The Geography of Job Creation and Destruction in the U.S. Manufacturing Sector, 1967–1997

Abstract

Much geographical research focuses on the causes of geographical differences in net employment change. While net employment change is an important indicator of regional growth and decline, it masks the underlying process of creative destruction resulting in tremendous employment turnover rates. Labor economists and those who work in industrial organization emphasize the importance of theorizing the individual processes of job creation and destruction to uncover technological and institutional differences among industries and over the business cycle. This work demonstrates that over the course of a year, 10 percent of all jobs are destroyed and close to 10 percent of jobs are newly created. This article bridges the work of geographers on net employment change and the nongeographical work in industrial organization and provides the first comprehensive account of regional and metropolitan differences of job creation and destruction in U.S. manufacturing. Plant level data of the U.S. Census Bureau show that geographical differences in net employment change are primarily the result of differences in job creation rates and to a lesser extent in job destruction rates, that different types of plants drive the job creation process in different metropolitan areas, and that the snowbelt/sunbelt dichotomy collapses in the 1990s with selected midwestern urban areas among the fastest growing areas in the country.

Key Words:

• job creation
• job destruction
• employment change
• metropolitan change
• plant-level data

Acknowledgments

I would like to thank the anonymous reviewers, David Rigby, and the “People, Place, and Region” editor Audrey Kobayashi for valuable comments on an earlier draft of the article. If there are any remaining errors, they are the sole responsibility of the author. The empirical part of this research was conducted during research for the National Science Foundation project, grant number SBR97-19037, at the Center for Economic Studies and the California Research Data Center. The article is based on plant-level information from the Longitudinal Research Database of the U.S. Census Bureau at the Center for Economic Studies. The views expressed in this article do not necessarily reflect those of the U.S. Census Bureau or the Center for Economic Studies.

Notes

INCGR=expanding incumbent plants; ENT=new firm entrants; OPEN=new branch plant openings; INCDE=shrinking incumbent plants; EXIT=firm exit; CLOSE=closure of branch plant.

Percentages for plants are calculated as shares on the total number of plants in the base and end year; The net change for the number of plants is calculated as (firm entry+plant opening) – (firm exit+plant closures); Percentages for employment are calculated as ratio to base year employment.

Source: Longitudinal Research Database, Bureau of the Census.

Source: Longitudinal Research Database, Bureau of the Census.

Inc=incgr+incde; Firms=entry+exit; Bp=open+close.

Source: Longitudinal Research Database. Bureau of the Census.

Source: Longitudinal Research Database, Bureau of the Census.

For the first period, metropolitan areas were ranked according to their average manufacturing employment in 1967 and 1982. For the second period they were ranked based on their average employment in 1982 and 1997. Quintile 1=largest metro areas; Quintile 5=smallest metro areas.

Source: Longitudinal Research Database, Bureau of the Census.

Notes: The U.S. Census Bureau provides data for the U.S.: 4 regions, 9 census regions, 51 states, approximately 275 metropolitan areas, and over 2,000 counties with some metropolitan areas crossing state and census region boundaries. In this article, the Northeast and Midwest denote the snowbelt, while the South and the West comprise the sunbelt. Alaska and Hawaii are excluded from this analysis.

1. SIC: Standard Industrial Classification.

2. An incumbent plant is a plant that is in operation in year t and t+1.

3. Notice that the job reallocation rate for the fifteen-year period would be considerably higher if measured on a quarterly or annual basis because most plants survive for less than five years. Jobs created after 1967 and destroyed before 1982 are not included in the calculation of the job reallocation rate in Table 1. Calculating job reallocation annually would yield an average annual job reallocation rate of 3 million. Multiplying this number by fifteen years would yield a total of 45 million reallocated jobs over a fifteen-year period.

4. Over the period 1967–1997, boundaries for metropolitan areas were redrawn several times. In order to allow compatibility over time, the 1993 metropolitan area classification has been applied consistently across all years. In the context of this article, “urban” refers to metropolitan counties as classified in 1993, while “rural” refers to the nonmetropolitan counties as defined in 1993.