A computational exercise on the bureaucratic corruption within the principal-agent approach with lotteries

This paper models a corruption problem through the principal-agent framework with lotteries where the bureaucrat is regarded as the agent and the society, as the principal. The model contemplates three variables: bureaucratic efficiency, compensation and level of honesty. Solutions for the model were given using linear program algorithms. The optimal contracts indicate that: (1) under an unmonitored scenario, it is possible that the society be better off in the presence of extreme corruption; (2) risk-loving bureaucrats can provide more righteous conducts and increase the society’s expected utility; (3) altruist bureaucrats makes the society better off while spiteful bureaucrats makes the society worse off; and (4) higher reservation utility tends to increase the probability of the bureaucrats receive a high compensation.