Unemployment rates vary over both time and space. In this article we seek to understand why the unemployment rate varies from area to area, using districtlevel data from the Midi-Pyr´en´ees region of France. The unemployment rate in any given area is determined more by national than by local effects (OCDE 1997), and most of the theoretical and empirical work on the subject reflects this (e.g., Artus and Muet 1995). Thus the unemployment rate in the Midi Pyr´en´ees region has evolved in line with the national unemployment rate, rising gradually from 3.2% in 1974 to 11.3% by January 1995. The rise in unemployment occurred despite robust growth in the Toulouse area, which was driven in large part by the rise of the aerospace sector.
Although regional unemployment rates move broadly in line with the national rate, differences across regions tend to persist over long periods of time, and 30% of regional movements in the unemployment rate cannot be explained at the national level (OECD cited by Tervo 1998). The correlation coefficient between the unemployment rates in France’s 22 regions in 1974 and the unemployment rates in 1994 was 0.62 (Maillard 1997, p. 156), which indicates a very substantial degree of persistence in inter-regional differentials in unemployment rates.
This persistence effect has been observed elsewhere. For instance during the 70 years prior to about 1990, unemployment rates were systematically higher in the north of England than in the south and southeast (Evans and McCormick 1994, p. 634). Devens (1988), Chapman (1991) and Martin (1997) have also documented the persistence of differences in unemployment effects in the UK. Similar, if somewhat weaker effects have been observed in the United States (Martin 1997, p. 241; Partridge and Rickman 1997). Disparities in unemployment rates across regions within Italy, Germany and the UK, at least during 1984–1994, were larger than the differences in unemployment rates among the countries (Taylor and Bradley 1997).
Persistence is also found at the sub-regional level, although this has not been studied as much. Pehkonen andTervo (1998) found that unemployment differentials tended to persist at the level of 423 municipalities in Finland, although less markedly than at the level of the country’s ten regions. The coefficient of correlation between the unemployment rates in the 174 hinterlands of the Midi-Pyr´en´ees region of France in 1982 and 1990 was 0.60, indicating a high degree of persistence in the spatial pattern of sub-regional unemployment rates over time.
One is naturally prompted to inquire into the explanation for these differences in unemployment rate across districts. That is the subject of this article. In a recent review of the French labor market, Gambier and Verni`eres (1991) note that it is ‘surprising that the study of economic geography has not been more fully applied to labor markets’ (p. 136).We hope to begin to fill this gap.
The next section reviews the relevant theory and literature and is followed by a description of the data, a discussion of the model to be estimated and the estimation issues, and the main results.
Download
PDF Ebook Explaining the pattern of regional unemployment: The case of the Midi-Pyr´en´ees region
