Nation or State? Where Should We Look to Measure Democracy's Effects on Health?

Abstract

The large body of research on regime effects on health primarily consists of cross-national studies and finds democracies to deliver better health than dictatorships. However, this research largely neglects the fact that state governments often have primary responsibility for social service provision. States matter and differ in their social policies. This is true for federal systems, which include populous nations such as China and India, as also countries which have experienced a ‘devolution revolution’. Based on a brief analysis of the health performance of India and its states and the findings of a small number of studies which have carried out subnational comparisons, the article calls for, following Richard Snyder, upgrading current research to include (1) within-nation comparisons and between-nation comparisons of subnational units from different countries or (2) using subnational comparisons in conjunction with cross-national comparisons. Until then, the validity of the democracy advantage thesis will remain in doubt.

Keywords:

• regime effects
• ‘democracy advantage’ thesis
• health
• state-level comparisons
• India

Acknowledgements

This paper emerged from an ongoing book project Democracy, Civil Society and Health in India (with Madhvi Gupta). Thanks are due to the two anonymous reviewers of the journal and to Madhvi Gupta for comments and suggestions to earlier versions of the article.

Notes

¹The term ‘devolution revolution’ was coined by Richard P. Nathan of the Rockefeller Institute of Government in Albany in his testimony before the Senate Finance Committee on 27 July 1995, to refer to shift in responsibility for social welfare programs from the federal government to the states (Moynihan, 1995).

²One of the main criticisms of cross-national comparisons is their assumption of ‘strong similarity’ between different countries when that is not the case (Jung and Marshall, 1985).

³Note, however, that these studies do not examine the role of political competition within a framework of the impact of regime type on health or human development more generally.

⁴Beer and Mitchell (2006) examined India's human rights record which is worse than predicted for a long-standing democracy. They found substantial inter-state variation with high human rights violations in India's border states than elsewhere which accounts for the country's overall poor human rights performance.

⁵However, human development outcomes in Bihar have so far not registered any significant improvements perhaps due to a time lag.

⁶Poverty matters for health. On poverty and infant mortality in southern US states, see Menifield and Dawson (2008). Furthermore, even comprehensive health services in fairly developed welfare regimes address primary health care concerns better than those health issues – diabetes, heart diseases and others – which are characteristic of countries like Chile or China which have experienced an epidemiological transition. Among other things, addressing primary health issues is relatively inexpensive whereas addressing the health concerns of an aging population is expensive.

⁷Inter-state variations may exist partly due to discrimination by national governments so that they favour some states over others. If that is so, it calls into attention the nature of centre-state relations and requires shifting focus away from the national government to its relationships with specific states. However, if there is a clear separation of power and responsibilities between the centre and the states, it may be more useful to focus on the states themselves.

⁸For example, Jeffrey Sachs has for long emphasized the connection between geography and disease (see for example, Gallup and Sachs, 1998).

⁹Past and current debates over the merits of cross-national and subnational comparisons are unlikely to be resolved in favour of one or the other. In the final analysis, both kinds of comparisons have a set of advantages and limitations. Cross-national comparisons remain a preferred method in quantitative research. Other than more ‘routine’ limitations of the subnational method (Snyder, 2001), newer ones have been identified. For example, Marks et al. (2008) have pointed out that most quantitative subnational comparisons use ‘rudimentary’ measures for the structure of the government. While a distinction is made between federal and non-federal countries, little is done to address, for example, issues such as the spheres over which the national and subnational government exercises authority. Finally, many subnational comparisons tend to be qualitative and therefore open to a different set of criticisms.