A Mixed-Methods Approach to Examining Decentralization and Public Service Outcomes in Pakistan
[Thesis]
Ansari, Muhammad Amir
Gershberg, Alec I.
The New School
2019
446
Ph.D.
The New School
2019
Over the last three decades, many developing countries initiated ambitious decentralization programs to improve public service outcomes. But while often supported by theory, in practice decentralization may not improve service outcomes. There is a range of factors that can potentially influence decentralization policies and distort the intended outcomes. Despite the high stakes, insufficient attention has been paid to developing methodically sound analytical frameworks to empirically evaluate decentralization and the factors affecting the associated outcomes. Against this backdrop, this dissertation examines Pakistan's significant decentralization program launched in 2001 to improve public service delivery. The decentralization involved the devolution of political power along with fiscal and administrative authority to hold district governments accountable for good governance, effective delivery of services and transparent decision-making through institutionalized participation of the people at the district level. I employ a mixed-methods research design (multilevel growth modeling and in-depth semi-structured individual interviews) to explore the association between decentralization and public service outcomes of gross enrolment ratio (GER), net enrolment ratio (NER), mother health (MH), and immunization against measles (MI) at district level in Sindh Province during the period 2004-12. The results indicate that different districts do not have similar initial levels and linear growth of GER and NER, over time. Moreover, this growth does not appear to be associated with the degree of public spending on education. Districts also do not have similar initial levels and growth of MH and MI, over time. But only MH is negatively associated with public spending on health. I do not find any empirical support for a positive relationship between fiscal decentralization and public service outcomes of GER, NER, and MI. On the contrary, there is significant evidence to suggest that fiscal decentralization has a negative relationship with the outcome of MH. In addition, a modified Soufflé Theory model and district vignettes each provide an in-depth understanding of contextual factors of decentralization by suggesting that decentralization outcomes are affected by political competition and incentives, improper design of decentralization, particularly relating to fiscal autonomy, the regulatory framework, and monitoring of performance. In sum, based on practical dimension of the decentralization involving political, institutional, and fiscal aspects, the empirical findings report that (i) public service outcomes are not positively associated with the degree of sectoral public funding under fiscal decentralization; and (ii) the outcomes of decentralization are likely to be affected by the relative strength of political, intuitional, and regulatory frameworks at different spatial jurisdictions with different characteristics, such as rural and urban.