We carry out a randomized controlled experiment in West Bengal, India to test three separate performance pay treatments in the public health sector. Performance is judged on improvements in child malnutrition. We exogenously change wages of government employed child care workers through either absolute or relative incentives. We also test for the impact of high and low absolute incentives. Results show that high absolute incentives reduce severe malnutrition by 6.3 percentage points over three months. Result is consistent with a reported increase in protein-rich diet at home in the high absolute treatment. There are no significant effects on health outcomes of other incentive arms. Results remain robust to propensity score matching, reversion- to-mean and a placebo check.