Racial composition, unemployment, and crime: Dealing with inconsistencies in panel designs

Racial composition and unemployment have appeared as either theoretically-relevant controls or variables of substantive interest in numerous studies of crime. While there is no clear consensus in the literature as to their statistical significance, the lack of consensus has been most apparent in panel analyses with unit fixed effects. One explanation for this is that racial composition and unemployment are fairly invariant, or slow-moving, which leads to collinearity with unit dummies. A number of pertinent studies are reviewed to illustrate how two slow-moving variables, percent black and percent unemployed, have behaved inconsistently. A fixed effects vector decomposition procedure [Plumper, V., Troeger, V. E., 2007. Efficient estimation of time-invariant and rarely changing variables in finite sample panel analyses with unit fixed effects. Political Analysis, 15, 124–139.] is used to illustrate how these variables’ coefficients appear positive and significant when the slow-moving process is accounted for.