Globalisation has raised living standards worldwide but has also fuelled concerns that international competition is eroding middle-class incomes and widening inequality within countries. The column shows that the international relocation of production – the shift of export capacity from higher- to lower-income economies – systematically increases within-country income inequality in the countries that originally specialised in relocated products by compressing the income shares of the middle deciles and raising the top share.
Globalisation has lifted hundreds of millions of people out of poverty and made consumer goods dramatically cheaper in rich countries. However, it has also generated deep anxiety about wages, job security, and fairness. Over the past three decades, many advanced economies have experienced factory closures, downward pressure on wages in manufacturing, and the rise of political movements demanding protection from foreign competition. Across advanced democracies, voters in regions most exposed to import competition have become more likely to support parties sceptical of openness and multilateralism, and governments in those regions have increasingly embraced protectionist or nationalist industrial policies (Ottaviano et al. 2021).
Economists’ answers to these questions have changed over time (Levell and Dorn 2022). Twenty years ago, the mainstream view in rich countries was that trade played only a modest role in rising inequality. Krugman (2007) argued, however, that it was “no longer safe” to downplay trade: import competition from low-wage countries, and especially from China, was having visible distributional effects inside advanced economies. Since then, a major body of research has focused on the ‘China Shock’, showing that the surge of Chinese exports after the 1990s caused persistent job losses, lower wages, and slower recovery in the most exposed local manufacturing labour markets in the US and other countries (Autor et al. 2013, Acemoglu et al. 2016, Pierce and Schott 2016). Further research has shown that the negative wage effects of exposure to intensified international competition are not limited to advanced economies but also arise in developing economies (Goldberg and Pavcnik 2018).
The China Shock, however, is only one manifestation of a broader process: the sustained expansion of exports from a wide set of developing economies over the last few decades. In our recent paper (Alcalá and Romeu 2025), we extend the analysis of the China Shock to a broader, multi-country international relocation of production and exports process (IRP). Using detailed trade data that cover roughly 5,000 products and a comprehensive panel of 168 exporting and importing countries over the period 1996–2017, we investigate whether the global expansion of exports from lower-income economies raised within-country inequality.
Exposure to international relocation and inequality
To analyse these distributional consequences, we construct an index of exposure to IRP. For each country, the index measures the extent to which the products the country initially exported subsequently became exported, on average, by richer economies (“IRP to the North”, which yields positive relocation indices) or by poorer economies (“IRP to the South”, which yields negative relocation indices). Our empirical strategy parallels the shift–share framework in Autor et al. (2013), who regress local labour market outcomes in the US on a weighted average of Chinese import shocks. We proceed in two steps to adapt this approach to a multi-origin, multi-destination environment. First, for each product we construct a relocation index that tracks whether, over time, the exports of that product are relocating toward higher-income or lower-income producers. Second, for each country we aggregate these product-level relocation indices using that country’s initial export structure as weights. This produces a country-specific measure of exposure to IRP, which we refer to as the country’s relocation shock. Furthermore, to identify the causal effect of IRP exposure on inequality, we employ an instrumental-variable strategy and apply dynamic-panel estimation techniques. As our main measure of inequality, we use the Gini coefficient from the World Inequality Database. We utilise pre-tax data since our objective is to capture market-driven distributional changes prior to the operation of fiscal systems and other policies.
We find that greater exposure to IRP to the South leads to statistically and economically significant increases in within-country inequality. Inequality tends to rise in those economies that, at the start of the period, were specialised in products whose global exports subsequently relocated, on average, to poorer economies.
How important is international relocation for explaining within-country inequality?
The magnitude of the effect is quantitatively important. Figure 1 illustrates this point for countries with GDP above US$400 billion (2017 purchasing power parity). For each country, it reports the estimated effect of IRP on the Gini index over the 1996-2017 period. Figure 2 relates this same estimated effect to the actual change in Gini over the period. For the largest world economies – the US, China, and Japan – the estimated effect of IRP is 1.13, 1.90, and 1.29 percentage points, respectively. Since the observed increase in Gini over 1996-2017 in these countries was 4.78, 8.09, and 3.49 points, respectively, exposure to IRP “to the South” can account for roughly 20-40% of the rise in inequality in these countries. For other major economies that also experienced notable increases in inequality, such as Germany and South Korea, IRP to the South can explain smaller but still economically meaningful fractions of the observed increase (on the order of 10-20%). IRP is not the sole driver of rising inequality – technology, domestic institutions, and structural transformations, obviously matter – but it is a first-order explanatory channel.
Figure 1 Estimated effect of the relocation shock on the Gini index over the 1996-2017 period (percentage points)
Note: Economies with average GDP greater than $400 billion (2017 PPP).
Figure 2 Estimated effect of the relocation shock on the Gini index and change in the Gini index over the 1996-2017 period (percentage points)
Note: Economies with average GDP larger than $100 billion (2017 PPP) and change in Gini greater than 1 percentage point.
Who gains and who loses from exposure to relocation? Evidence by decile
Relying on a single inequality index such as Gini could be misleading because different indices weight parts of the distribution differently and can register the same underlying distributional change in distinct ways (Ravallion 2018). To address this potential problem, the paper goes beyond synthetic inequality measures and analyses how relocation shocks affect each part of the distribution. Using decile-level income data from the World Inequality Database, we estimate how exposure to IRP affects the income shares of each decile within a country. This tests the robustness of the Gini-based results and characterises the channels through which IRP operates.
Figure 3 reports the estimated coefficients of the effect of IRP exposure on each decile’s income share, along 95% confidence intervals. The pattern is highly informative. The income shares of the lower-middle and middle groups – roughly the 10th through the 70th percentiles – tend to fall with exposure to IRP to the South (recall that IRP to the South corresponds to negative values of the relocation shock and, therefore, a positive coefficient in Figure 3 means that greater exposure to relocation to the South reduces that decile’s income share). By contrast, the income share of the top decile increases with relocation to the South. Taken together, these results indicate a redistribution of income away from the broad middle and towards the top. This configuration is consistent with Milanovic’s (2016) account of a “second Kuznets wave” in advanced economies, in which the national middle class is squeezed while top income shares rise. Our results identify exposure to IRP to the South as a substantive force behind both the hollowing out of the middle and the rise of top income shares.
Figure 3 Estimated coefficients on the relocation shock (RS) index, by decile of the income distribution
Note: Because IRP to the South corresponds to a negative relocation shock, a positive coefficient in this figure implies that exposure to relocation to the South reduces the income share.
Final remarks on policy
Specific regions, sectors, and groups of workers absorb large negative shocks from international relocation of production, and these shocks are not automatically offset by new opportunities elsewhere in the economy. The inequality costs associated with IRP have contributed to a strong political backlash and to renewed protectionist pressures. If the main distributive pressures arise from exposure to particular relocation shocks rather than from trade intensity per se, then the appropriate response is not blanket protectionism but targeted intervention. A general turn towards protectionism is unlikely to solve the problem. Broad barriers to trade are costly for consumers and downstream firms, and it is unclear that they could produce a durable recovery of the affected jobs. Efficiency and fairness require more precise adjustment policies. This includes active labour market policies (retraining, wage insurance, mobility support) and place-based industrial and infrastructure investment in regions most exposed to relocation shocks. It also includes redistributive tools capable of counteracting the upward shift in income shares toward the top. These interventions are not without distortionary effects, as shown by Antràs et al. (2017). But if globalisation is to remain politically sustainable, the gains from trade must be paired with credible support for those who bear its concentrated adjustment costs.
Source : VOXeu
