|Trends and Challenges for Work in the 21st Century|
Opinions and views in these papers are those expressed by the author(s). They are not to be taken as expressions of support for particular positions by the Department of Labor. Please do not cite these papers without prior permission of the author(s).
AN OVERVIEW OF ECONOMIC, SOCIAL,
AND DEMOGRAPHIC TRENDS
AFFECTING THE US LABOR MARKET
Robert I. Lerman
Stefanie R. Schmidt
The Urban Institute
Business investment in computers has grown very rapidly over the last 25 years. In 1973, business computer purchases accounted for .005 percent of GDP, and in 1996, they accounted for 3.1 percent of GDP. Computers now comprise about one-fifth of businesses new capital investment. This investment in technology has meant that one-half of American workers are likely to use a computer in their work, compared with one-fourth in 1987 (Madrick, 1998).
On average, technology has increased the demand for workers who have good math, communication, and teamwork skills. However, the impact of technology has not been uniform; it varies by occupation and industry and among firms within the same industry. Technology increases the skill demands of many manufacturing jobs, but may deskill some clerical work. Industries vary in their investment in information technology. Evidence shows that industries that invest the most in technology have the highest demand for skilled workers, and pay a higher wage premium to workers than other industries. Manufacturing industries that invest the most in technology have a higher demand for non-production workers (managers, secretaries, and janitors) relative to production workers (line workers).
The impact of technology varies considerably among firms within the same industry because firms differ widely in their organizational responses to technology. Technology is not simply an input into the production function. It changes the production function by affecting work organization and expanding the types of possible outputs. In the long run, a firms type of product, its work organization, and its level of investment in technology are complementary. Although firms choose a myriad of organizational responses to technology, it is heuristically useful to think of two types of strategies, customization and more traditional work practices (Piore, 1998). With customization, firms produce products in which quality and/or control of the production product are key. They invest heavily in information technology and employ a cluster of work organization practices that both expand workers decisionmaking power and flexibility and compensate workers for their performance. Firms that follow the customization strategy require the most skills from their workers and also experience the largest productivity gains from adopting computers.
Firms that choose more traditional work organization produce products that are less customized and they invest less in technology. Such firms do not fully transform their work organization in response to technology; rather, their method of work organization is closer to mass production. Firms that follow the second strategy do not demand as many skills from their workers as firms that follow customization and do not experience as large a boost in productivity from adopting new technology. They may adopt some new workplace organization practices, but do not fully transform their workplace. The choice of strategies varies widely among firms within the same industry.
The future impact of technology on the labor force will depend on whether more firms adopt the customization strategy in response to new technologies. Competitive pressures may push more firms to change their workplace organizations. However, many firms will find it difficult to transform themselves because of the resistance of workers, lower- to mid-level managers, and/or labor union leaders.
Recent Changes in Work Organization
Accompanying the growth in computer investment has been a movement away from the Ford/Taylor model of mass production, once a major form of work organization in the U.S. Under the mass production model, workers performed a narrow set of tasks and had little incentive or need to understand how their work fit into the larger picture. Monitoring quality was the responsibility of supervisors.
No single form of work organization has emerged as dominant, but most firms have adopted flexible forms of production. Workers often perform a broad range of tasks, thereby gaining the ability to respond quickly to changes in the production process and assuming some responsibility for the quality of the product (Piore, 1998).
Information technology was not the only factor contributing to movement away from the mass production model. During the 1970s, several changes in the macroeconomic environment led to increased levels of uncertainty and volatility in the demand for products. The macroeconomic changes included the opening of the economy to trade and deregulation. Additionally, the change from fixed to fluctuating exchange rates led to wide swings in the value of the dollar. The 1970s also witnessed more rapid shifts in consumer demand, which led to quicker product cycles, more products, and quicker supplier responses to changes in consumer tastes (Piore, 1998). Consumer demand also became more fragmented; the number of new food and household products introduced each year has increased 15- or 20-fold since 1970.
Heuristically, it is again useful to think of two types of strategies that were responses to the changing economic environment: customization and more traditional work arrangements. The nomenclature is based on Piores (1998) analysis of the case study literature on manufacturing firms; Bresnahan, Brynjolfsson, and Hitt (1998); and quantitative analysis of data from 303 Fortune 1000 firms in both manufacturing and services.
First, firms that chose the strategy of customization responded to rapid changes in consumer demand by producing products in which service was an important component that could be easily customized by controlling the production process. These firms have been heavy investors in information technology and have transformed their work organization practices. They organize work into self-directed teams, which gives workers more decisionmaking authority (particularly over pace and method of work). They also rotate workers among jobs, improving their understanding of the production process. In addition, these firms implement incentive systems to reward and encourage performance. Such firms also require workers to perform highly skilled tasks, such as reprogramming equipment for a change in the product line, so firms typically invest heavily in worker training and screening (Piore, 1998). According to one estimate, nearly 40 percent of firms follow the customization model (Osterman, 1994). The positive relationship between information technology and new work organization practices cannot be explained by industry type alone; the relationship exists both within and between industries (Bresnahan, Brynjolfsson, and Hitt, 1998).
Second, some firms choose to produce products in which service and producer control are not as important. These firms tend to use fairly traditional forms of work organization. Most firms have implemented at least one but not all of the nontraditional human resource practices that exist in the customization strategy (Osterman, 1994). Appelbaum and Batt (1994) find that the fraction of Fortune 100 firms with at least one employee involvement practice rose from 70 percent in 1987 to 85 percent in 1992.
Firms that follow the customization strategy experience greater productivity gains from the adoption of information technology than other firmseven firms within their own industry (Beede and Young, 1996; Brynjolfsson and Hitt, 1998). The customization strategy also results in large improvements in productivity when new organization practices deeply penetrate the firm (Cappelli et al., 1998; Ichniowski et al. 1997). However, firms that implement only some of the new work organization practices experience only marginal improvements in productivity.
Will More Firms Adapt Their Work Organization to Fully Exploit Technology?
Technologys future impact on the demand for skills and on productivity will likely depend on how many additional firms adopt customization as a form of work organization. David (1990) argues that a new technology can take decades to have its full impact on productivity because it takes some time for firms to adopt work organization to the technology. Such was the case with the adoption of electricity in the early twentieth century.
Although competitive pressures may push additional firms to adopt customization, many firms may find it difficult to reform their work organization. Piecemeal changes to work organization are easier for firms with a tradition of mass production to implement than customization (Piore, 1998). Management and union leaders may oppose such changes because they threaten their power and workers may have difficulty conforming to new modes of work organizations because old work norms have become ingrained (Cappelli et al., 1998; Bresnahan, Brynjolfsson, and Hitt, 1998). Also, decades of animosity between management and labor may prevent the cooperation necessary to reform work. Because they lack an institutional memory, greenfield (newly built) plants are more likely to adopt customization than other plants (Ichniowski et al., 1995).
Does Technology Change the Demand for and Returns to Skills?
The growth in information technology and changes in work organization have likely contributed to the rising demand for and returns to higher-order cognitive skills. Anecdotal evidence shows that new technologies and workplace practices often require workers to have good writing and verbal skills, good math skills, and good problem-solving skills. In addition, given the increased autonomy and responsibilities in many environments, there is evidence that so-called soft skillsmotivation, work habits, and so onhave become more important. Although there has been an economy-wide increase in the demand for skills, some jobs may have been deskilled by the changes in work organization and growth in information technology. In particular, some clerical work has become less skilled now that office technology, such as word processing software, can perform many of the most skilled clerical tasks.
Because of other sweeping economy-wide changes that were occurring at the same time as technological change and the change in work organization (including globalization, deregulation, and the decline of unions), it is difficult to empirically identify the magnitude of technologys effect on the rising skill premium.
Technology has played a role in the rising demand for and return to skill. Three papers (Bound, Berman, and Griliches, 1994; Allen, 1997; Haskel and Slaughter, 1998) find that industries that invest the most in technology have the highest demand for skilled workers and pay the highest relative wages to skilled workers. When the three studies are taken together, the result is robust across the manufacturing and service sectors, across different measures of technology, and across different proxies for unskilled and skilled workers.
The shortcoming of the industry-level studies is that they cannot say whether the adoption of computers per se increases the demand for skilled workers, or whether the adoption of computers is simply a proxy for other organizational changes in the firm. Recent evidence from worker-level data suggests that it is not the use of the computer per se that raises the relative earnings of skilled workers; the use of a computer is simply a proxy for whether the individual sits and thinks at his or her job (DiNardo and Pischke, 1997).
Mishel and Bernstein (1997) argue that technological change cannot explain the rising demand for and returns to skill. They believe that in order for the relationship to hold true, technological change should have accelerated between the 1980s and the 1970s, because the wage gap between skilled and unskilled workers grew more quickly during the 1980s than during the 1970s. They find no evidence of an acceleration in investment in technology in the 1980s. The discrepancy in timing between the introduction of technology and wage inequality may be explained by the fact that the new work organization practices that exploited technology became more widely adopted in the 1980s. Bernard and Jensen (1994) also argue that technological change cannot account for much of wage inequality; they show that the within-industry changes in the employment shares and wage shares of skilled workers were larger than the between-industry changes. They assume that all firms within the same industry would have similar technologies, which has been shown to be empirically false.
Many papers addressing the issue of skill-biased technical change have simply assumed the change was the residual that could not be explained by other factors, such as changes in demographics, changes in the composition of industries, or trade.
Given that whether and how firms adopt technology depends on their response to changes in the external environment (e.g., trade, deregulation), a research program that attempts to allocate contributions to wage inequality may nevertheless fail to achieve a consensus. For policy purposes it is more important to first determine what skills are valued in the economy, then to go about training workers in those particular skills.