Biased Contests

We examine the effects of providing more accurate information to a political decision-maker who is lobbied by competing interests. Conventional wisdom holds that such a bias in the direction of the correct decision improves the efficiency of government. We provide a formal definition of bias which is derived from the same fundamentals that give rise to a contest model of lobbying. Efficiency of government is measured by both the probability of taking the correct decision and the amount of social waste associated to lobbying activities. We present a benchmark model in which increasing the bias always improves the efficiency of government under both criteria. However, this result is fragile in the sense that reasonable alternative assumptions in the micro-foundations lead to slightly different models in which -due to different strategic effects of bias- under either criterion there is no guarantee that more accurate information improves government.