Despite numerous policy interventions, women still earn substantially less than men in most countries around the world. The income difference, however, can only partially be explained by observables. This paper estimates the causal effect of a relatively inexpensive anti-discrimination policy introduced in Switzerland in 2006 consisting of a simple regression framework that firms can use to monitor their wage policies. Random checks are implemented by the government and sanctions may lead to exclusion from public procurement. Only firms with more than 50 employees are subject to the random controls and our data span periods before and after the introduction of the policy. Using a combination of regression discontinuity and difference-in-differences methodologies, I found that this policy has reduced significantly the unexplained gender wage gap.