We introduce a novel way to infer employer search behaviour, through deadweight loss incidence in wage subsidy schemes. Using a data set on british firms participating in such schemes we can distinguish between intensive and extensive employer search. These data also allow us to separate the negative economies of scale effect from the positive monitoring effect of firm size on intensive search costs. Our findings further indicate that ‘hours worked’ and ‘supervisory tasks’ increase extensive search costs. Since intensive and extensive search behaviour affect the job find probability of the (long-term) unemployed, our conclusions have significant labour market policy relevance.