Das Dokument ist öffentlich zugänglich im Rahmen des deutschen Urheberrechts.

Despite the growing skepticism regarding the efficacy of antidepressants, global consumption has increased at an unprecedented rate with unknown implications for society. We estimate the causal effect of this increase on mental health outcomes using an instrumental variable strategy that exploits pharmaceutical company local market power and the availability of detailed drug sales data from Switzerland between 2002 and 2014. Our main instrument, a modified version of the popular shift-share instrument, relies on the national growth in antidepressant sales for pharmaceutical companies (the shift) - mainly due to product innovation - and assigns it locally using regional non-antidepressant market shares. Our estimates show that an increase in antidepressant sales causes a sharp increase of hospital admissions related to depression symptoms. An alternative instrument, which exploits prescribing practice spillovers from neighboring countries, leads to very similar point estimates providing further evidence about the validity of our results.