We examine individual-level compensating differentials for commuting distance in a quasi-natural experiment setting by examining how wages respond to changes in commuting distance induced by firm relocations. This set-up enables us to test for the relevance of job search frictions within labour market models. Due to the quasi-experimental set-up, we are able to avoid a range of endogeneity issues. We demonstrate that a 1 km increase in commuting distance induces an almost negligible wage increase in the year after the relocation but a more substantial wage increase of about 0.15% three years later. © 2013 Royal Economic Society.