How Economic Inequality Structures Vote Buying Across Africa

• In Event: Clientelism and Distributive Politics 1

Abstract

Countries with more clientelism and vote buying have less public investment (Keefer 2007; Khemani 2015). The basic problem is straightforward. Providing public services is less attractive to politicians seeking reelection because they are subject to the free-rider problem: both supporters and opponents can benefit regardless of vote choice. One classic solution to the free-rider problem is to provide selective incentives (Olsen 1965). As a result, mobilization through clientelism and vote buying is attractive because leveraging private or club goods to incentivize support allows for the exclusion of citizens who do not support politicians electorally (Kitschelt and Wilkenson 2007). However, candidates face a budget constraint when deciding to provide public goods or selective incentives in the form of vote buying. In this paper, we consider how politicians navigate this tradeoff, and its implications for empirical patterns of vote buying in Africa.
Because vote buying research has typically taken clientelistic politics as a given and focused on the dynamics between the politician, voter, and broker within a single case, key questions remain. In particular, the role of national factors - beyond poverty - that may explain the use of vote buying to garner political support. We argue that the tradeoff between public goods and targeted direct transfers is shaped by the distribution of income within each country. In particular, the level of economic inequality in a country determines the relative efficiency of providing public goods vs private transfers as a strategy for attracting votes: Demands from citizens are relatively uniform in more egalitarian societies compared to unequal societies. As a result, it is relatively more efficient to provide public goods in more egalitarian countries. The inability to exclude citizens means that small investments in public services can have large distributional impacts. Public goods also avoid much of the rent seeking endemic in vote buying. In unequal societies however, there is a more heterogeneity in demand, and in turn, fewer citizens will want to consume any particular public good or service provided by the government. As a result, public goods will be relatively less appealing to politicians in this context.
While politicians in many low-income settings choose vote buying as an electoral strategy, there does exist wide variation in its prevalence, even in the countries for which we would most expect it. For example, Rosas et al. (2013) find that in local jurisdictions that provide homogenous political support to Venezuelan incumbents, voters tend to receive public goods. However, those same incumbents mobilize heterogeneous jurisdictions with clientelistic strategies such as vote buying. This result suggests that politicians may prefer to provide public goods when the free-riders are their own constituents. Similarly, Kramon and Posner (2013) study the provision of political favoritism toward co-ethnics in six African countries and find evidence that public service provision is higher to households whose ethnicity is the same as the president’s. The findings in this study also suggest that there are circumstances where politicians would prefer to forego the use of targeted transfers, provide public goods, and not worry about whether some non-supporters benefit. In this paper, we add to the literature by building a theory of how income inequality explains variation in levels of vote buying.
We test our hypotheses derived from our theory using inequality data from the World Bank’s World Development Indicators dataset and Afrobarometer’s Wave 5 data, which includes data from 37,101 respondents in 29 sub-Saharan African countries on whether they received a gift or money for their vote in the last election. Our results demonstrate that inequality impacts vote buying, even after controlling for GDP per capita, individual-level household poverty, and other variables identified in the literature. Moreover, the statistical analysis tests and rejects alternative explanations, such as whether the relationship is a result of improved campaign targeting in unequal societies. We consider robustness of our findings to endogeneity concerns through instrumental variable designs following recent work on effects of inequality.
By focusing on inequality, this paper moves beyond a common finding linking vote buying to poverty (Jensen and Justesen 2014, Hicken 2011), and explains why politicians in relatively wealthy nations may continue to engage in vote buying, while those in some low-income countries engage in less vote buying than one might expect. Our paper thus helps explain the “stickiness” of vote buying, even as countries develop economically. Moreover, the paper redirects theory away from individual-level determinants of vote buying, for example that poorer people value transfers more than others, toward structural determinants vote buying.

Authors

• Joseph Amick, Credence Management Solutions

• Henry Pascoe, IE University