Housing markets are subject to search frictions. The matching function is a widely used tool in macroeconomics to summarize labour-market search frictions and has been estimated in an extensive literature. Limited data on home-buyers has precluded estimation of housing-market matching functions. This paper fills that gap using a novel dataset identifying buyers and their search behaviour in viewing properties. An event study based on the staggered removal of restrictions on housing purchases provides the first causal estimation of the elasticity of matches with respect to buyers for a constant-returns-to-scale Cobb-Douglas matching function, which is found to be around 0.5.