This paper investigates the empirical validity of the setup of a large-scale government neighbourhood investment programme in the Netherlands. Selection of neighbourhoods into the programme was determined by their score on a predetermined index. At first sight this is a textbook example for the application of a regression discontinuity (RD) design to estimate the causal effect of the programme on neighbourhood outcomes. However, at the discontinuity threshold we find a large gap in the share of non-Western immigrants. In addition, the pattern of non-compliance with the assignment rule is consistent with investing in neighbourhoods with a high share of non-Western immigrants. Finally, the way of selecting neighbourhoods into the programme could be a likely explanation for the imbalance at the discontinuity threshold. This case illustrates that RD designs can become invalid even when treatment and control groups have no influence on the assignment.