A randomized evaluation of two teacher incentive programs was conducted in a nationally representative sample of 420 public primary schools in Guinea. In 140 schools, high-performing teachers were rewarded in-kind, with the value of goods increasing with level of performance. In another 140 schools, high-performing teachers received a certificate and public recognition from the government. After one year, the in-kind program improved learning by 0.24 standard deviations, while the recognition treatment had a smaller and statistically insignificant impact. After two years, the effect from the in-kind program was smaller (0.16 standard deviations), and not significant: we provide evidence that the reduction was likely due to the onset of an Ebola outbreak. The effects of the recognition program remained small and insignificant. The effects differed by teacher gender: for female teachers, both programs were equally effective, while for male teachers, only the in-kind program led to statistically significant effects.