Decentralization, Democratization, and Public Service Delivery Günther Schulze and Bambang Suharnoko Sjahrir University of Freiburg, Germany Indonesia Update Conference 2013 Canberra, 20-21 September 2013
Outline Introduction Decentralization and Democratization in Indonesia Analytical approach Research questions Approach PSD input Governance / Technology PSD output Concluding remarks
Decentralization in Indonesia Indonesia underwent a big-bang decentralization and rapid democratization Start of direct elections of district heads Major events Political/fiscal/administrative decentralization Democratic elections: legislative power to local parliaments Suharto stepped down 1993 1998 1999 2001 2004 2010 Year
The democratization process The timing of democratically indirect and direct elections was fairly exogenous No. districts 0 100 200 300 400 500 1993 1997 1999 2001 2004 2007 2010 Year Appointed district heads Indirectly elected district heads Directly elected district heads
Districts gained fiscal authority Local sectoral expenditure p.c. increased considerably Average Level ln Mean p.c. expenditure 7 8 9 10 11 12 13 INFR. EDUC. HEALTH 1994 1997 20002001 2004 2007 2009 Years Source: SIKD, Min. of Finance
Public service delivery did not improve much over time Average Level PSD Level.1.2.3.4.5.6.7.8.9 1 INFRA HEALTH EDU 1994 1997 20002001 2004 2007 2009 Years Source: Susenas, Podes (BPS)
Central questions Have things become better for the people? Better governance? Better services? Better preference matching, higher transparency and participation or local capture? Heterogeneity of districts: convergence or divergence? Methodological challenge: missing counterfactual
This paper Reports on the existing large scale empirical research by Bambang Suharnoko Sjahrir, joint with Krisztina Kis-Katos and Günther G. Schulze. Panel dataset: District level, 418 out of 477, excl. NAD, Papua and DKI Balanced and unbalanced, from 1994 to 2009 Fiscal variables for 3 sectors (MoF) Levels of public physical infrastructure (BPS: Susenas, Podes) Ratio of jun. secondary schools to jun. School-aged children Ratio of health clinics (Puskesmas) to population Share of villages with paved roads Timing of local elections (exogeneously determined), national elections and local parliamentary compositions (KPU, MoHA, TAF, WB) Other controls (BPS, MoF) 8
The analytical approach Three complementary approaches to measure the effect of institutional change on PSD PS (health, edu, infra.) Output Has PSD output improved? Technology/ Governance Has the process of PSD become more efficient/ less wasteful? Input Have inputs become more correlated with PSD level? 9
Input : public expenditure (1) Source: Kis-Katos and Sjahrir (2013) RQ : Had DEC and DEMO an effect on expenditure patterns for core services? Method : balanced panel 1994-2009, SUR FE on health, education and infrastructure Results : District spends more if its PSD level is relatively low after decentralization No systematic changes with the democratization process
Governance/technology : Administrative spending (1) Government administrative spending is relatively high p.c. Public expenditures (in '000 Rupiah) 0 500 1000 1500 Govt. Administrative Health Education Infrastructure Agriculture Other sectors 2001 2002 2003 2004 2005 2006 2007 2008 2009 Years Source : Sjahrir et al. (2013)
Governance/technology: Administrative spending (2) Source : Sjahrir, Kis-Katos and Schulze (2013) RQ : what are the determinants of administrative spending? Method : unbalanced panel 2001-2009, Pooled OLS, FE Results : Splitting is not the main reason, lack of accountability is Democratic accountability not (yet) effective Higher in districts with lower political competition
Output : PSD level (1) Source: Schulze and Sjahrir (2013) RQ : Convergence or divergence? Have DC and DEMO altered PSD? Determinants of PSD Method : unbalanced panel 1994-2009 SUR FE on health, education and infrastructure Synthetic counterfactual: prior trend
Output : PSD level (2) Convergence in secondary education Growth of jnr. secondary schools ratio 1994-2000(in percentage point) -1 0 1 2 3 Growth of jnr. secondary schools ratio 2001-2009(in percentage point) -1 0 1 2 3 0.2.4.6.8 1 0.2.4.6.8 No. of junior secondary schools (jsc) per 100 jsc-aged Children No. of 1994 junior secondary schools (jsc) per 100 jsc-aged Children 2001 Source : Schulze and Sjahrir (2013)
Output : PSD level (3) But not in health Growth of puskesmas ratio 1994-2000(in percentage point) -1 0 1 2 Growth of puskesmas ratio 2001-2009(in percentage point) -1 0 1 2 0.5 1 1.5 No. of health clinics (Puskemas) per 10000 people 1994 0.5 1 1.5 No. of health clinics (Puskemas) per 10000 people 2001 Source : Schulze and Sjahrir (2013)
Output : PSD level (4) And only slightly in infrastructure Growth of Sh. villages w/ paved roads 1994-2000(in percentage point) -1 0 1 2 3 Growth of Sh. villages w/ paved roads 2001-2009(in percentage point) -1 0 1 2 3 0.2.4.6.8 1 Share of villages w/ paved roads 1994 0.2.4.6.8 1 Share of villages w/ paved roads 2001 Source : Schulze and Sjahrir (2013)
Output: PSD level (5) Results Convergence. Lagging districts catch up Money matters. Richer districts provide better services Decentralization tends to make things better No uniform effects of direct election: depends on financial endowment. Source: Schulze and Sjahrir (2013)
Concluding remarks First systematic econometric approach to measure the effects of decentralization and democratization on PSD Covers the first decade Democratic accountability still insufficient DEC and DEMO have non-uniform effects, but overall tend to be slightly positive Effects may still unfold The best may yet to come
Thank you very much for your attention