What are the labour market effects of globalisation? This question dates back, at least, to the seminal work by Stolper and Samuelson (1941), but even today relatively little is known about the micro-level impacts of trade shocks on the job biographies of single workers. How are different individuals affected? Does it depend on their initial sectoral affiliation, location, and personal characteristics? Do they systematically adjust to globalisation by moving across industries, regions, or plants to mitigate import shocks or to benefit from export opportunities? Does this endogenous mobility occur smoothly, or does it involve disruptive unemployment spells? And what are the cumulated long-run effects of trade shocks? These are central concerns for policymakers who worry about the distributive consequences of trade liberalisations.
International trade theory shows that rising trade freeness reinforces the countries' specialisation patterns according to their comparative advantage. With labour-abundant countries like China or India on the rise, traditional Western market economies (like the US or Germany) will exhibit rising exports in skill- and technology-intensive manufacturing industries, and rising imports of goods from labour-intensive sectors. Traditional models with homogeneous workers and firms then predict these trade shocks to induce cross-industry worker flows out of the declining import-competing and into the expanding export-oriented industries. Newer approaches highlight intra-industry reallocations towards more productive firms and the corresponding worker flows within industries as an additional channel. The baseline models do not feature any wage dispersion or a differential impact of trade across equivalent workers, but those features arise once labour market frictions are introduced.
In particular, recent models feature assortative matching of heterogeneous firms and workers within industries, and sorting of workers across sectors given their industry-specific productivity realisations. A common theme of these structural models (e.g. Caliendo et al. 2015) is that globalisation induces workers to adjust to the exogenous shocks by sorting across industries, local labour markets, or plants. Empirical evidence on these differential effects is rare, however, especially about whether trade shocks cause endogenous mobility responses.
Rising trade exposure and individual earnings profiles: The empirical approach
In recent research, we analyse high-quality German micro data which allow us to follow the detailed job biographies of individual manufacturing workers over the period from 1990 to 2010 (Dauth et al. 2016). We set up a panel model where we trace the short-run impacts of the continuous rise in import and export exposure on their yearly earnings, and we also study the cumulated effects over a longer time horizon in a complementary cross-sectional analysis. Different estimation techniques are considered to address confounding industry-specific shocks, in particular an instrumental variable approach close in spirit to Autor et al. (2013) using third-country trade flows. Moreover, we include interacted individual-level fixed effects to capture unobserved heterogeneity, and to restrain the variation that identifies our central coefficients.
We start from an encompassing approach with dummies for every worker, and successively move to more demanding specifications with worker region, worker (local) industry, or even worker plant-fixed effects. When we only exploit the variation within worker-establishment spells, we eliminate all earnings effects stemming from sorting based on time-invariant characteristics and come as close as we can to the direct effects of trade. A comparison to the results of a model with plain individual-fixed effects – which exploits the workers' total earnings variations and thereby also captures indirect compositional effects – allows us to gauge how important endogenous worker mobility is in the adjustment to trade shocks. Moreover, by not only identifying effects within industries or regions, but across plants, we shed light on the relative importance of different adjustments.
The asymmetric response to export and import shocks
Our first key finding is that rising international trade exposure was beneficial for German manufacturing workers at large. Interestingly, this conclusion is quite different from the results of an influential study by Autor et al. (2014), who argue that the rise of China had strongly negative effects on the cumulative earnings (and other labour market outcomes) of American manufacturing workers. Comparing our results for Germany with their results for the US, we consistently find that rising import penetration per se adversely affects individual earnings.
Yet, unlike in the US case, this is more than offset by a positive causal effect of rising export opportunities. The average manufacturing worker in Germany benefited from the rising trade exposure, but globalisation has at the same time led to rising inequality within Germany.
Turning to the novel issue of endogenous mobility, our data show that rising import competition triggers substantial ‘push effects’ out of the exposed industries. However, the corresponding ‘pull’ effects of rising export opportunities appear much weaker in comparison. Other types of induced mobility, such as intra-industry adjustments across regions or plants, appear to play a weaker role in the response to trade shocks.
Figure 1. Labour mobility patterns of German manufacturing workers
We graphically illustrate the mobility patterns in Figure 1. The dark areas show that, over both decades in our observation period, around 60% of the workers in our sample have not switched their employer – that is, they work for the same plant at the beginning and at the end of the decade. That leaves roughly 40% of individuals having some kind of job change, which decomposes into different types of mobility. Less than 10% switched jobs within their original manufacturing industry (the medium grey area), which includes intra-industry plant switches in the same and in different regions. More than 30% changed industries when switching their employer. That is, conditional on switching, more than 75% of this mobility occurs was across industries, and less than 25% was within the same industry. This mobility can mean taking up a job in another manufacturing industry (light grey area), or moving to the service industry (white area). This latter response is by far the most common mobility pattern, and this also seems to be the explanation for the asymmetric response to trade shocks – displaced workers who are pushed out of import-competing industries do not primarily churn within the manufacturing sector and get pulled into export-oriented branches; they are pushed out of manufacturing altogether.
The push effects of import shocks: Who gets hit? How do they perform?
Digging deeper into the induced mobility responses, we then investigate in more detail who leaves the import-competing industries, and how the movers perform relative to the stayers in their subsequent careers. Figure 2 illustrates the answer to these two important questions.
Figure 2. Industries with top 25% net import exposure – movers versus stayers
We focus on workers in highly import-exposed industries, and distinguish between those who stayed within their original industry and those who eventually switched at some point. The left panel plots the distribution of their unobserved industry-specific ability levels, as proxied by the worker plant-fixed effect from their original match, separately for the movers and the stayers. These distributions show that the movers tend to be negatively selected – import shocks push the relatively less able workers out of the exposed sectors, while workers with higher industry-specific ability are retained.
The right panel of Figure 2 then compares the labour market performance of the movers, before and after their move, relative to the stayers. There we follow an event study design, and define the industry switch as the respective ‘event’ (at time 0) in the employment biography. Figure 2 shows that both movers and stayers experience earnings declines which reflect their affiliation in highly import-exposed sectors. The decline in the pre-event period is much sharper for the later movers, however, which is consistent with them being negatively selected. More importantly, for movers we find that their earnings profile becomes flat in the years after the event, while for the stayers the decline continues. In other words, after the move, the respective worker is now less import-exposed, and is therefore able to stop the previous earnings drain. This is different for the stayers, for whom the high import exposure and the earnings decline go on. However, the representative mover is not able to make up for the cumulated earnings losses from the pre-event period by the relatively better performance later on. Eventually, the movers even seem to catch up to the stayers, but over the full decade the movers realise lower total earnings than the stayers.
Summing up, our key insight is that German manufacturing workers benefited at large from this particular globalisation episode. Yet, there have been winners and losers. Moreover, we find that individuals systematically adjust to globalisation, and we discover a notable asymmetry in the individual labour market response to positive and negative shocks.
Trade shocks do not seem to trigger much ‘voluntary’ sorting from import-competing to export-oriented manufacturing industries. The patterns we observe in the data are more consistent with a type of mobility that is ‘forced’ by job displacement and unemployment. The push effects out of import-competing manufacturing industries are not mirrored by comparable pull effects into export-oriented branches. Often, the direction of the move is to the service sector.
We also find strong heterogeneity across different types of workers. Younger and less-skilled individuals, for example, are hit harder by import shocks but also benefit more from export opportunities. Women and men are affected similarly from import penetration, but men seem to materialise the benefits of rising export exposure better than women, thereby fuelling the gender-wage gap. The basic upshot is that trade causes strong distributional effects in the German labour market, and our micro-level empirical findings might be informative for policymakers to design targeted policies that aim at those who benefit the least from globalisation.
Autor, D H, D Dorn and G H Hanson (2013) “The China syndrome: Local labor market effects of import competition in the United States”, American Economic Review,103: 2121–68.
Autor, D H, D Dorn, G H Hanson and J Song (2014) “Trade adjustment: Worker level evidence”, Quarterly Journal of Economics, 129(4): 1799–1860.
Caliendo, L, M Dvorkin and F Parro (2015) “Trade and labor market dynamics”, NBER, Working Paper 21149.
Dauth, W, S Findeisen and J Suedekum (2014) “The rise of the East and the Far East: German labor markets and trade integration”, Journal of the European Economic Association, 12(6): 1643–1675.
Dauth, W, S Findeisen and J Suedekum (2016) “Adjusting to globalisation: Evidence from worker-establishment matches in Germany”, CEPR, Discussion Paper 11045.
Stolper,W F and P A Samuelson (1941) “Protection and real wages”, Review of Economic Studies, 9: 58–73.