Among policymakers, educators and economists there remains a strong, sometimes heated, debate on the extent to which good schools matter. This is seen, for instance, in the strong trend towards establishing accountability systems in education in many countries across the world. In this paper, in line with some recent studies, we value school quality using house prices. We, however, adopt a rather different approach to other work, using a policy experiment regarding pupils' choice to attend high schools to identify the relationship between house prices and school performance. We exploit a change in school choice policy that took place in Oslo county in 1997, where the school authorities opened up the possibility for every pupil to apply to any of the high schools in the county without having to live in the school's catchment area (the rule that applied before 1997). Our estimates show evidence that parents substantially value better performing schools since the sensitivity of housing valuations to school performance falls significantly by over 50% following the school choice reform.