In recent decades, the US bank market has been exposed to several waves of mergers, resulting in concerns about branch presence and consumer access to financial services. This paper examines the effects of bank mergers on branch density in the period 2000–2020. To do so, we use panel regressions and matching techniques at the census tract level to study the impact of inter- and intrastate mergers before and after the Great Recession of 2007. To generate plausible exogenous variation for mergers, our analysis focuses on transactions involving large entities, and we consider the within-tract variation in exposure to mergers. A comparison of exposed and unexposed tracts shows that in the period under study each merger reduced branch density by around 3%. Moreover, interstate mergers reduced branch density at the tract level across the whole period but had an expansionary effect on the number of branches at the county level before the crisis. Intrastate mergers, in contrast, had a consolidation effect across the whole period, an impact that was more intense in rural tracts and in tracts where merging entities operated overlapping branch networks. Finally, we show that the reduction of bank branches was stronger in tracts with a relatively higher penetration of broadband Internet services, but we find no evidence that the adoption of FinTech services intensified branch closures.