We present the first capitalization study to look at crime fighting and house prices using the causal inference technique of regression discontinuity. It is also the first study on the link between police spending and housing sales volume. Voting to increase police taxes and spending by 15% is not linked to house prices or transaction volume overall. However, increasing spending causes 13% higher housing prices in low-income cities and at least a 14% decrease in house prices in high-income cities. The effects persist through all five years after the vote that we study. The sale price results are most consistent with overfunded police departments, rather than Tiebout sorting, neighborhood instability, or signaling. Our results suggest that the small or non-existent link between house prices and crime found by the literature really just reflects the sum of large but opposite moves in house prices in different market segments.