Persistence of Suburban Violent Crime Rates: An Ecological Analysis

Abstract

A lagged endogenous variable structural model of 1970 suburban violent crime rates is constructed and evaluated using a sample of 252 large suburbs for which 1960 and 1970 official crime data were available. It was found that suburban crime rates are persistent, that employment specialization and percent low income affect change in violent crime rates, and that suburban age, density, employment specialization, percent low income, percent Black, and region have substantial indirect effects on 1970 crime rates through earlier areal crime rates. Despite theoretical expectations, population growth and age composition exerted no significant direct or indirect effects on violent crime or changes in violent crime rates.