We study neighbourhood externalities caused by large public investments in poor neighbourhoods. A stylised theory of a linear city is proposed to guide interpretation of the magnitude and attenuation of the external effects generated by these public investments. We use a large Dutch nationwide dataset and compare house price changes over time between targeted and control neighbourhoods. The control neighbourhoods are selected through a nonparametric propensity score method. The identifying assumption is that time-varying unobservable locational variables are uncorrelated to the choice of location for the revitalisation program. We relax this assumption by identifying the external effect based on the difference in price changes between targeted and control neighbourhoods within a municipality. To test for robustness of the results, we also identify the external effect within the neighbourhood. It is shown that public investments have lead to an increase in house prices of about 3 percent. Using a novel method to estimate a decay parameter within the empirical model, it is shown that the external effect decreases with 50 percent every 250 metres from the targeted area.