Technology, Trade Sensitivity, and Labor Displacement.

• Author: Addison, John T.

John T. Addison [*]

Douglas A. Fox [+]

Christopher J. Ruhm [++]

This study examines the relationship between international trade, technology, and exposure to job displacement, using data on displaced workers as well as those at risk of job dislocation, for the two-year sample periods 1986-1987 and 1990-1991. Workers employed in manufacturing industries with elevated import penetration or high shares of R&D personnel appear to have increased rates of job loss. However, the risk of job loss is materially reduced when a relatively high proportion of employees report working with computers. The opposing effects on displacement probabilities of R&D employment intensity and computer-use carry over to the nonmanufacturing sector.

1. Introduction

   The last 25 years have witnessed a substantial growth in inequality in most industrialized nations. In the United States this has taken the form of an increasingly unequal distribution of earnings, whereas in Europe, it has been manifested in a dramatic rise in joblessness. [1] The sources of these trends have been the subject of much study, with two factors, (accelerating) technological change and heightened international trade, receiving most attention.

   Most economists have emphasized the role of technology, arguing that skill-biased technological improvements have reduced the relative wages of unskilled workers in the United States and, given lower levels of economic flexibility, have resulted in substantially higher unemployment for these groups in Europe. Only a minority of researchers believe that international trade has been of key importance, holding that the growth in trade with less developed countries has led to a dramatic increase in the stock of available unskilled workers, thereby reducing the returns to "raw" labor. Trade considerations have nevertheless dominated public debate and preoccupied policymakers; for example, much of the debate over the North American Free Trade Agreement (NAFTA) reflected concerns over its likely effects on domestic wages and job destruction, and on side agreements motivated by a desire to avoid a race to the bottom.

   Trade or technology may of course increase wage inequality without causing any unemployment. For example, once the Stolper-Samuelson theorem (1941) is bolted on to the factor price equalization model, the rise in the skill premium in rich countries--with the shift in their industry mix toward skill-intensive output--will cause firms to reduce the ratio of skilled to unskilled labor. The equilibrium differential will be one that induces a growth in unskilled worker employment in skill-intensive production that exactly offsets the shrinkage in unskilled worker employment in the declining, labor-intensive sectors. Although the biased technological change story can in principle be deployed to explain almost any pattern of relative wages, the most obvious impact of new machines that do the work of unskilled labor for less is to widen the distribution of earnings.

   Moreover, in addition to employment declines in the affected sectors, there will also be (unless wages display an exceptional degree of downward flexibility) permanent layoffs. Such displacements are of particular concern because they have been linked to increases in unemployment and, more fundamentally, to substantial and lasting reductions in earnings. [2] Yet significantly, despite the lack of research confirming the contribution of increased economic integration to labor displacement, special assistance has for almost 40 years been available through categorical programs geared to those displaced by international trade (namely through Trade Adjustment Assistance and more recently via the NAFTA Transitional Adjustment Assistance Program). [3] By contrast, special support has generally not been provided to workers dislocated by reason of changes in technology, even though this may be a more important source of job loss and reemployment difficulty. Rather, such support has applied more generally to displaced workers through such vehicles as Title III of the Job Training Partnership Act.

   This paper focuses on the link between trade sensitivity, technology, and exposure to job loss. Our goal is to provide a more detailed analysis of probability of displacement than is available in extant treatments that have largely confined their attention to trade variables. Subsequent research should examine the consequences of trade and technology-induced displacement by considering pre- and postdisplacement outcomes for displaced workers alongside the earnings development of the nondislocated population who may nevertheless have been affected by heightened international trade or changes in technology. The present exercise represents a first step toward understanding the job-loss component of changes in total employment.

   To anticipate our findings, persons employed in manufacturing industries with high, if not rising, import shares and in branches with more advanced technology (as indexed by R&D employment intensity and the computer investment ratio) confront a higher risk of displacement than their counterparts in sectors that are more sheltered from imports and less technologically advanced. Conversely, holding these factors constant, individuals working with computers are consistently less likely to be displaced. Thus, just as with recent explanations of the growth in the return to education, computer facility on the part of the worker appears to deliver a competitive edge. Interestingly, as with manufacturing, risk of displacement in nonmanufacturing industries is positively related to R&D employment intensity and is negatively associated with working with computers.

2. Previous Literature

   In recent years, extensive literature has investigated the impact of increases in international trade on the relative wages--and less frequently, the employment levels--of less skilled workers. The approach commonly adopted by labor economists has involved calculating the "factor content" of imports and exports to estimate the changes in factor endowments resulting from trade. These changes are then combined with assumed demand and supply elasticities of labor to determine how trade has affected equilibrium wages. [4] Conversely, trade economists have more frequently focused on changes in relative prices, arguing that the threat of import competition could affect prices, and thence domestic wages and employment, even in the absence of trade flows. [5]

   Generally, the effect of technological change on inequality has not been estimated directly. Rather, the proportion of the time trends not attributed to changes in trade or other observed factors (e.g., demographic characteristics and institutional events such as union decline) is assumed to result from skill-biased technological change. This residual approach is unsatisfactory, not least because errors in calculating the effects of the other factors will lead to potentially serious biases in the estimated effects of technology. As a result, researchers (Berman, Bound, and Griliches 1994; Machin 1996; Berman, Bound, and Machin 1997) have increasingly attempted to include direct proxies for the effects of technology. This is the approach used in our analysis. The technology studies generally report that increases in the ratio of skilled to unskilled worker employment are primarily an intraindustry phenomenon--rather than being explained by sectoral differences as might be the case if trade were the cause--and that the proxies for technology are important determinants of these within-industry movements. [6]

   As noted above, employment reductions, whether caused by trade or technology, are likely to be accompanied by falling wages and job terminations. Other things being equal, greater wage flexibility will be associated with smaller numbers of permanent layoffs. Conversely, increasing employment may coexist with rising import shares if foreign production is more elastic than domestic supply. Nevertheless, if product demand is controlled, growth in imports will generally be associated with some displacement. The effect of technological change is more ambiguous. New technologies reduce the demand for labor, at given levels of output, but product demand will increase as cost savings are passed on to customers. Displacement per se is yet more complicated because net employment could decline without permanent layoffs if firms are able sufficiently to reduce the rate of hires or increase quits. [7]

   The consensus of the early literature (Grossman 1987; Mann 1988; Dickens 1988) was that rising imports were responsible for, at most, only a small proportion of the domestic employment reduction observed in trade-sensitive industries. A number of more recent studies of the effect of trade on low skilled worker wages and employment also point to very modest effects (Sachs and Shatz 1994). Taken in conjunction with the phenomenon of an increased relative use of skilled workers across industry at a time of rising skill differentials, most observers have concluded that biased technological change, rather than international competition, is the dominant factor explaining changes in employment and relative wages over the last 15 years. [8] But if this is the orthodox view, a number of empirical analyses, using both quantity-based and price-based measures of import competition. point to much stronger trade disemployment effects (Freeman and Katz 1991; Revenga l992). [9]

   In marked contrast with the burgeoning literature on the "distributional" consequences of heightened trade and technological change, there has been less investigation of the causes of displacement. The issue of trade and displacement has attracted some scrutiny, but we know of no studies investigating the contribution of technology to worker dislocation.

   Using data from the Displaced Worker Supplement (DWS) to the January 1984 Current Population Survey (CPS), in conjunction with the...