This paper is the first to estimate the impact of a direct measure of firm-level upstreamness on productivity, wage costs and profits (i.e. productivity-wage gaps). To do so, we merged detailed Belgian linked panel data, covering all years from 2002 to 2010, to a unique data set developed by Dhyne et al. (2015), which contains accurate information on the position of (almost) each commercial firm in the value chain at each year. We rely on the methodological framework that has been pioneered by Hellerstein et al. (1999) to estimate dynamic panel data models at the firm level. Our estimates show that if upstreamness increases by one step (that is, by approximately, one standard deviation), productivity rises on average by 5%. They also indicate that productivity gains associated to upstreamness are shared almost equally between wages and profits. However, upstreamness is found to be more beneficial for workers’ wages in less competitive environments, where the price-elasticity of demand for firms’ products is typically smaller. Overall, these findings are compatible with the assertion that firms should move up the value chain to be more productive and profitable, but also that being higher in the value chain is likely to facilitate firms’ control over strategic downstream activities. Our results can also be understood through the application of the Melitz (2003) model to the value chain framework.