Deindustrialization of rural America: Economic restructuring and the rural ghetto

ABSTRACT

Since 2000, there has been a significant decline in the number of manufacturing jobs in the US. The manufacturing sector has provided good jobs and benefits, and it has a stronger multiplier effect on regional economies than other industries. Post-industrial theory suggests that there is a natural transition from agriculture to manufacturing, and ultimately, to services. Rural areas, especially those that are small and not adjacent to a metropolitan area, lack the population density and size to support many service sector jobs. Deindustrialization has led to the rise of “rural ghettos.” These places face many of the same problems as urban ghettos, such as high levels of underemployment, selective out-migration, social isolation, drug use, and lack of services.

KEYWORDS:

• Deindustrialization
• rural development
• ghetto

Introduction

For several decades now, industrial recruitment has been a key rural development strategy, especially attracting establishments in low-skilled, low-wage industries (Summers & Selvik, 1979). Manufacturing firms have been attracted to rural areas by the availability and cost of land and labor (compared to urban areas), and to a lesser extent, the tax breaks and other incentives provided by state and local governments. Local officials have viewed rural industrialization as the primary means of creating jobs, increasing income, and providing new opportunities for the poor (especially those who had less education and few skills).

Rural communities have historically been dependent on extractive industries (e.g., forestry, mining, fishing, and agriculture). During the 1960s and 1970s, manufacturing employment grew rapidly in many rural areas, especially in the Midwest and Southeast US (Summers, Evans, Clemente, Beck, & Minkoff, 1976). Growth of manufacturing jobs offered pathways to the middle class for the rural workforce. Manufacturing jobs have traditionally provided higher wages and better benefits than other types of jobs available in rural areas (Green & Sanchez, 2007). The manufacturing sector usually has strong linkages with local suppliers and service firms, and thus tends to have a large multiplier effect in the region (Cohen & Zysman, 1987). Manufacturing jobs in rural America typically required little job training or work experience, and young workers in many cases could take the jobs immediately out of high school. Finally, manufacturing jobs in rural areas have been critical to the survival of family labor farms because they are an important source of off-farm income (Browne, Skees, Swanson, Thompson, & Unnevehr, 1992).

Since 2000, however, the percentage of jobs in the manufacturing sector has declined from 25% to less than 10%. The Midwest has been hit especially hard–manufacturing employment dropped from 35% to 15% of the jobs in the region. Job losses in the manufacturing sector accelerated during the last two recessions. There appear to be several reasons for this decline. Technological change has contributed to the decline in manufacturing jobs, as productivity has increased rapidly during this period. Globalization and outsourcing have decreased the number of manufacturing jobs. Acquisitions and mergers have led to the closure of many manufacturing plants as firms consolidate their operations. There is a general belief that the manufacturing jobs lost over the past few decades will not return (Reich, 2009).

In addition to these employment changes, the US is experiencing a significant population shift away from nonmetropolitan areas, even in those areas adjacent to metropolitan regions (Hertz, Kusmin, Marre, & Parker, 2014). Aside from the Intermountain West region, most rural areas have experienced a population loss since the Great Recession. Twice as many nonmetropolitan counties experienced population loss compared to gains. Outmigration was worse during the 1980 recession, but the natural increase was lower after the most recent recession, which meant that population increases could not compensate for the losses. Even during 2015 when middle-class income grew at a record pace and poverty rates declined in the US, rural areas continued to suffer declines (Tankersley, 2016). Population losses coupled with employment changes present new obstacles to rural development.

There is a rather large body of literature on the impacts of deindustrialization on individual workers (Brand, 2015; Bluestone & Harrison, 1982; Dudley, 1994; High, 2003; Mah, 2012). Much of this literature has focused on the immediate and short-term impacts of job loss on individual wages, benefits, health, and family structure (Brand, 2015). Most of this research has been conducted in urban settings (such as Detroit or Flint), while there are fewer studies on the impacts of deindustrialization on rural areas. The loss of a major employer in a rural area, however, may have more devastating local impacts because of the lack of economic diversification and the commuting distance to urban labor markets for workers (Lyson & Falk, 1993). Research suggests that most workers who lose their job due to plant closings remain in their community. Lack of information about available jobs and depressed housing values tend to restrict the options available for most dislocated rural workers.

Twenty-five years ago, Glasmeier and Howland (1995, p. 221) raised an important question about the future of rural America: “Will service-led growth be the basis for the next phase of development in rural America?” Their analysis demonstrated that there was little growth in export-oriented services, such as producer services, in rural America. Goe (2002) did find development of producer services in a relatively few growth nodes in rural areas during the 1980s, but it was not widespread. The size and density of the population in rural areas are major constraints to the growth of the service sector. Glasmeier and Howland argued that deliberate policies were needed to integrate rural communities into the information age for them to benefit from the rise in the post-industrial society, but to a large extent this has not happened.

Post-industrial theory suggests that there is a natural evolution from agriculture to manufacturing to services (Bell, 1973). Although this restructuring is occurring at the national level, there may be considerable variation across space. This shift from manufacturing to services is supposed to be accompanied by the rise of a professional and technical class and the centrality of technical knowledge. The service sector includes a variety of industries, including personal, business, communications, health, and educational services. Educational requirements for these industries are often higher for these industries than the traditional manufacturing sector.

The rural workforce may lack the required skills for many of these positions. Growth of the service sector remains strongly linked to goods-producing activities (agriculture and manufacturing). As labor was replaced by capital and new technology in farming, new service industries expanded to support this sector (e.g., fertilizer sales, tractor repairers, crop dusters, truckers). Agriculture was not off-shored, but was automated. As the number of farms has declined dramatically since the 1940s, employment in industries supporting agriculture on both the input and output/processing sides have increased (Fite, 1981).

Manufacturing industries also are strongly linked with a variety of service industries, including trucking, communication, accounting, etc. Technological change, especially computerization, has increased productivity and led to fewer workers in manufacturing. Globalization and the loss of routine jobs in the manufacturing sector to low-wage countries have contributed to the economic restructuring (Bluestone & Harrison, 1982). The result of these processes is that there is a smaller manufacturing base to support service industries. In most metropolitan areas, the service sector has grown rapidly, especially in sectors such as health care and professional services.

Technology, especially the internet, has begun to have an impact on employment in many industries within the service sector. Several retail chains have made decisions to close stores due to changing consumer patterns. These stores closings often have a disproportionate impact on smaller communities with less consumer demand. In 2016, Walmart closed 154 stores in the United States accounting for the loss of more than 10,000 jobs (Tabuchi, 2016). Store closings were concentrated in small, rural communities in the South where Walmart had experimented with their Express stores (smaller versions of the larger Supercenters). For more than a century, Sears has played a pivotal role in small town economies has shuttered stores in the last few years. In 2016, the retailer announced it would close at 68 Kmart locations and 10 Sears Stores. Less than a year later, it closed another 150 stores (Peterson, 2017a). JC Penny also has experienced declining sales and closed 138 stores (14% of its locations) accounting for about 5,000 jobs in 2017 (Peterson, 2017b).

Deindustrialization has scarred many rural communities in the Midwest and the Southeast. Osha Gray Davidson (1990) argued that the disintegration of rural communities was leading to the rise of “rural ghettos.” The process parallels William Wilson’s (1987) analysis of urban ghettos in the United States. For Wilson, the rise of the urban ghetto was a result of the loss of manufacturing jobs and population decline in the inner city. As a consequence, these neighborhoods were socially isolated, and poverty became more concentrated.

Although the concept of ghettos usually is applied only to communities with predominately a minority population (Duneier, 2016), I believe the concept also can be applied to rural places inhabited predominately by the working class today. The loss of manufacturing jobs has weakened the economic base of manufacturing-dependent rural communities and accelerated the selective out-migration of educated, young workers (Carr & Kefalas, 2009). Families that remain face increased social isolation, which can result in less social support, insufficient job information, and less access to external resources. In addition, many of these rural places are experiencing more social problems, especially drug use, and crime. Deindustrialization stigmatizes rural ghettos and makes it more difficult to attract businesses because employers may consider workers’ skills as outdated. This stigmatization of rural areas increases the likelihood that businesses avoid investing there and youth are less likely to remain.

A formidable obstacle in many rural regions to developing the service sector is the communication system, especially broadband, which is a key to the infrastructure for post-industrial society. Just as the highway and rail systems were central to the growth of goods producing industries, the broadband has become the most important infrastructure element for the information society.

Nonmetropolitan areas also have difficulty promoting service industries because of the low density of consumer demand and distance to metro areas. It has long been recognized that service industries can be delivered more easily, and profitably, to areas with higher population concentrations (Johansen & Fuguitt, 1984). Major stores, restaurants, and specialty shops are more likely to be in dense, urban areas. Similarly, many professional services are in urban areas are close to major business clients.

There are a few exceptions to this trend in service employment in rural areas. Recreation and retirement areas have experienced significant growth over the past 20 years. In a few cases, producer services have concentrated their activities in rural areas, especially the type of back office activities that can be easily moved to these areas.

In this paper, I examine the trends in manufacturing and service employment in nonmetropolitan areas of the US from 2000 to 2017. I am especially interested in assessing how communities that lose major manufacturing facilities transition to the service sector and the factors related to the growth of the service sector in nonmetropolitan areas.

Data and methods

This analysis focuses on nonmetropolitan areas in the US. I use the 2013 census definition of nonmetropolitan areas so as not to include regions that have been recently included in the metropolitan definition. I also use the Economic Research Service (ERS) typology of economic dependence to define different types of rural communities. Manufacturing-dependent counties that are defined as counties as having annual earnings of 25% or more of total county earnings derived from manufacturing from 1998 to 2000. There are 585 nonmetropolitan counties that were considered manufacturing dependent. Although the ERS updated these files for 2015, I use the older definition. The number of manufacturing-dependent nonmetropolitan counties has declined significantly. In the 2015 definition (based on 2010–11 data), there were only 351 manufacturing-dependent counties in nonmetropolitan areas.

Data on employment and wages are drawn from the Quarterly Census of Employment and Wages (annualized data) that is collected from the Bureau of Labor Statistics (BLS). I am especially interested in the period over the past two post-recessionary periods (2000 to 2017). I include additional data on education, poverty, income, and other demographic data drawn from the American Community Survey (ACS).

Analysis

I first examine changes in the manufacturing sector in nonmetropolitan areas during the 2000 to 2017 period (Table 1). More than three-fourths of the nonmetropolitan counties lost manufacturing jobs during these years. The average nonmetropolitan county lost more than 600 manufacturing jobs. During the time period, manufacturing-dependent counties lost, on average, almost one-third of their manufacturing jobs in these regions. Non-manufacturing counties lost a much smaller proportion of jobs – about 14%. In addition, wage increases were much lower in these manufacturing-dependent counties over this period – 57% versus 69%.

Table 1. Economic restructuring and manufacturing employment in nonmetropolitan areas, 2000 to 2017

	Employment Change	Establishment Change	Wage Change
Manufacturing Dependent	−.30	−.16	.57
Non-manufacturing Dependent	−.14	−.10	.69

It is useful to identify specific counties that have been devastated by the loss of manufacturing jobs (Table 2). In Table 2, I identify the ten nonmetropolitan counties that lost the highest number of manufacturing jobs from 2000 to 2017. As you can see, most of the largest manufacturing losses were in the rural South. Surry County, North Carolina is one of the most depressed regions in the state and suffered from the loss of textile jobs throughout this period. In Lee County, Mississippi, the Sara Lee Corporation closed its pork processing facility, which employed as many as 2,500 workers at one time. Throughout the region, the low-skilled jobs, such as textile production and meat processing, were the most likely to be lost.

Table 2. Top ten nonmetropolitan counties change in manufacturing employment, 2000–2017

County	Manufacturing Employment
Surry County, NC	−8,612
Lee County, MS	−7,591
Muskingum, OH	−7,266
Franklin County, PA	−10,229
Rutherford County, NC	−6,755
Henry County, VA	−6,466
Litchfield County, CT	−6,433
DeKalb County, AL	−6,349
Pittsylvania, VA	−6,083
Cleveland, NC	−5,834
Scotland County, NC	−5,449

Most of these counties in the rural South have historically had a low educational base. Out-migration was selective, with the younger and more educated workers moving to other region where there were opportunities for economic mobility. As a result, manufacturing jobs in these counties tended to be low-skilled and low-paying. For example, the textile industry relied heavily on low-wage workers and many of these jobs were lost to regions in the global economy that offered cheaper labor.

Race is also strongly related to the loss of manufacturing jobs in these counties. Many of the regions experiencing the loss of large number of jobs tended to have a large minority population.

Another way to look at this issue is to examine the percentage of manufacturing jobs lost over this time period (Table 3). All top ten counties have lost more than three-fourths of their manufacturing jobs during this period. Most of the counties are relatively small, so the loss of a major employer can devastate the regional economy. Two examples are Jenkins County, Georgia, and Monroe County, Ohio, both of which lost almost their entire manufacturing base during this period. Jenkins County lost most of its manufacturing workforce during the Great Recession. This has had a ripple effect throughout the region as retail stores closed, and even the hospital had to make layoffs.

Table 3. Top ten nonmetropolitan manufacturing-dependent counties percentage change in manufacturing employment, 2001–2015

County	% Change Mfg. Employment
Monroe County, OH	−.99
Harney County, OR	−.98
Chicot County, AR	−.98
Lassen County, CA	−.97
Jenkins County, GA	−.96
Clay County, AR	−.95
Bledsoe, TN	−.94
Perry, KY	−.94
Massac, IL	−.93
Wolfe, KY	−.93

Monroe County suffered a major loss of manufacturing jobs later, in 2013. Ormet Aluminum closed its operation and approximately 1100 workers lost their jobs in the region. Ormet located in Monroe County in 1956, primarily because of the low electricity prices. The company experienced several ownership changes and low aluminum prices, but ultimately it was a rate increase for its electricity that ultimately led to the decision to close the plant. The facility has been sold and a steel mill plans to operate there, but it will only hire a small number of workers.

Obviously, these counties have less diversified economies. Location of a large manufacturing firm in a region may pull workers from other employers and make it more difficult for small firms to survive. At the same time, the lack of diversification makes it much more difficult to grow a sustainable local economy.

Of course, there are exceptions to this general pattern of manufacturing job loss in rural areas. Following the Great Recession, Midland County, Michigan experienced a significant growth in manufacturing during this period. Midland is the home of Dow Chemical and the firm serves as an incubator for new startups in the region.

Next, I examine the change in average weekly wages for manufacturing workers from 2000 to 2017 (Table 4). Change in manufacturing wages can be the result of new firms adding jobs, existing firms expanding or declining, or existing firms reducing their workforce or leaving altogether. As can be seen from Table 4, wage loss was not as concentrated in specific regions, but is much more widespread. In a few cases, the decline in manufacturing wages appears linked to the loss of major employers, such as in Monroe County, Ohio.

Table 4. Top ten nonmetropolitan counties weekly manufacturing wage loss, 2000–2017

County	Average Weekly Wage Loss ($)
Mason County, TX	−481
Monroe County, OH	−308
Morehouse Parish, LA	−275
Jack County, TX	−253
Henry County, IN	−193
Mingo County, WV	−165
Harney County, OR	−139
San Juan County, UT	−122
Perry County, KY	−111
Bath County, VA	−80

In the next analysis, I examine the relationship between change in manufacturing and services employment. Approximately three-fourths of the nonmetropolitan counties experienced a decline in manufacturing jobs from 2000 to 2017. At the same time, more than 37% of the counties lost service sector jobs. Approximately 30% of the nonmetropolitan counties lost both manufacturing and service sector jobs during this time period. Almost all of the counties experiencing both manufacturing and service employment decline were in either the Black Belt of the rural South or the Upper Midwest.

These counties with declining employment in both manufacturing and service employment are much more likely to be losing population than other nonmetropolitan counties. From 2000 to 2010, these counties were losing population at a rate of 2%, while other nonmetropolitan counties were growing at close to a four percent rate. From 2010 to 2015, the population loss increased in most nonmetropolitan counties, but was larger in these counties losing manufacturing and service employment. These disparities were due almost entirely to differences in migration rates, not birth or death rates.

A good example of this decline in both manufacturing and service employment is Wood County, Wisconsin. Since 2000, Wisconsin’s paper industry has collapsed, losing about 35% of the paper mill jobs during this time. Among the 34 mills that have closed, two are in Wood County. The county lost about 50% of its manufacturing jobs during this period. The loss of these jobs had a dramatic impact on the community. The closing of Domtar paper mill in the Village of Port Edwards resulted in a loss of 750,000 USD in property taxes for the Village. At the same time, Wood County lost approximately 2300 jobs in the retail sector from 2000 to 2017. In the most recent announcements of retail closings, JCPenney is closing two stores in Wood County.

To examine the broader trends between the manufacturing and service sectors, I regress change in service sector employment from 2000 to 2017 on change in manufacturing employment, while controlling for several other independent variables (Table 5). Education should be positively related to job growth in the service sector. Many service sector jobs require a college degree, so I include a measure of the percent of adults with a bachelor’s degree or higher (2012–16). Conversely, many of the service sector jobs, especially in the retail sector, rely heavily on workers with less than a high school degree. I also include a variable for the percent of adults with less than a high school diploma (2012–16). In addition, population density and change should also influence service sector growth. Communities with higher population density and are experiencing higher rates of population growth should have higher rates of service sector job growth.

Table 5. Regression analysis of service sector employment growth, 2000–2017

	Unstandardized Coefficient	Standard Error	Beta
Constant	−999.460***	189.211
Mfg. Employment Change	−.106***	.030	−.081
Population Density (2010)	1.568***	.486	.076
Population Rate Change
(2000–2010)	44.199***	3.840	.286
Population Rate Change
(2010–2015)	66.273***	7.420	.219
% Adults with bachelor’s
Degree	52.263***	6.331	.245
% Adults with less than
HS Degree	21.710***	6.242	.099
R-Square	.298

*** Significant at.001 level.

Overall, change in manufacturing employment is negatively related to change in service sector employment over the 2000 to 2017 period (Table 5). Although the relationship is statistically significant it is not exceptionally strong, which indicates that other factors are influencing the growth or decline of the service sector in these regions. As expected, population change rate and density are positively related to job growth in the service sector. Similarly, education level, especially the percent of adults with a bachelor’s degree, is positively related to job growth.

Conclusions

The loss of manufacturing jobs (and population loss) and the lack of growth in the service sector presents new obstacles to rural development. The findings suggest that manufacturing still does matter in most rural communities. A significant proportion of the rural communities that lost manufacturing jobs also lost service sector jobs. The manufacturing sector has a strong multiplier effect and is linked to suppliers and other service firms in the region. Over time, however, the impact of the manufacturing industries may decline as the proportion of jobs in this sector declines.

The service sector is also undergoing a major restructuring process. The loss of traditional department stores may create a significant job loss in many rural areas. The pandemic of 2020 could exacerbate these problems as many small businesses may have a difficult time surviving the health crisis.

In response to deindustrialization, rural communities tend to adapt in several ways that exacerbate conditions. As local government revenues decline, officials often decrease spending on education and infrastructure that is so critical to economic development (Lyson & Falk, 1993). Desperate to create new jobs, rural communities may offer tax incentives and subsidies to attract low-wage employers (Green, Haines, Dunn, & Sullivan, 2002). As the competition becomes steeper to attract these employers, the costs often exceed the benefits. These responses may lock in the spiral downward in rural communities; the ghetto in effect becomes self-perpetuating.

The evidence suggests that not only are most rural communities losing manufacturing jobs since 2000, a large percentage are facing a shrinking service sector as well. Almost one-third of the nonmetropolitan counties lost both manufacturing and service sector jobs during this period. For the most part, these regions do not have a large agricultural sector. In response to the declining economic base, out-migration accelerated during the last decade. Individuals with less education and fewer job skills are less likely to migrate.

There is a question of whether we should attempt to save rural ghettos or just let them die. Some economists have argued that rural residents are often willing to tradeoff wages for noneconomic benefits (such as recreation opportunities and environmental quality) of living in rural areas, so there is no need for public policy to protect rural communities (Roback, 1988). This argument, however, ignores the fact that many rural residents (especially children) are not making a choice to remain there, but must pay the costs of lack of access to health care and educational resources. It also overlooks the critical role these communities play in managing our natural and cultural resources in these regions.

The typical response to the condition of rural America is that we should encourage dislocated workers to move to cities where there are more job opportunities. Much of our public policy today is premised on an individual mobility principle–it is easier to encourage people to move to where the jobs are than increase economic activity in these areas (Imbroscio, 2012). Yet, many people affected by deindustrialization remain in their communities due to social support, lack of resources, or the absence of job information or prospects in other places. The evidence also suggests that workers who lose their jobs are not necessarily better off when they move (see Huttunen, Moen, & Salvanes, 2018). Many times, they have strong social support networks that enable them to survey through the tough times. Moving to a city without any social support can increase workers’ vulnerability. Thus, we need to focus on alternative strategies for these communities that are not likely to be competitive in global markets.

What are some alternative local responses to rural ghettoization? In response to the negative impacts of capital mobility, some rural communities have turned to worker- or community-ownership (Williamson, Imbroscio, & Alperovitz, 2002). Most manufacturing firms moving to low-cost regions are still profitable, but they are moving because the firm can make even higher profits by moving or consolidating production facilities. Although worker- and community-owned firms account for a relatively small share of the US economy, they can be critical to communities where they are located. They are also rooted in the community and not likely to move elsewhere. Worker- and community-ownership model can be an effective strategy for promoting service sector businesses as well. In some small towns, there may be only one restaurant, grocery store, or hardware store. These types of businesses may be retained through a cooperative model.

Increased immigration to rural communities is another strategy for countering the population and employment trends in rural regions. Several states and cities (e.g., Dayton, Ohio and Utica, New York) have developed immigration-friendly campaigns to attract new residents. Immigration can help communities maintain their schools, build new businesses, increase demand for housing, and bring new capital and ideas to the region. It also may provide the labor supply that may be lacking in rural communities.

Another common response to deindustrialization is job training of dislocated workers. There are several government programs that enable workers who lose their job through plant closings to receive benefits for retraining. The research on the effectiveness of this re-training suggests that it may have a limited impact on the reemployment of displaced workers. One of the problems is that the training programs are not tightly linked to the local demand for workers. In other words, workers may be trained for jobs that are not available in the local labor market. Workers either migrate to where the jobs are or take jobs in the local area that do not require the skills in jobs for which they were trained. Linking training programs to local opportunities will improve the chances of workers remaining in the community and continuing to contribute to the local economy. Although rural communities have lost manufacturing plants over the past two decades, many of them remain largely dependent on the manufacturing sector. The jobs that remain, however, tend to demand workers with a higher level of skills than has been the case in the past. Innovative strategies are needed to provide job training in these settings where the population density is low and there are a relatively small number of positions in any occupation (Green, 2007). In particular, it is important to provide employers with more incentives to participate in apprenticeship programs and to increase the job training they provide to their employees.

It is probably unrealistic to think that these alternatives to deindustrialization can effectively overcome the structural forces that are affecting rural communities today. They do provide alternatives that cushion the impacts and provide a new logic for economic development in these settings. The focus becomes less on attracting capital and population to the community and more on building a local economy that is resilient and sustainable.

Disclosure statement

No potential conflict of interest was reported by the author.

Additional information

Funding

This research was funded through the support of the University of Wisconsin-Madison, Office of the Vice Chancellor for Research and Graduate Education.