Elsevier

Journal of Urban Economics

Volume 68, Issue 3, November 2010, Pages 247-259
Journal of Urban Economics

Crime and urban flight revisited: The effect of the 1990s drop in crime on cities

https://doi.org/10.1016/j.jue.2010.05.002Get rights and content

Abstract

The ‘flight from blight’ and related literatures on urban population changes and crime have primarily considered times of high or increasing crime rates. Perhaps the most cited recent work in this area, Cullen and Levitt (1999), does not extend through 1990s, a decade during which crime rates declined almost continuously, to levels that were lower than experienced in decades. This paper examines whether such declines contributed to city population growth and retention (abated flight). Through a series of population growth models that attempt to identify causality through several strategies (including instrumental variables) we find at best weak evidence that overall city growth is affected by changes in crime. We find no evidence that growth is differentially sensitive to reductions in crime, as compared to increases. Focusing more narrowly on within MSA migration, residential decisions that are more likely to be sensitive to local conditions, we do find evidence supporting abatement of ‘flight’ – that is, lower levels of crime in central cities in the 1990s are associated with lower levels of migration to the suburbs. This greater ability to retain residents already in the city does not appear to be accompanied by a greater ability to attract new households from the suburbs, or from outside of the metropolitan area.

Introduction

For most of the 20th century, concerns about safety and high crime rates have beset US cities. Researchers and policymakers pointed to these high urban crime rates as one of the chief ‘urban blights’ from which middle class, mobile (and typically white) households fled during the post-war period, fueling suburbanization. But this picture changed dramatically in the 1990s, a decade during which the crime rate in the US fell by a remarkable 30%, and crime rates in many US cities declined even further.

This paper builds on the ‘flight from blight’ literature, and considers what effect (if any) the 1990s drop in crime rates had on urban population changes. While the empirical evidence is somewhat mixed about the effect of crime levels on population changes, work by Cullen and Levitt (1999) suggests that changes in crime during the late 1970s and 1980s may have contributed to central city flight, particularly of more affluent households and those with children. Their work, however, does not extend through the 1990s. Unlike the two previous decades, the 1990s was a period of almost continuous declines in crime rates, to levels that were lower than experienced in decades. Did this drop in crime reverse, or at least abate, urban flight? It is not obvious a priori that the relationship between crime and residential decisions would be symmetric – while increases in city crime may push residents away from cities, similarly-sized reductions may not attract or retain them.

We examine this more recent time period using census data for 1980 through 2000, for more than 100 central cities and their surrounding metropolitan areas, in combination with more than 20 years of Uniform Crime Report data. We first consider the effect of crime on overall changes in city population, attempting to identify causality through models with lagged measures of crimes, and through an instrumental variables strategy based on state-level measures of stringency of state criminal justice systems. Relying on models with lagged measures of crime, we find at best weak evidence that overall city growth is affected by changes in crime, and considerable evidence that researchers using similar data need to take seriously a number of measurement and estimation issues. We find no evidence that growth is differentially sensitive to increases in crime, as compared to decreases.

We then examine the relationship between city crime and particular sources of population change (migration into cities from surrounding suburbs and migration out of cities into surrounding suburbs) that are more likely to be sensitive to local conditions. Here we find evidence supporting an abatement of ‘flight’ – that is, lower levels of crime in central cities in the 1990s are associated with lower levels of migration to the suburbs. This greater ability to retain residents already in the city does not appear to be accompanied by a greater ability to attract households from the suburbs, or from outside of the metropolitan area.

Section snippets

Background

High crime rates in cities have long been viewed as one of the chief causes of suburbanization and urban flight (for overviews, see Bradford and Kelejian, 1973, Mieszkowki and Mills, 1993). However, empirical support for the importance of crime rates in shaping residential location decisions is fairly mixed. Indeed, even a cursory review of the literature provides evidence of a positive relationship between central city crime and suburbanization or flight (Marshall, 1979, Grubb, 1982), an

Data and empirical strategies

To address our two primary research questions, we employ two slightly different data sets. Our first data set is a city-level panel extending from 1980 through 2000, for all central cities in 1980 with populations of at least 100,000. Annual data on city and county crime rates are taken from the Federal Bureau of Investigation’s Uniform Crime Reports (UCR). We use data on violent crimes, property crimes, and all index crimes. The annual city population data used to determine the UCR per capita

Do crime rates affect overall population changes for cities?

To start, we examine annual changes in central city population, over a 20-year time period beginning in 1980. As Fig. 1 shows, average central city crime rates varied quite a bit during this period, experiencing both increases and decreases.6 After 1992, however, mean crime rates declined almost continuously. To our knowledge, no paper

Do reductions in crime rates reverse central city flight?

The annual models of city population changes consider net changes in populations, and thus include some changes potentially less likely to be related to crime. Moreover, this aggregate measure combines the two key residential decisions that we think are likely related to crime – retention and attraction – into a single net effect. Yet it is quite possible that decisions to remain in the central city are more/less sensitive to changes in crime than decisions to move into a central city from

Conclusion

We began this paper with an interest in assessing whether the decline in crime rates in the 1990s had a positive impact on city growth. Our preliminary results suggested that in aggregate, they did not (and indeed may have depressed growth). However, those results suffered from multiple estimation issues, most notably measurement error bias and the endogeneity of crime rates. When we focused on the number of homicides (controlling for population separately) or estimated decade models of

Acknowledgments

We gratefully acknowledge funding support from the Wagner School and the Furman Center at NYU. We are especially appreciative of the research assistance provided by Michael Lens, Keren Horn, Diana Beck, and Emre Edev. We thank Ted Joyce, Sanders Korenman, and Karl Kronebusch for helpful comments. And a final thanks to Steven Raphael and Rucker Johnson for providing state data on prisons and crime in electronic form.

References (23)

  • W. Frey

    Central city white flight: racial and nonracial factors

    American Sociological Review

    (1979)
  • Cited by (43)

    • The spatial reordering of poverty and crime: A study of Glasgow and Birmingham (United Kingdom), 2001/2 to 2015/16

      2022, Cities
      Citation Excerpt :

      Yet, given the theoretical framing of this investigation it is also important that the spatial unit of analysis comprises meaningful ontological settings (i.e., neighbourhoods). Studies have found the ‘flight from blight’ (Foote, 2015), the suburbanisation of non-poor city dwellers, to be driven by the concentration of poverty and crime in inner urban areas (Ellen & O’Regan, 2010). What then might the counter urbanisation of non-poor, and the suburbanisation of poor, city dwellers imply for the spatial patterning of crime?

    • I read the news today, oh boy: The effect of crime news coverage on crime perception

      2020, World Development
      Citation Excerpt :

      We are going to focus on (ii) to (iv), although a deep engagement with these issues is left as an avenue for future research. Individuals may react to a higher crime perception by changing their habits inefficiently affecting both time and money allocations, which may further restrict mobility, labor decisions and even housing choices (Ellen and O’Regan, 2010; Hamermesh, 1999). To explore how individuals change their habits in response to crime perceptions we exploit some survey questions from the enapres that were only available in 2017.

    • Has falling crime invited gentrification?

      2019, Journal of Housing Economics
    • Crime and private investment in urban neighborhoods

      2018, Journal of Urban Economics
      Citation Excerpt :

      The converse does not appear to be true, however. Declining crime rates have not been found to induce in-migration by those living in the suburbs, though they do appear to limit the extent of out-migration from urban neighborhoods (Ellen and O’ Regan, 2010). Regarding business location decisions, Rosenthal and Ross (2010) is one of a small number of studies that has considered how crime affects such decisions.

    • The effect of violent crime on economic mobility

      2017, Journal of Urban Economics
    View all citing articles on Scopus
    View full text