Spatial hysteresis and the clustering of urban unemployment

Abstract

Researchers into poverty have become increasingly concerned by the growing number of individuals whose social and economic isolation from mainstream society has been exacerbated by the geographic clustering within urban areas brought on by the recessions of the 1980s and 1990s. Evidence from the United States, Britain, Europe, and Australasia indicates that in almost every case such exclusion is closely associated with growing unemployment. The concern of this thesis is that such geographic concentrations may in fact inhibit the ability of the men and women concerned to re-enter the workforce once the economy begins growing again. It is this idea which is present in the concept of spatial hysteresis: namely, that inertias built up in recessions become cumulative if the unemployed are spatially clustered, leading to even longer periods of unemployment, associated poverty, welfare dependence, and the movement of the employed out of such areas due to the costs imposed on them by the clustering of the unemployed. According to this argument, geographic clusters of the unemployed that increase in intensity and expand during recessions fail to shrink or lessen and may even expand during periods of subsequent growth. This thesis tests this argument by tracing the development of geographic clusters of the unemployed in Auckland and Wellington over the period 1986 to 1996. The spatial effects of the recession that raised aggregate unemployment during the 1980s is contrasted with the changing geography of the unemployed during the period of economic growth in the early 1990s. Employing a 'local indicator of spatial association', the analysis of unemployment rates at the meshblock level of the New Zealand census provides strong evidence for the spatial hysteresis effect. The major clusters of meshblocks with high unemployment rates relative to the city average not only failed to dissipate as the metropolitan economies grew again in the 1990s, but actually expanded. While individuals outside these areas appeared to be re-entering the workforce, labour adjustment within the clusters appeared to be retarded. This result underscores the international concern over the geographic characteristics of social exclusion. While the main concern of the thesis was to assemble evidence for the presence of spatial hysteresis, some attention is also paid to the processes underlying the observed clustering. An exploratory analysis distinguishes between site specific factors such as the heavy dependence of certain areas on manufacturing employment and on rent subsidies through state housing, and the process by which vulnerable areas loose those (re)entering employment and attract additional unemployed thus aggravating the contagion effects. Support for the last process was evident but not conclusive, which in turn raises important future research questions about the role played by contagion effects and exactly how they work on vulnerable populations in New Zealand.