IV. Trade and European Labor Markets

International Monetary Fund
Published Date:
August 1997
  • ShareShare
Show Summary Details
Charles Wyplosz

This paper examines the effect of trade with developing countries on European labor markets. First, it is important to note that the evolution of developing countries over the last ten years has been dramatic. One can imagine the impact on the world economy of China and India should they become as productive as Hong Kong and Singapore. Add to that the potential impact of the former centrally planned countries, which lie very close to Western Europe and are another potential source of inexpensively produced goods. One can well understand the perceived fear of workers and politicians alike of job displacement and reduced incomes—concerns that have led us to produce this paper.

Over the last two decades, world trade has increased rapidly—much faster than GDP—and competition from developing countries has increased. These developments have come at a time when labor markets are under pressure in the West. Real wages of unskilled workers have declined by more than 20 percent in the United States, and unemployment has increased considerably, especially among unskilled workers in Europe. The coincidence of this increase in trade and pressure in labor markets in the West is seen by many as a causal relationship. Issues such as dumping and unfair competition have become matters of concern; the European Commission, for instance, is actively tracking down cases of dumping and levying duties against offenders. There is a clear feeling that Europe is under threat. People are asking, “Is it true that this trend is going to accelerate? How bad is the situation?”

Economists tend to believe that trade is always welfare enhancing and that such fears must be misplaced. Yet it is well known that while trade is welfare enhancing in reality, there may be losers that need to be compensated; otherwise, political problems can ensue. It is not enough to point reassuringly to the traditional theory that free trade is always good; it is also necessary to look at the results. Table 1 shows that imports to Europe from the non-OECD area have actually declined as a percentage of GDP. Imports from East Asia have increased enormously, but they started from near zero (rates of growth are always enormous when they start from zero). Imports from the rest of the non-OECD area have either stagnated or declined. Thus, the first result is that presumably the trade effects must have been very small, because the trade volumes themselves are small. Unfortunately, there is very little research that attempts to assess, based on past trends, the effect of trade from developing countries on Europe. There is a small but quickly growing body of literature on this issue in the United States. And a number of papers typically find small and negligible effects (Lawrence and Slaughter 1993). Wood (1994) finds strong effects, but this result is very controversial.

Table 1.Imports(In percent of GDP)
Imports to:Imports from:19621992
North America2.31.8
East Asia 10.21.1
Other Asia0.30.2
Latin America0.70.4
Central/Eastern Europe0.60.7
United StatesOECD1.75.3
North America0.61.7
East Asia 10.11.8
Other Asia0.30.1
Latin America0.41.0
Central/Eastern Europe0.00.0
Source: OECD (1994).

East Asia includes Taiwan, South Korea, Hong Kong, Taiwan, and China.

Source: OECD (1994).

East Asia includes Taiwan, South Korea, Hong Kong, Taiwan, and China.

We find similar results: negligible effects—not strong, but significant—at the industry level. These effects are not easy to interpret, so the paper Damien Neven and I have put together is really the beginning of a complex investigation. The results may not be robust. Essentially, there are a variety of approaches, all of which have limits. The first question is how trade has affected European labor markets in terms of both employment and wages. Answering that question requires a counterfactual situation that allows researchers to observe what happens in the absence of trade with the less developed countries. Lacking such a situation, it is necessary to disentangle a number of effects that have changed the labor market situation and that have all occurred simultaneously. There are effects on trade, not only with the developing countries but also within the European Union and with other industrialized countries. There have also been labor supply effects, and there has been technological progress, both of which have affected the data on employment and wages. What we suspect (and I will try to give examples as I go along) is that these imports are not exogenous.

Since trade may well trigger technological progress, it is not enough to say that trade is of little importance. There is a problem of data in general, in particular in Europe, that limits the available evidence. Trade theory implies that, under a number of conditions (chiefly access to the same technology) trade will result in the equalization of factor prices or of wages. There are fears that, say, Italian workers will end up receiving the same wages as workers in Bangladesh. While this example is extreme, the tendency worries many people.

I have said that it is necessary to disentangle different effects occurring at the same time; let me give an example from our data. According to standard theory, Europe is characterized by more human and physical capital than the developing world, and it can be expected that output and employment in labor-intensive industries will decline and that imports of labor-intensive goods will increase when trade is developing. Labor-intensive goods can be characterized as originating in industries with a large number of blue-collar workers; this characterization is an approximation that is not without its difficulties, but it is also one that others have used before. For the two countries for which data are available, and for the industries characterized as intensive in unskilled labor (textiles, wood, chemicals, and metals), there is a decline in output and employment in some (but not all) areas, and therefore the results are not clear (Figure 1). Turning to price and wage data (Figure 2), and assuming that low wages correspond to low skills and high wages to high skills, a positive relationship can be expected between wage levels at the beginning of 1977 and relative price changes in each industry over the 1977-91 period, if trade is a driving force. However, the regression shows only a weak relationship, indicating that trade is not responsible for the outcome. In terms of technological progress, this result can be interpreted as biased toward industries where skills are higher (high-tech industries) and within which prices have declined. Such an interpretation explains a negative relationship but is not very strong. Looked at this way, as others have looked at the United States, the data suggest that technological progress dominates.

Figure 1.Import penetration - UK

Figure 1.Import penetration - Germany

Figure 2.Relative price changes (77-91) and initial wage (1977) - France

To cope with the problem of how to decide what comprises unskilled labor, we have divided all industries into three sectors: labor intensive; high tech; and services. Labor-intensive sectors, as I have mentioned, are made up primarily of blue-collar manufacturing. High-tech industries typically pay high wages and employ a large proportion of white-collar workers. The services sector is typically the sector of unskilled white-collar labor. Our research looks at how the relative prices of goods produced by these three sectors have evolved in three countries—Finland, France, and Norway—the only countries for which we have enough observations to be able to make meaningful comparisons. The data, from the OECD database, indicate that the relative prices of goods produced by both high-tech and labor-intensive industries have declined (Figure 3). A natural interpretation is that high-tech prices have declined because of technological progress and that labor-intensive or unskilled labor industries have seen their prices decline because of trade pressure. This interpretation may be correct, but the opposite may also be true. An examinination of output (Figure 4) and employment (Figure 5) shows the expected results for high-tech industries: assuming favorable technological change, prices will decline and output and employment will rise. For unskilled labor-intensive industries, trade shocks cause prices and output to decline relative to the average, and in fact output and employment slow. So our earlier conjuncture seems to be supported: a number of things seem to be happening simultaneously, and trade and growth are moving forward at the same pace.

Figure 3.Relative prices - Finland

Figure 3.Relative price - France

Figure 3.Relative prices - Norway

Figure 4.Real Production indices - Finland

Figure 4.Real production indices - France

Figure 4.Real production indices - Norway

Figure 4.Real Production indices - Switzerland

Figure 5.Employment indices - Finland

Figure 5.Employment indices - France

Trade theory predicts convergence in factor prices and especially in wages, because factor proportions tend to converge and the same factor proportions emerge, especially the ratio of capital to labor. We use the methodology from the endogenous growth literature, which looks at the convergence of output and income per capita and the capital-labor ratio, to explore this issue. The results show little evidence of convergence: there is some within Europe as a whole, and with Southeast Asia, but none between Europe and developing countries. Even when the developing countries are broken down into four groups—the World Bank classification—the methodology never works, except for the two intermediate groupings. The findings underscore what researchers have found in the endogenous growth theory: that very small groups of countries tend to move together, but most countries are proceeding at their own pace.

These findings work against the prediction that trade is exerting pressure on industrial countries but do not provide a direct measurement. Previously, our analysis has looked mainly at quantities—employment and output—but trade should not be looked at through quantities but through prices. The OECD job studies have done exactly that by looking at import penetration and trying to detect its effect on relative employment. However, this approach is problematic. For instance, developing countries could increase their competitiveness in the production of watches and challenge the industrial countries. If the industrial countries reduce their prices and costs in response, they may actually be able to prevent an increase in or to block imports altogether. Yet there could also be downward pressure on prices and wages that would not be evident in an analysis of quantities. Thus, the methodology used here takes import prices as a measure of competitive pressure from imports and adopts a simple general equilibrium model proposed by Grossman (1987) that offers industry-level reduced form equations. Wages and employment are estimated by sector as a function of variables meant to capture everything that affects wages and employment in each sector, and an import price index is included as the channel through which import competition enters. The databases used (from the OECD and the European Commission) allow for the differentiation of import prices according to whether the imports come from OECD or developing countries.

Problems aligning the databases preclude the possibility of presenting information on all industries at this stage. Instead, results are presented from a selection of 14 industries in the 4 largest countries of the European Union: Germany, the United Kingdom, France, and Italy. The industries were selected using two criteria. In the period 1975-80, exports in the industries must have been declining or imports increasing as fast or faster than the average of manufacturing. Second, there must have been an indication of trade pressure—either an export decline or an export increase (if imports had increased). Rigid criteria were not adopted for this selection process; instead, data were examined to identify those industries that were clearly under pressure.

The results are presented in Tables 2 and 3. In Table 2, the dependent variables are wages and employment. Both are estimated separately, but account is taken of possible correlations between the error terms. On the right-hand side are a number of variables likely to affect wages and employment, but only the effect of import prices—the measure of foreign competition—is not reported. A positive coefficient for both wages and employment is likely: if import prices go down, the assumption is that wages and employment will go down, too. Thus, a positive coefficient means that import competition is reducing either wages, employment, or both in these countries.

Table 2.Effect of Import Prices on Employment and Wages: Constrained Estimates, 1976-87 1
Dependent Variable: EmploymentDependent Variable: Wage
Prices of imports to−0.05
Prices of imports to
South East Asia0.0620.04−0.01−−0.000.00
Lagged dependent0.8720.6420.6720.8120.4220.3520.6220.512
Source: Neven and Wyplosz (1994).

Number of observations: France: 139; Germany: 147; Italy: 142; UK: 141.

Significant at the 5 percent confidence level.

Source: Neven and Wyplosz (1994).

Number of observations: France: 139; Germany: 147; Italy: 142; UK: 141.

Significant at the 5 percent confidence level.

Table 3.Effect of Import Prices on Employment and Wages: Unconstrained Estimates, 1976-87 1
Dependent variable: EmploymentDependent variable: Wages
(Imports from industrialized countries)
NACE 247
NACE 2520.3030.1130.11−0.1120.072
NACE 2600.233−0.1630.163−0.053
NACE 3110.3530.1530.1730.063
NACE 314−0.293−0.0720.042
NACE 321−0.353−0.293−0.2630.032
NACE 3450.032
NACE 3510.133
NACE 4360.443
NACE 4670.2030.063
NACE 4710.192
NACE 473−0.223−0.3830.112
NACE 4810.133
NACE 4830.343
(Imports from developing countries)
NACE 247
NACE 2520.183
NACE 2600.072−0.2030.083
NACE 311−0.383−0.142
NACE 3140.2830.1430.042
NACE 3210.283−0.1630.2930.263
NACE 3450.183
NACE 3510.143−0.143
NACE 436
NACE 467
NACE 471
NACE 4730.15−0.5030.142
NACE 481−0.113
NACE 483
Source: Neven and Wyplosz (1994).

NACE codes: 247: glass; 252: basic chemicals; 260: fibers; 311: foundry; 314: metallic construction; 321: equipment for agriculture; 345: electronic equipment (radio, TV, etc.); 351: automobile construction; 436: apparel; 467: wooden furniture; 471: pulp, paper, and cardboard; 473: printing; 481: rubber; 483: plastics.

Significant at the 100 percent confidence level.

Significant at the 5 percent confidence level.

Source: Neven and Wyplosz (1994).

NACE codes: 247: glass; 252: basic chemicals; 260: fibers; 311: foundry; 314: metallic construction; 321: equipment for agriculture; 345: electronic equipment (radio, TV, etc.); 351: automobile construction; 436: apparel; 467: wooden furniture; 471: pulp, paper, and cardboard; 473: printing; 481: rubber; 483: plastics.

Significant at the 100 percent confidence level.

Significant at the 5 percent confidence level.

Table 2 looks at what happens when the same effect is imposed on all 14 industries. Put differently, it is assumed that all the industries have the same wage and employment elasticities to import price changes. (These elasticities appear in the two first lines of Table 2.) By and large, nothing is found. Although most of the signs are positive, as expected, very few coefficients are significant, lending support to those who say that import pressure has few effects. In Table 2 each industry is allowed to react differently to import prices—that is, the same elasticity is not imposed across industries. The results are discussed below, first by sector and then by country.

Table 3 shows that two sectors are basically never affected by import prices and therefore are not affected by the pressure of import competition: glass and apparel. The 14 selected industries (listed in the notes to Table 3) are those we would expect to see affected by import competition. Some industries seem to be affected slightly: wooden furniture, and pulp, cardboard and paper. In some sectors, trade competition actually leads to increases in wages and/or employment. This development is the opposite of trade pressure: the more foreign import prices fall, the more wages and/or employment increase. (These are the sectors with negative coefficients.) All in all, there are 53 positive coefficients—the normal trade pattern—but 28 are negative coefficients, a number of them being strongly significant. These findings create a problem of interpretation. One possibility is that technological progress is also taking place, so that the findings capture something completely different from import competition. Another interpretation is that in response to trade pressure, a number of industries have hired more skilled workers and successfully expanded output in what resembles a trade reversal, which occurs when an industry adapts and moves away from its reliance on unskilled labor.

This table also shows that employment tends to respond more than wages to import pressure: there are more entries in the employment column in Table 5 than in the wages column. This result is to be expected when wages are sticky (as the prevailing view in Europe holds they are). Import pressure can also be expected to work itself out through employment instead of wages. Another important result is that there are more entries when import competition from OECD countries is considered than when import competition from developing countries is taken into account. This result matches the initial observation that competition from developing countries so far has not been significant. Finally, the details for a number of industries demonstrate the main conclusion: what is happening at the individual level is very different from what is happening in the aggregate. Thus, to understand the pressure from import competition, the analysis should not stop at the aggregate level (where indeed little is happening or the news is mostly good) but should go to the industry level, where things are very different.

For example, agricultural equipment is the industry most affected and has the largest number of entries. However, there is something troublesome here, for the OECD effect is favorable, with mostly negative signs. The more pressure there is from other OECD countries, the more wages and employment go up. But the effect from developing countries is negative for wages and employment. Firms in OECD countries that feel pressure from developing countries have changed their product mixes and started competing among themselves, with some success, and have probably become more efficient. Similar but less clear-cut results are shown for metallic construction and printing, but the results for the foundry industry are the exact opposite. Here, the effects are favorable for employment and wages with pressure from developing countries and unfavorable with pressure from industrial countries. These results are hard to explain.

Finally, the results differ from country to country. The country most affected is Germany, which not only has the largest number of entries, but shows consistent adverse effects. When trade competition pressure increases in Germany, it has negative effects on wages and employment. France is the country least affected, and most of the time it is affected favorably—that is, import competition has positive effects. What conclusions can be drawn from these results? What are the implications? The circumstantial evidence suggests that it is hard to find aggregate effects, or that aggregate effects are not powerful. There are fairly clear industry-level effects, but they are not systematic across industries or countries. Briefly put, each industry is a special case. These are hypotheses or conjectures for future work.

Technological progress is likely to be partly endogenous, and the number of firms that may be affected by trade pressure is probably small. So when people talk about trade pressure costing jobs and lowering wages, I am not sure which firms or industries are meant. Perhaps a few industries have reacted to import competition, and individual firms may have made some intelligent adjustments that have created a favorable industry-level effect; likewise, some firms may have failed to adjust, creating an unfavorable industry-level effect. Thus, the absence of systematic effects may be due simply to the fact that different firms react in different ways. However, it is very hard to draw general conclusions at this stage, except to say that findings show that trade does have an effect. The absence of systematicity means that it is impossible to make general predictions. At this stage, it certainly cannot be predicted which industries will suffer and which will benefit. But fairly significant effects are occurring, and even though, as theory holds, trade may be generally welfare enhancing, there exist sectoral distresses that can generate political opposition.


    GrossmanGene M.1987. “The Employment and Wage Effects of Import Competition in the United States.Journal of International Economic Integration 2 (Spring) pp. 1-23.

    LawrenceRobert Z. and Matthew J.Slaughter.1993. “International Trade and American Wages in the 1980s: Giant Sucking Sound or Small Hiccup?Brookings Papers on Economic Activity 2 pp. 161-226.

    NevenDamien and CharlesWyplosz1994. “Trade and European Labor Markets.Unpublished.

    WoodAdrian.1994. North-South Trade, Employment and Inequality: Changing Fortunes in a Skill-Driven World. Oxford: Clarendon Press.

This article is based on a longer paper, “Trade and European Labor Markets,” by the author with Damien Neven (1994). Available from the author.

    Other Resources Citing This Publication