The evolution of regional manufacturing employment: gross job flows within and between firms and industries

New England Economic Review, Summer, 2002 by Scott Schuh, Robert K. Triest

The distribution of manufacturing employment across regions of the United States has changed tremendously over time. Shares of manufacturing employment in older, northern regions of the country have declined markedly relative to shares in the Sunbelt regions. (1) But the shifting of manufacturing employment shares goes beyond the well-known migration of population to the South and West. Manufacturing employment relative to population has also fallen in northern regions, and even the absolute number of manufacturing jobs has declined in these areas as well.

Anecdotal evidence suggests that some of the shift in the distribution of manufacturing employment is due to the movement of particular firms and industries to the Sunbelt in search of lower costs of production and increased proximity to customers. However, other forces driving the shift between regions are also often cited. The fast-growing Sunbelt regions may have benefited from specialization in newer, faster-growing manufacturing industries than those clustered in the North. And the Sunbelt may also have been the preferred location for entrepreneurial manufacturing startups.

This study focuses on two particular questions. First, what is the importance of job shifts within a firm but across regions in explaining regional differences in manufacturing employment growth? Second, to what degree are the varying fortunes of regions due to employment reallocation within industries?

Our investigation measures and examines gross job flows (job creation and job destruction). Most studies of the evolution of regional manufacturing employment, like most studies of population migration, focus on the net changes in employment between regions. Much less is known, however, about the distribution of gross job flows across geographic regions and how that distribution changes over time. The primary evidence on regional gross job flows comes from Eberts and Montgomery (1994, 1995). Using regionally aggregated data, these studies argue that cross-region variation in employment is more closely associated with job creation than job destruction. (2)

In this article, we reexamine the process of job creation and destruction at the regional level to gain a better understanding of regional employment dynamics. In particular, we use plant-level data from the Longitudinal Research Database (LRD) at the Census Bureau to construct quinquennial gross job flows for the period 1963 to 1992. This sample is notably longer than those in previous studies and yields a dearer view of the secular changes that dominate regional flows. Furthermore, the richness of the plant-level LRD provides the opportunity to disaggregate the data and construct measures that are not available from more aggregated data. Of particular relevance to this study, the LRD allows us to calculate gross job flows that occur across regional boundaries but within the same firm or industry. (3)

Our central finding is that although intrafirm and intra-industry job reallocation between regions makes up relatively small shares of total gross employment flows, these flows account for a substantial portion of the differences in manufacturing employment growth rates across regions. Reflecting the constant churn of plant births, expansions, contractions, and closures, regional gross job flow rates are much larger than are regional rates of net employment growth. So, job reallocation between regions can be both small relative to total gross flows and large relative to net employment growth. Our results are consistent with anecdotal evidence of firms shifting jobs from older, northern facilities to newer plants in the Sunbelt states. And our results refute the hypothesis that manufacturing employment in the Sunbelt areas has grown faster than in the North because of the Sunbelt's specialization in fast-growing industries.

The paper is organized as follows. First, we provide a historical context for our analysis of regional manufacturing reallocation, with a brief review of trends in regional manufacturing growth over the twentieth century. Section II follows with an exposition of the job-flow concepts that we use in our subsequent analysis, along with a discussion of data and measurement issues. Section III presents our evidence on regional plant-level and firm-level job flows, followed, in Section IV by our analysis of intrafirm and intra-industry job flows between regions. Section V concludes.

I. Regional Manufacturing Trends in Historical Perspective

Although manufacturing was concentrated in the northeastern part of the United States at the start of the twentieth century, as the century progressed, the distribution of manufacturing activity tended to move south and west. Trends in the distribution of manufacturing across regions of the United States are shown in Figures 1 through 4.(4) Figure 1 shows that throughout the twentieth century there was a steady drop in the share of manufacturing jobs located in New England and in the Middle Atlantic states, while the share of jobs in the southern and western regions of the country grew steadily. In 1899, New England accounted for 19 percent of manufacturing employment, and the Middle Atlantic, 36 percent; by 1992, these shares had declined to 6 percent and 12 percent, respectively.