Electoral Outcomes and Provision of Public Goods: A Case Study of India

ABSTRACT

This paper analyses the implications of political factors on provision of public goods by state governments in India. Using data on state governments’ expenditure and constituency results of states' Legislative Assembly elections during the period 1971–2005, it demonstrates that the greater the spread of ruling party's strength across legislative constituencies in a state, the higher the share of developmental expenditure in revenue budget. It also documents that the share of developmental expenditure in revenue budgets is positively associated with voter turnout and with a change in the political party in power. Interestingly, delivery of public goods in a state does not appear to have any significant relation to the form of the government – single party or coalition. It also shows that economic liberalization has reduced governments’ responsiveness to provision of public goods.

KEYWORDS:

• Revenue budget allocation
• developmental expenditure
• public goods
• state government
• political factors

JEL CLASSIFICATIONS:

• H72
• O23 D72
• H40

Acknowledgements

We would like to thank Debraj Ray, Tim Besley, Karl Schlag and Elena Esposito for helpful comments and discussions. Remaining errors, if any, are our responsibility.

Notes

¹These numbers can also change over time subject to delimitation of constituencies, which is the redrawing of boundaries of assembly constituencies with the aim to have equal number of people in each constituency. Although Article 82 of the Indian Constitution directs Parliament to enact a Delimitation Act after every census, during the period of study such delimitation of constituencies took place only twice, in 1973 and in 2002.

²Economic liberalization in India refers to a process that was initiated in 1991. Its objective was to reduce the role of the government and increase the role of private players in the economy.

³Winning margin = .

⁴ We note here that, in the existing literature, a constituency is often considered as the stronghold of a party, if that party has won all previous elections during a given period in that constituency. See, for example, Khemani and Keefer (2009). However, such categorization does not appear to be appropriate because of the following reasons. First, the results are likely to be different for different length of the period considered, and there is no obvious way to choose the length of the period. Second, it is quite plausible that a party wins a constituency in all the elections held during a particular period of time by a negligible margin or merely by chance. Third, delimitation of constituencies, which leads to significant changes in geographical boundaries of constituencies as well as voters’ composition in constituencies to a large extent, has taken place twice during the period of study. As a result, it becomes difficult to trace a particular constituency over a reasonable time period.

⁵We mention here that, since all the political variables used in the regression analysis were determined before the budget allocation decision took place in any given year, the problem of endogeneity is less likely to occur in the present analysis. Nonetheless, we recognize that the possibility of erogeneity problem cannot be ruled out altogether, unless we consider appropriate instrumental variables.

⁶These results are available from the authors on request.