Convergence and Divergence Between UCR and NCVS Trends

• In Event: Using the UCR and NCVS to Enhance Understanding of Crime Trends and Offender Behavior

Abstract

Regardless of whether they are measured in the National Crime Victimization Survey (NCVS) or the Uniform Crime Reports (UCR), crime trends in the United States are non-stationary: their mean and variance change over time. The linear combination of non-stationary time series, however, may be stationary. If so, they are said to be “cointegrated.” Cointegrated time series have several desirable statistical and substantive properties. Prior research has shown that NCVS and UCR crime rates are not cointegrated, but that research is limited in statistical power and substantive scope. The current study applies several cointegration tests, with and without structural breaks, to aggregate and disaggregated NCVS and UCR crime trends. The implications of the results for future research and policy guidance are discussed.

Authors

• Richard Rosenfeld, University of Missouri - St. Louis

• Janet Lauritsen, University of Missouri - St. Louis