The impact of fiscal decentralization on infant mortality rates: Evidence from OECD countries

Abstract

This study re-examines the hypothesis that shifts towards more decentralization would be accompanied by improvements in population health on a panel of 20 OECD countries over a thirty year period (1970–2001). Decentralization is proxied using a conventional indicator of revenue decentralization and a new measure of fiscal decentralization that reflects better than previous measures the existence of autonomy in the decision-making authority of lower tiers of government, a crucial issue in the decentralization process. The results show a considerable and positive effect of fiscal decentralization on infant mortality only if a substantial degree of autonomy in the sources of revenue is devolved to local governments. The proportion of health care expenditure on GDP and, in particular, education, were found to have a larger contribution to the reduction of infant mortality in the sample of OECD countries analysed over the period of study.

Introduction

Despite the large number of countries worldwide that are increasingly devolving responsibilities for their health care systems to local levels of government, there is insufficient evidence about the potential impact of such reforms. To date much of the literature on this topic has concentrated on theoretical discussions about the potential advantages and disadvantages of transferring decision-making and economic resources from central to local levels of government.

However, in absence of quantitative measures of the magnitude of the effect of decentralization, there is little that can be said in terms of its benefits and its costs to the health system. The most appropriate level of decentralization of health services is therefore a central policy issue that to date remains largely unresolved.

Many of the proponents of decentralization claim that decentralization could strengthen accountability of policy makers to local electors and transparency, thus allowing a better tailoring between the provision of goods and services and local needs (Oates, 1999). With respect to the health services in particular, decentralization is expected to improve access to health care services, and ultimately, population health (Khaleghian, 2004, Robalino et al., 2001, Uchimura and Jütting, 2009).

However, in spite of the identified advantages of devolution of policy making in health care services, there is limited empirical evidence about the impact of this intervention on population health. One of the main reasons for this is that much of the existing literature focuses on case studies or on middle and low-income countries. In addition, current studies on decentralization and health outcomes have generally relied on fiscal data from the Government Financial Statistics (GFS) of the International Monetary Fund (IMF). While providing a consistent dataset across countries and over time, decentralization measures based on data from the GFS fail to properly reflect the level of autonomy in policy making of sub-national tiers of government (Ebel and Yilmaz, 2002, Rodden, 2003).

The main objective of this paper is to use empirical analysis to inform the debate about the most appropriate degree of decentralization of the health system. We contribute to the research literature by using an improved fiscal decentralization indicator on a panel of 20 OECD countries to test the extended hypothesis in the research literature that more decentralization leads to better health outcomes. The next section discusses the theoretical considerations to understand how decentralization could influence population health. Section three presents an overview of the current evidence on the impact of decentralization on health outcomes. Section four describes the new data on fiscal decentralization, while section five presents the empirical specification used to model the relationship between decentralization and health status and the data and variables employed in this study. In section six the results are shown and section seven concludes.