Corruption in NREGA: a theoretical analysis

Abstract

Purpose Niehaus and Sukhtankar (2013a, 2013b) find that in response to an increase in the public wage rate in National Rural Employment Guarantee Act (NREGA), underpayment of wage and embezzlement rates increase monotonically. This paper aims to investigate theoretically whether the empirical finding of Niehaus and Sukhtankar (2013b) is valid in the long run. The author also verify whether their observation extends to the situation where multiple types of corruption coexist. Design/methodology/approach The author builds a theoretical model comprising of three individuals – a government official, a landlord and a representative worker. First, the author defines a no-corruption equilibrium where the official honestly reports the employment period to the government. In a two-stage game, the landlord announces the private wage rate in stage one, followed by the worker deciding on the allocation of time between private and public works and leisure. Then the author considers a single type of corruption where the official embezzles a part of public money by over-reporting the employment period. The landlord-worker game remains the same as before. Finally, the author considers multiple types of corruption where alongside embezzlement, the official takes a bribe from the worker for employing her. Here also, the author considers a two-stage game where in stage one, the official and landlord simultaneously determine embezzlement and bribe rates and the private wage rate, respectively. In stage two, the worker decides on the allocation of time. In both types of corruption, the probability of detection increases with the corruption rate. Findings In the case of a single type of corruption, the embezzlement rate behaves nonmonotonically in response to an increase in the public wage rate. A similar kind of result is obtained with multiple types of corruption as well. Originality/value First, to the best of the author’s knowledge, no paper prior to this has theoretically modeled corruption in NREGA. Also, the observations of Niehaus and Sukhtankar (2013a, 2013b) is neither valid in the long run nor extend to multiple corruptions.