We aim to estimate the effect of search imperfections on the length of the average commute. We start from the assumption that the commute of the self-employed is the result of a search process for vacant workplaces, whereas employees search for vacant jobs. Because the arrival rate of workplaces is much higher than the arrival rate of jobs, the self-employed minimize the commute, whereas employees may have to accept jobs with a longer commute. In the empirical analysis, the extent of the ‘wasteful’ or ‘excess commuting’ is identified by estimating the difference in the commute of employees and self-employed individuals. Our estimates indicate that about 40 to 60% of the observed commute may be considered ‘excess’ due to search imperfections. We reject a range of alternative hypotheses as to why the self-employed have a shorter commute than employees (self-selection of not working from home, preference for residence and workplace locations, characteristics of workers which are difficult to observe).