Association for Public Policy Analysis & Management

Business Improvement Districts (BIDs) are an increasingly popular example of a public-private initiative employed to revitalize urban commercial corridors and maintain a reasonable level of neighborhood services.  BIDs are enabled by the public sector, but established, operated, and funded by private property owners, and one of their primary responsibilities is maintaining the safety of the streets.  While cities across the world have fostered, and even promoted, the formation of BIDs as a way to keep the streets safe and vibrant (without additional expenditures on their parts), there is little empirical evidence to support their effectiveness.  There is reason to believe that BIDs will decrease crime due to their supplemental safety measures and increased street-level vigilance.  However, we have little evidence to show that BIDs actually reduce crime.  Further, it is possible that any abated crimes or nuisances simply take place somewhere outside of the BID instead.  Is there a shift in the type of crime committed in the BID, i.e. away from violence towards property or lower level nuisance?  Is any crime reduction countered by the increased street activity that BIDs often bring, which could also introduce more opportunities for crime?  What exactly are the mechanisms through which BIDs influence crime—via professional crime control services (like cameras or patrols) or as a platform to communicate with local precinct and city officials?

In this paper, we rely on econometric analyses to answer these questions and to test for the magnitude and nature of any effect of BIDs on crime in New York City, which has more BIDs than any other city in the country (71 total and 20 with formation dates that we observe in our sample).  We have access to rich data, spanning a decade, on BID characteristics (specifically their formation dates, boundaries, budgets and services), point-specific crime reports, nuisance complaints, and neighborhood demographics that we can exploit to more precisely identify the relationship between the BIDs’ activities and investments and criminal activity.  Furthermore, we employ two counterfactuals in testing for the BIDs’ effect: first, we exploit the fact that we can compare point-level crime data inside the BID boundary to that immediately outside the BID boundary (but in the same broader neighborhood context) and second, we identify a sample of commercial clusters that do not yet have BIDs and compare their crime-related outcomes to those of the actual BIDs.  Our sample includes a wide range of neighborhoods and BIDs (with respect to their size, spatial layout and service portfolio), which means we can test for any variation in effects depending on the context and structure of the BID.