This paper studies the impact of university-industry collaboration on academic research output. We analyze the channels through which the degree of industry collaboration may be affecting research output. We exploit a unique longitudinal dataset on all the researchers in all the engineering departments of 40 major universities in the UK for the last 20 years. We use an innovative measure of collaboration based on the fraction of public research grants that include industry partners. Our empirical findings corroborate that the relationship between collaboration degree and publication rates is curvilinear, and shed some light on the selection mechanisms at work. Our results are robust to several econometric methods, measures of research output, and subsamples of academics.