitary as well as in the federal states). Research Question Could political decentralization affect the loss of the central government legitimacy? Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 2 / 19

Key aspects of the research Initial Hypothesis Key aspects of the research Political Decentralization could have a number of negative externalities on some socio-economic and political institutions ( federalism disillusion concept); Political Decentralization could stimulate the strength of the regional parties (D.Brancati,2008); Regional authorities turns out to be more effective on the regional and local decision-making process ( New regionalist literature). Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 3 / 19

Key aspects of the research Initial Hypothesis Key aspects of the research Political Decentralization could have a number of negative externalities on some socio-economic and political institutions ( federalism disillusion concept); Political Decentralization could stimulate the strength of the regional parties (D.Brancati,2008); Regional authorities turns out to be more effective on the regional and local decision-making process ( New regionalist literature). Initial Hypothesis: The growth of Political Decentralization could lead to the loss of the legitimacy by the national tie of government Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 3 / 19

Theoretical aspects of the research Main concepts Definitions Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 4 / 19

Theoretical aspects of the research Main concepts Definitions Decentralization It means the institutional process of power distribution between national and subnational ties of government (in administrative, political and financial sense) Legitimacy 1 Civil acceptance of the current authorities (M. Weatherford (1992) Measuring Political Legitimacy ); 2 Political activity (B. Useem, M. Useem (1979) Government Legitimacy and Political Stability ); Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 4 / 19

Theoretical aspects of the research Main concepts Definitions Decentralization It means the institutional process of power distribution between national and subnational ties of government (in administrative, political and financial sense) Legitimacy 1 Civil acceptance of the current authorities (M. Weatherford (1992) Measuring Political Legitimacy ); 2 Political activity (B. Useem, M. Useem (1979) Government Legitimacy and Political Stability ); 3 Justice, responsibility and accountability of the current government (B.Gilley (2006) The meaning and measure of state legitimacy: Results for 72 countries ) Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 4 / 19

Theoretical aspects of the research Measuring the key concepts Data Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 5 / 19

Theoretical aspects of the research Measuring the key concepts Data Decentralization the subnational expenditure share (Sourse: International Monetary Fund) Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 5 / 19

Theoretical aspects of the research Measuring the key concepts Legitimacy Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 6 / 19

Theoretical aspects of the research Measuring the key concepts Legitimacy the integral index of the legitimacy, calculated with principle component analysis using the following indicators: 1 political interest (Sourse: World Value Survey - WVS); 2 civil confidence (WVS); 3 police confidence (WVS); 4 political activity (WVS); 5 independence of the judicial branch; (Source:Cingranelli-Richards Human Rights Data Project - CIRI); Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 6 / 19

Theoretical aspects of the research Measuring the key concepts Legitimacy the integral index of the legitimacy, calculated with principle component analysis using the following indicators: 1 political interest (Sourse: World Value Survey - WVS); 2 civil confidence (WVS); 3 police confidence (WVS); 4 political activity (WVS); 5 independence of the judicial branch; (Source:Cingranelli-Richards Human Rights Data Project - CIRI); 6 extrajudicial killing (CIRI); Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 6 / 19

Theoretical aspects of the research Measuring the key concepts Legitimacy the integral index of the legitimacy, calculated with principle component analysis using the following indicators: 1 political interest (Sourse: World Value Survey - WVS); 2 civil confidence (WVS); 3 police confidence (WVS); 4 political activity (WVS); 5 independence of the judicial branch; (Source:Cingranelli-Richards Human Rights Data Project - CIRI); 6 extrajudicial killing (CIRI); 7 political imprisonement (CIRI) Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 6 / 19

Theoretical aspects of the research Data and Methods Data Time-series cross-section data (75 worldwide countries) for the period of 1981 till 2009 Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 7 / 19

Theoretical aspects of the research Data and Methods Data Time-series cross-section data (75 worldwide countries) for the period of 1981 till 2009 Methods pooled regression with robust standard errors; Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 7 / 19

Theoretical aspects of the research Data and Methods Data Time-series cross-section data (75 worldwide countries) for the period of 1981 till 2009 Methods pooled regression with robust standard errors; panel models with fixed effects with robust standard errors; Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 7 / 19

Theoretical aspects of the research Data and Methods Data Time-series cross-section data (75 worldwide countries) for the period of 1981 till 2009 Methods pooled regression with robust standard errors; panel models with fixed effects with robust standard errors; cluster analysis Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 7 / 19

Empirical Results Regression Analysis Model Pooled Regression with Robust SE Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 8 / 19

Empirical Results Regression Analysis Model Pooled Regression with Robust SE Y i = β 0 + β 1 X i + C i + ε i Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 8 / 19

Empirical Results Regression Analysis Model Pooled Regression with Robust SE где: Y i - legitimacy index; Y i = β 0 + β 1 X i + C i + ε i Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 8 / 19

Empirical Results Regression Analysis Model Pooled Regression with Robust SE где: Y i - legitimacy index; X i - level of decentralization; Y i = β 0 + β 1 X i + C i + ε i Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 8 / 19

Empirical Results Regression Analysis Model Pooled Regression with Robust SE где: Y i - legitimacy index; X i - level of decentralization; C i - controls: Y i = β 0 + β 1 X i + C i + ε i Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 8 / 19

Empirical Results Regression Analysis Model Pooled Regression with Robust SE где: Y i - legitimacy index; X i - level of decentralization; C i - controls: population; urban rate; GDP p/c PPP; infant mortality; ethnical fragmentation Y i = β 0 + β 1 X i + C i + ε i Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 8 / 19

Empirical Results Regression Analysis Results-1 Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 9 / 19

Empirical Results Regression Analysis Results-1 VARIABLES Index of Legitimacy Decentralization 0.302*** (0.052) Constant 1.210*** (0.182) Observations 1,092 R-squared 0.553 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 9 / 19

Empirical Results Regression Analysis Model Panel model with fixed effects (FE) Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 10 / 19

Empirical Results Regression Analysis Model Panel model with fixed effects (FE) Y i = β 0 + β 1 X i + C i + ε i Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 10 / 19

Empirical Results Regression Analysis Model Panel model with fixed effects (FE) где: Y i - legitimacy index; Y i = β 0 + β 1 X i + C i + ε i Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 10 / 19

Empirical Results Regression Analysis Model Panel model with fixed effects (FE) где: Y i - legitimacy index; X i - level of decentralization; Y i = β 0 + β 1 X i + C i + ε i Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 10 / 19

Empirical Results Regression Analysis Model Panel model with fixed effects (FE) где: Y i - legitimacy index; X i - level of decentralization; C i - controls: Y i = β 0 + β 1 X i + C i + ε i Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 10 / 19

Empirical Results Regression Analysis Model Panel model with fixed effects (FE) где: Y i - legitimacy index; X i - level of decentralization; C i - controls: population; urban rate; GDP p/c PPP; infant mortality; ethnical fragmentation Y i = β 0 + β 1 X i + C i + ε i Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 10 / 19

Empirical Results Regression Analysis Results-2 Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 11 / 19

Empirical Results Regression Analysis Results-2 VARIABLES Index of Legitimacy Decentralization 0.169 (0.166) Constant 1.170* (0.620) Observations 1,092 R-squared 0.038 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 11 / 19

Empirical Results Regression Analysis Discussion Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 12 / 19

Empirical Results Regression Analysis Discussion Decentralization regression coefficient is not significant on the TSCS Model with FE; Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 12 / 19

Empirical Results Regression Analysis Discussion Decentralization regression coefficient is not significant on the TSCS Model with FE; No significant effect of decentralization on central government legitimacy on the whole sample; Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 12 / 19

Empirical Results Regression Analysis Discussion Decentralization regression coefficient is not significant on the TSCS Model with FE; No significant effect of decentralization on central government legitimacy on the whole sample; Check the model -> sub-sampling Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 12 / 19

Empirical Results Model Sustainability Cluster Analysis Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 13 / 19

Empirical Results Model Sustainability Cluster Analysis Input: legitimacy, political decentralization, ethnical fragment; Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 13 / 19

Empirical Results Model Sustainability Cluster Analysis Input: legitimacy, political decentralization, ethnical fragment; Details: hierarchical cluster analysis using Wald method; Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 13 / 19

Empirical Results Model Sustainability Cluster Analysis Input: legitimacy, political decentralization, ethnical fragment; Details: hierarchical cluster analysis using Wald method; Result: 5 clusters Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 13 / 19

Empirical Results Model Sustainability Sub-sampling Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 14 / 19

Empirical Results Model Sustainability Sub-sampling Regression Analysis for each of the 5 clusters: Panel model with fixed effects (FE) Y i = β 0 + β 1 X i + C i + ε i Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 14 / 19

Empirical Results Model Sustainability Sub-sampling Regression Analysis for each of the 5 clusters: Panel model with fixed effects (FE) Y i = β 0 + β 1 X i + C i + ε i Significant coefficients were received just on two sub-samples Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 14 / 19

Empirical Results Model Sustainability Sub-sampling TSCS Regression Models Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 15 / 19

Empirical Results Model Sustainability Sub-sampling TSCS Regression Models CL 1 CL 2 CL 3 CL 4 CL 5 IL IL IL IL IL Decentr -0.324 9.206*** 0.849-4.389*** -4.857 (2.535) (2.667) (1.463) (1.302) 3.445 Const 1.921*** 1.930*** 1.827*** 1.809*** 2.040*** (50.57) (31.00) (38.24) (26.03) (17.86) Obs 217 302 270 112 164 R-sq 0.955 0.949 0.983 0.996 0.972 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 15 / 19

Empirical Results Model Sustainability Second Sub-sample Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 16 / 19

Empirical Results Model Sustainability Second Sub-sample Sub-sample of strongly centralized states, most of them are unitary ones (former Post-Soviet and Eastern European states); Positive significant effect; No regional ties of government, weak local authorities -> any power redistribution is estimated as a new level of freedom and self-governance -> stimulates the growth of the central government legitimacy. Fourth Sub-sample Federal (Canada, Switzerland, USA) and strongly decentralized states (Spain); Negative significant effect -> statistical support of the initial hypothesis; Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 16 / 19

Empirical Results Model Sustainability Second Sub-sample Sub-sample of strongly centralized states, most of them are unitary ones (former Post-Soviet and Eastern European states); Positive significant effect; No regional ties of government, weak local authorities -> any power redistribution is estimated as a new level of freedom and self-governance -> stimulates the growth of the central government legitimacy. Fourth Sub-sample Federal (Canada, Switzerland, USA) and strongly decentralized states (Spain); Negative significant effect -> statistical support of the initial hypothesis; Ethnical or religious fragmentation, resource heterogeneity, the desire of the national independence, strength of regional parties -> growth of the decentralization -> the loss of the central government legitimacy. Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 16 / 19

Conclusion Conclusion Remarks Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 17 / 19

Conclusion Conclusion Remarks Initial hypothesis is rejected on the whole sample Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 17 / 19

Conclusion Conclusion Remarks Initial hypothesis is rejected on the whole sample Three sub-samples showed the absence of any correlation between decentralization and central government legitimacy; Negative effect of decentralization is noticed only on the historical federations and strongly decentralized states Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 17 / 19

Conclusion Conclusion remarks Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 18 / 19

Conclusion Conclusion remarks Causal Inference: decentralization is stipulated by national and ethnic cleavages (the existence of regional parties) -> loss of the legitimacy is the consequence; Regional problem and the problem of the sovereignty units in decentralized states; No effect on the unitary cases -> power distribution here does not affect the legitimacy; Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 18 / 19

Conclusion Conclusion remarks Causal Inference: decentralization is stipulated by national and ethnic cleavages (the existence of regional parties) -> loss of the legitimacy is the consequence; Regional problem and the problem of the sovereignty units in decentralized states; No effect on the unitary cases -> power distribution here does not affect the legitimacy; Prerequisites to the change of the legitimacy model: from national to subnational ties of government. Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 18 / 19

Conclusion Thank you for your attention! Olga Gasparyan (HSE) MPSA Conference 2014 April 6, 2014 19 / 19