Association for Public Policy Analysis & Management

Are rentier states more likely to use public sector employment as a redistribution mechanism? A standard probabilistic voting model by Robinson & Verdier (2013) predicts they are: in rentier states politicians have incentives to prefer public employment to other means of redistribution, because it yields a device for credible mutual commitments with voters. Our paper provides the first empirical test of this model, exploring the predicted link between natural resources rents (as a proxy for rentierism) and public sector employment shares as a proxy for clientelistic redistribution. Our cross-national analysis, covering 92 countries for the 1990-2015 period, confirms that public sector employment increases with low productivity and poor institutions. However, contrary to Robinson and Verdier’s predictions, we find that public sector employment size does not correlate to natural resources revenues.

The use of public employment as a redistribution mechanism is widespread, across countries with different political institutions, income levels and cultures. However, are resource rich governments more likely to prefer public employment as a redistribution mechanism? Could this be a popular clientelistic vote-buying device despite the existence of alternative, better targeted programmatic redistribution approaches? The empirical literature offers no large-scale attempts to explicitly link natural resources, public employment size and clientelism. This study aims to fill this gap. Based on Robinson & Verdier's model, we empirically tests under which conditions public sector employment is used as a preferred redistribution mechanism.

Contrary to Robinson and Verdier’s hypothesis, we find no clear relationship between the size of public workforces and rents extraction capacities; nonetheless, public sector employment does increase in low productivity and poor institutions setups. This finding appears robust to different sample sizes, definitions and measures of government employment and rent-seeking, the addition of control variables and alternative econometric specifications. We draw on cross-sectional data from multiple sources at the country level, and the main empirical specification is an OLS with regional dummies, averaged by 5 years sub-periods.

Studying under which circumstances public employment is used for income redistribution, clientelism and rent-seeking, is policy-relevant. It can inform country policy-makers’ and donors’ risk-mitigation strategies as they design redistributive programs or public service reforms. If, resource rich countries have particularly large public workforces, this can undermine social welfare, among others by crowding out investment expenditures. Further, redistribution mechanisms other than public employment could perhaps better target low productivity individuals. Finally, rent seeking bureaucracies can also distort markets (Jaimovich & Rud, 2014), by hurting private investment and misallocating human resources, leading to segmentations in the labor markets.

Our research contributes to four separate strands of the literature: to the inefficient redistribution literature by testing a central comparative static from Robinson and Verdier’s model; to a closely related literature on clientelism and public employment (Calvo & Murillo, 2004; Rodrik, 2000); to research on the political economy of the natural resource curse (Beblawi, 1987; Dunning, 2008), by testing one key mechanism; and to a small emerging empirical literature on the impact of natural resources on the size of public employment (See Ali & Elbadawi, 2016).