OECD Employment Outlook 2026

Geographic Disparities in Jobs and Incomes

Report

OECD Employment Outlook

7 July 2026

Available in:

English
français

Download PDF

3. Places and people: The impact of structural change on local labour markets

Copy link to 3. Places and people: The impact of structural change on local labour markets

César Barreto

Alexander Hijzen

Carlo Menon

Antonela Miho

Abstract

The capacity of places to adapt to profound and rapid structural change in the labour market varies significantly, with some emerging stronger and more prosperous, while others experience persistent economic decline. These divergent trajectories carry important implications for the well-being of residents and contribute to growing geographic inequalities. This chapter documents how structural change is reshaping local labour markets by focussing on trade and technology shocks and their effects on the careers of individual workers (people), the composition of employers (firms) and the economic fortunes of regions (places). The chapter concludes by discussing the implications for place‑based labour market, industrial and regional policies.

In Brief

Copy link to In Brief

Local labour markets are experiencing deep and rapid structural change, driven by digitalisation, advances in artificial intelligence, evolving global trade patterns, and the energy transition. This chapter explores how structural transformation reshapes local labour markets and considers the implications for public policy and institutions. It focusses on local exposure to major global shocks, including China’s accession to the WTO, EU enlargement to Central and Eastern Europe, and the rapid spread of digital technologies such as artificial intelligence and robotics. The analysis assesses the impacts of these changes on regions, workers and firms, drawing on linked employer – employee data from 14 OECD countries over the period 2000‑2022 (Austria, Canada, Denmark, Finland, France, Germany, Hungary, Italy, the Netherlands, Norway, Portugal, Spain, Sweden, the United States).

Exposure to structural change

A region’s exposure to structural change depends crucially on its industrial mix.

Regions vary considerably in the degree of industrial specialisation. In the average region across the countries covered by the analysis, about 20% of activity would need to shift across sectors for the region to mirror the national industrial structure. However, disparities within countries are substantial. Some regions closely resemble the national average, while in others the share of activity that would need to be shifted can be 50% or more.
Regional differences in industrial specialisation lead to differences in exposure to structural change. A region’s exposure to trade is measured by the increase in import penetration in an industry from China and Central and Eastern European countries between 2000 and 2007 weighted by the initial employment share in the region (import competition from low-wage countries). Likewise, a region’s exposure to technological change is measured by global patents for digital technologies related to an industry between 2012 and 2021, weighted by its initial employment share in the region (digital innovation).
Regional differences in worker and firm composition lead to uneven exposure to import competition from low-wage countries and digital innovation. Higher-educated and foreign-born workers, who are more likely to sort into urban areas, tend to be more exposed to technology shocks and less to trade shocks. Large and high-wage firms are more exposed to trade shocks in manufacturing and technology shocks in all sectors, since such firms tend to be concentrated in urban areas.

Effects of structural change on places

Regional exposure to import competition from low-wage countries and digital innovation are important drivers of a region’s structural employment trends.

Regional exposure to import competition from low-wage countries is associated with persistent job losses in manufacturing while job gains elsewhere only materialise after some time. A one standard-deviation increase in regional trade exposure, which corresponds to moving from a region with average exposure to one around the 80th to 90th percentile, is associated with a reduction in manufacturing employment relative to the initial population of 0.5 percentage points (p.p.) across the Western European countries considered in the analysis and 1.25 p.p. in the United States by 2018. To put these effects in perspective, this amounts the loss of about one in ten manufacturing jobs in Western Europe and one in seven in the United States. However, job losses in manufacturing are more than offset eventually by job gains in non-manufacturing, especially producer services.
Regional exposure to emerging digital technologies is associated with a sustained increase in regional employment. These gains are driven by producer services and non-routine occupations. However, it is associated with job losses in regions exposed to digital innovation in industries with a high share of routine‑manual occupations (e.g. computers, machinery). This suggests that digital innovation is labour-augmenting overall, boosting demand for workers in complementary tasks, while acting as a labour-saving force in routine‑manual roles where tasks are more easily automated.

Effects of structural change on people

The effects of regional exposure to trade and technology on regional employment are decomposed in its different flow components: in or out of employment, local job reallocation across sectors, and internal migration across regions. For both trade and technology shocks, movements in and out of work represent the main margin of adjustment, while mobility between sectors and regions plays a comparatively minor role.

Changes in regional employment by broad sector in response to structural change are largely brought about by flows in and out of national employment. Trade and technology shocks lead to persistent increases in joblessness among incumbent manufacturing workers. The existing literature shows that the resulting social costs can go beyond unemployment and may include deprivation, social exclusion, chronic health issues and premature retirement. Job gains in non-manufacturing largely reflect inflows from inactivity, including young labour market entrants making the transition from school to work. Adjustment to structural change is largely generational, with older cohorts leaving and younger cohorts entering employment in different sectors.
Mobility between industries and regions plays only a minor role in adjusting to structural change. Net worker flows from manufacturing to non-manufacturing account for less than 15% of the decline in manufacturing employment due to exposure to low-wage import competition, while net worker flows between regions typically account for even less (except in Germany). Prior evidence for the United States suggests that workers who change industry had higher initial earnings, whereas those that did not had lower initial earnings and faced larger and more persistent earnings declines. For regional exposure to digital innovation, the contribution of net worker flows between regions to the increase in employment between 2012 and 2017 is even smaller (mobility between industries is not considered).

Effects of structural change on firms

The effects trade and technology shocks on regional employment also differ across groups of firms, e.g. high/low-wage industries, high/low-wage firms and small/large firms.

Import competition from low-wage countries shifts employment away from high-wage industries. Import competition disproportionately reduces manufacturing employment in industries with more generous wage‑setting practices, while non-manufacturing job growth is spread evenly across high- and low-wage industries. Digital innovation supports job creation evenly across high and low wage industries, while routine‑manual job losses are concentrated in low-wage industries.
Import competition from low-wage countries shifts employment to better firms within industries. When industry employment contracts, this tends to be concentrated in firms with wage premia below the average of their industry, typically lower productivity and smaller firms. By contrast, when industry employment expands, this tends to be concentrated in firms with above average wage premia, often high productivity and larger firms. This process of job reallocation supports industry-level wage and productivity growth.

Policy implications

On average, structural change requires considerable reallocation in exposed regions but leads to net job gains eventually. However, positive regional outcomes in the medium to long term can mask persistent negative effects on specific groups of workers due to limited geographical and sectoral mobility. From a policy perspective the main challenge is how to limit local vulnerabilities and anticipate change by fostering industrial diversification, investing in education and infrastructure and anticipating skills needs, while alleviating the social costs by supporting displaced workers, removing barriers to job mobility across sectors and regions and promoting job opportunities. This highlights the importance of combining employment and social policies with place‑based industrial and regional policies.

Introduction

Copy link to Introduction

Labour markets are undergoing profound and rapid structural transformations driven by digitalisation, advances in artificial intelligence, shifts in global trade dynamics, as well as the energy transition in many countries. Both open trade and technological progress have been key drivers of economic growth, rising living standards and well-being improvements. However, the capacity of places (or regions) to adapt to these changes varies significantly, with some emerging stronger and more prosperous, while others experience persistent economic decline. These divergent trajectories carry important implications for the well-being of residents and contribute to growing geographic inequalities – see Chapter 2. Addressing this multifaceted challenge requires a policy framework that integrates regional and place‑based industrial policies with employment and social policies that are equally place‑sensitive. Such an approach is essential to promote inclusive growth and mitigate the uneven impacts of structural change.1

This chapter examines the labour market consequences of structural change across industries in local labour markets and their implications for policies and institutions. It focusses on regional exposure to major trade and technology shocks, including China’s accession to the WTO, EU enlargement to Central and Eastern Europe, and the rapid diffusion of digital technologies such as artificial intelligence and robotics. These shocks have been central drivers of job reallocation between sectors, with effects that vary significantly across regions depending on their initial industrial composition and workforce characteristics.2 Adaptable regions adjust to structural change by creating sufficient jobs to sustain high employment levels, limiting the social costs of adjustment through anticipation and effective support for displaced workers, and promoting the reallocation of jobs toward high-quality employers.

The effects of structural shocks are analysed separately for regions, individuals, and firms, drawing on recent work by Dustmann et al. (2025[1]) and Autor et al. (2025[2]). Since the composition of workers and firms can change over time, structural change can have qualitatively different implications at the regional, worker and firm level. For instance, Autor et al. (2025[2]) show for the United States that, although regional employment recovered within a decade after the China shock, incumbent manufacturing workers suffered persistent earnings losses. This divergence reflects responses by firms and workers: manufacturing firms contracted through job losses and retirements, while service firms expanded by hiring new labour market entrants. In other words, structural change was partly accommodated through a generational shift, while job mobility between declining and expanding industries or regions remained limited. A key contribution of the chapter is to assess whether similar patterns emerge across countries in response to trade shocks and to what extent the effects of digitalisation shocks differ.3

The empirical analysis in this chapter makes uses of linked employer-employee data covering around 15 OECD countries over the period 2000‑2022. These include Canada, Norway, the United States, as well as 11 European Union countries (Austria, Denmark, Finland, France, Germany, Hungary, Italy, the Netherlands, Portugal, Spain and Sweden). Local labour markets are measured using small administrative units (TL3 regions) or commuting zones.

The chapter is organised in three parts consistent with the main building blocks of the conceptual framework visualised in Figure 3.1. It begins with a descriptive overview of the exposure of local labour markets to structural change in the form of trade and technology shocks (Section 0). The chapter then proceeds with an empirical analysis of the effects of trade and technology shocks on regions, workers and firms (Section 3.2). The chapter concludes with a discussion of the implications for policies (Section 3.3): i) regional policies that seek to promote regional employment or productivity (e.g. investments in infrastructure and education); ii) employment and social policies that provide support to vulnerable workers (e.g. income support policies, active labour market policies); iii) place‑based industrial policies directed at industries or firms (e.g. policies to promote the adoption of AI or renewable‑energy technologies).

Figure 3.1. Regional, industrial, employment and social policies are needed to address the local impact of structural change
Copy link to Figure 3.1. Regional, industrial, employment and social policies are needed to address the local impact of structural change

3.1. Regional specialisation and regional exposure to structural change

Copy link to 3.1. Regional specialisation and regional exposure to structural change

Regional exposure to structural change is analysed by focussing on the degree of regional specialisation across industries and how this shapes a region’s exposure to country-wide industry shocks due to trade and technology. In this section, local labour markets are defined using small administrative regions or, where possible, commuting zones. Small administrative regions (Territorial Level 3, TL3) typically correspond to the smallest subnational administrative units after municipalities, such as départements in France or provinces in Italy. Regions that are part of the same metropolitan area are combined following the OECD classification of metropolitan areas (Fadic et al., 2019[3]). Commuting zones are defined based on detailed information on the location of the workplace and the residence of the worker.4^,5

3.1.1. Regional specialisation

Economic activities tend to be spatially concentrated

Industrial restructuring is inherently a local process, driven by the spatial concentration of economic activities. The degree of industrial specialisation in a region is measured here using the Krugman index of regional specialisation (see Box 3.1 for details). On average across countries and regions, the index of regional specialisation is about 0.2 (Figure 3.2). This means that on average, 20% of employment would have to be reallocated across sectors for a region to mirror the sectoral structure of the national economy. However, there is substantial variation in the industrial specialisation across regions within countries, as indicated by the minimum and maximum levels of specialisation in each country. In some regions the industrial mix closely resembles that of the national economy, as indicated by the low minimum values of the Krugman index in the figure. In others, regional specialisation is much more pronounced, with maximum values around 0.5 or above.6

Figure 3.2. Industrial specialisation is very high in some regions
Copy link to Figure 3.2. Industrial specialisation is very high in some regions

Rescaled Krugman index of regional specialisation averaged across regions by country [0‑1]

The spatial concentration of economic activities arises from a combination of locational and technological factors. Location-specific characteristics include natural resources, infrastructure and skills, as well as public services. For example, mining is concentrated in areas abundant in natural deposits and logistic services cluster in port cities. Tradeable sectors, including manufacturing and tradeable services, especially benefit from spatial specialisation, by exploiting economies of scale and agglomeration effects (Ellison and Glaeser, 1997[5]; Rosenthal and Strange, 2004[6]). Geographic proximity among firms in related industries facilitates interactions within supplier networks and fosters knowledge spillovers. At the same time, such spatial concentration also creates disproportionate vulnerabilities to global shocks and shapes local socio‑economic conditions for extended periods in some places (Rice and Venables, 2021[7]).

Box 3.1. Methodology I: Measuring regional specialisation and exposure

Copy link to Box 3.1. Methodology I: Measuring regional specialisation and exposure

The exposure of regions to structural change depends on the extent by which regions specialise in industries that are subject to country-wide industry demand or supply shocks.

Measuring regional specialisation

The Krugman index of regional specialisation quantifies how different a region’s industry structure is from the national average (Krugman, 1992[8]). It is calculated as the sum of the absolute differences between the regional employment shares s of 3‑digit industries j in region r and their national counterparts:

$K_{r} = \frac{1}{2} \sum_{j = 1}^{J} | s_{j r} - s_{j} |$

The Krugman index ranges from zero to one, taking the value zero when a region’s industrial structure matches the national average. When the national average includes the region itself, the index is strictly less than one, but under complete specialisation the maximum converges to one as the number of regions becomes large. An index value of for example 0.2 would imply that to equalise the industry mix in the region to that of the rest of the country, 20% of the workforce would need to be shifted from overrepresented to underrepresented industries. Suppose that the industries in which the region is specialised represent overall 40% of the workforce, against 20% in the rest of the country, which implies that the other industries represent 60% in the region and 80% in the rest of the country, then K=½*(0.4‑0.2) +½*(0.8‑0.6)=0.2. The employment-weighted average across regions provides an indication of the degree of regional specialisation of a country. To ensure that the index does not depend on the number of regions in a country and ensure its comparability across countries it is rescaled by R/(R‑1).

Measuring local exposure to structural change

A region’s exposure to country-wide structural change $Δ X_{r}^{}$ is captured using a shift-share approach which weighs national-industry level shifts by a region’s industry shares in employment in the base year:

$Δ X_{r}^{} = \sum_{j} {s_{r j τ} g}_{j}$

where $s_{r j τ}$ refers to the employment share in region r of industry j in base year $τ$ , and $g_{j}$ is a national industry-level shift in a variable capturing the aggregate component of structural change that is common across regions using 2‑ or 3‑digit industry-level information.

Trade shocks are measured based on country-wide changes in industry-level import penetration from China and Central and Eastern European Countries (“East”) between 2000 and 2007 ( $Δ {I P}_{j 2000 - 2007}^{H o m e \leftarrow E a s t}) .$ A region’s exposure to trade shocks is calculated using equation (2) with $g_{j} = Δ {I P}_{j 2000 - 2007}^{H o m e \leftarrow E a s t}$ . Import data for 2‑digit industries is obtained from the OECD Trade in value‑added (TiVA) database. Changes in import penetration are measured by dividing changes in import volumes by initial domestic industry absorption (domestic industry production plus imports minus exports) in 2000.

Technology shocks are measured based on global patent applications for digital technologies between 2012 and 2021 by 3‑digit industry ( $P a t D i g i j)$ following Prytkova and Petit (2025[9]) and Prytkova et al. (2024[10]). Patent applications linked to industries come from the TechXposure database. A region’s exposure to technology shocks is calculated using equation (2) with $g_{j} = P a t D i g i j$ . The digital technological exposure score of an industry reflects semantic similarities between patents and the descriptions of industries. It is also possible to estimate digital exposure separately for routine‑manual and non-routine occupations, as the TechXposure database includes exposure measures for both sectors and occupations.1

1. In practice, this is done by i) aggregating occupational exposure across occupational groups (routine‑manual and versus other) for each technology; ii) retaining only the top 25% technologies to which each occupational group is most exposed; iii) calculating average exposure across the top 25% technologies for a given industry and occupational group. Regional exposure is measured by weighting sector‑level exposure, on average across occupations or for a given occupational group, by regional sector employment shares.

3.1.2. Regional exposure to structural change

Differences in regional specialisation generate differences in the exposure of regions to country-wide structural change, including those driven by growing import competition from low-wage countries or digital innovation (see Box 3.1 for details).

Exposure to structural change varies strongly across regions

Regions are exposed very differently to import competition from low-wage countries and digital innovation due to the spatial concentration of economic activities (Figure 3.4). While the present analysis focusses on a region’s exposure to import competition during the 2000s and that to digital innovation in the 2010s, its exposure to structural change may have changed significantly since then, as trade patterns and the nature of technological change have evolved, with potentially important consequences for regional outcomes, as discussed in Box 3.2.

Box 3.2. The effects of trade realignment and artificial intelligence

Copy link to Box 3.2. The effects of trade realignment and artificial intelligence

Recent OECD work and broader research highlight how trade realignment and technological change, especially artificial intelligence (AI), are shaping labour market outcomes across regions and workers in OECD countries.

Trade realignment

The effects of trade with China and Central and Eastern European countries may be changing importantly as their role in the world economy evolves. There are signs that China is reducing its imports from advanced economies as it is becoming increasingly self-reliant and, at the same time, competing more directly with exports from advanced economies. The latter is most apparent for the euro area, notably in automotives (de Soyres et al., forthcoming[11]). Similarly, OECD (forthcoming[12]) documents that Central and Eastern European countries are increasingly integrated in the supply chains of EU firms, notably in car manufacturing.

New or recent increases in tariff barriers and heightened trade tensions are forecast to slow economic activity and weigh on labour markets (OECD, 2025[13]). Along with increased policy uncertainty, increased tariffs act as a drag on trade flows, investment and output growth and ultimately may also dampen employment growth. While labour markets have shown resilience in the short term, these effects can become more visible with time as firms shift from absorbing increases in costs to making longer‑term adjustments in production and investment, particularly in highly specialised regions dependent on international trade.

A recent multiregional input – output study simulating the increase in tariffs finds that it could lead to widespread employment losses, with cumulative global job losses exceeding 23 million in adverse scenarios, disproportionately affecting informal and low‑skilled workers (Ernst, Michelena and Bertin, 2026[14]). Historical evidence further supports these findings: IMF research using long‑panel country data from 1963 to 2014 shows that tariff hikes are associated with declines in domestic output and productivity, along with higher unemployment and rising inequality (Furceri et al., 2019[15]).

Artificial intelligence

The share of workers with jobs exposed to generative AI varies widely across places, with estimates ranging from about 16% in some regions to more than 70% in others, reflecting differences in industrial composition and occupational structures (Figure 3.3).¹ Regions with high concentration of employment in industries such as ICT, finance and education and occupations using cognitive and non‑routine tasks tend to have higher AI exposure. Regional differences in AI exposure could exacerbate existing urban – rural divides in incomes and labour market outcomes (see Chapter 2).

Evidence on the effects of AI on local labour markets is only starting to emerge (see the discussion in Chapter 1 of this report). In a widely publicised study, Brynjolfsson et al. (2025[16]) finds that while overall employment continues to grow, employment among younger, early-career workers in occupations most exposed to generative AI has declined relative to less‑exposed roles. Similarly, Hampole et al. (2025[17]) finds that AI adoption has not yet resulted in widespread job loss but instead is associated with productivity gains and employment growth in high‑exposure occupations. Gal et al. (2024[18]) estimate that AI could increase annual aggregate productivity growth by between 0.25‑0.6 p.p. over the next 10 years.

Figure 3.3. Regional exposure to generative AI
Copy link to Figure 3.3. Regional exposure to generative AI

Share of employment highly exposed to Gen-AI now or in the near future, latest available year

The impact of trade on regions is considered by focussing on sudden changes in import competition from low-wage countries, due to the admission of China to the World Trade Organisation (WTO) in 2001, or the expansion of the European Union to Central and Eastern European Countries (CEECs) in 2004‑2007.7 The local exposure of a region is measured by the rise in import penetration (i.e. the degree to which domestic demand is satisfied by imports rather than by locally produced goods) from both China and CEECs between 2000 and 2007 weighted by its industry shares in regional employment. To the extent that import competition leads to lower prices, profit margins, and employment in exposed industries, this is likely to disproportionately affect more exposed regions and particularly those where workers face barriers to switching industries or regions. Evidence from the “China shock” literature documents substantial job and income losses in more exposed local labour markets along with persistent scarring effects among incumbent workers (Autor, Dorn and Hanson, 2013[20]; Dorn and Levell, 2024[21]; Autor et al., 2025[22]). While all countries considered in this chapter have been exposed to a sharp rise in import competition during the early 2000s, it varied widely across regions within countries (Figure 3.4, Panel A).

The impact of technological change is examined by focussing on digital innovation during the 2010s. Digital technologies influence local labour markets through technology and scale effects, which depend on the way they complement or replace different groups of workers, the pattern of regional specialisation as well as other regional characteristics (e.g. intangible assets, universities). Digital technologies in the form of automation and software tend to reduce the demand for workers conducting routine tasks, while they complement workers specialisation in non-routine and interpersonal tasks (technology effect). Moreover, by raising productivity, they typically lead to lower prices and stronger product demand, with potentially positive effects for the overall demand for labour in an industry (scale effect). Regions with a high share of digitally complementary sectors, stronger skills bases, and greater intangible investment are more likely to benefit. Regional exposure to digital innovation is measured by global patent applications in 40 digital technologies from 2012 to 2021 linked to industries based on the semantic similarity of patent and industry descriptions weighted by regional employment shares in each industry at baseline in 2010 (Prytkova et al., 2024[10]; Prytkova and Petit, 2025[9]) (Panel B). Digital technology exposure is further broken down by the relevance of different digital technologies for different occupations (e.g. routine‑manual vs. non-routine).

The present analysis of structural change in regions complements ongoing OECD work that identifies places exposed to intensified industrial restructuring pressure and maps their adjustment patterns across local labour markets in six OECD countries over the period 1975 to 2023, highlighting distinct trajectories of labour market adaptation. More than half of places showed resilience or bounced back within a decade. Nevertheless, three‑quarters of all places experienced a decline in employment or productivity. Places with a stronger presence of higher education institutions, a broader service sector and younger populations were better equipped to absorb shocks and adapt to changing conditions (Ahrend et al., 2026[23]).

Figure 3.4. There are large differences in regional exposure to structural change
Copy link to Figure 3.4. There are large differences in regional exposure to structural change

Regional exposure to import competition from low-wage countries and digital innovation by country, minimum, maximum and median z-scores

3.1.3. Regional exposure across workers and firms

Regions do not only differ in their exposure to trade and technology shocks, but also in the composition of workers and firms. Geographical differences in exposure to trade and technology shocks, therefore, also translate into differences in exposure across workers and firms (Figure 3.5). Differences in worker composition in part reflect the fact that some groups of workers, e.g. the younger or more educated workers, are more mobile than others. Differences in firm composition are to a large extent driven by initial location choice as firms’ relocations are rare (OECD, 2023[24]).

Manufacturing workers are concentrated in regions that are more exposed to trade shocks (as may be expected since trade shocks only exposed manufacturing workers by construction), while non-manufacturing workers are concentrated in regions that are more exposed to digital innovation. In 2000, the average manufacturing worker lived in a region with a trade exposure of +0.3 standard deviations above the average, compared to ‑0.1 standard deviations below the average for the average non-manufacturing worker (Figure 3.5, Panel A). Similarly, in 2011, the average manufacturing worker lived in a region with an exposure to technology shocks of ‑0.1 standard deviations below the average, while this was it was just above the average for the average non-manufacturing worker (Figure 3.5, Panel C).

College‑educated workers are concentrated in regions that are more exposed to technology shocks and less exposed to trade shocks in both manufacturing and non-manufacturing (Figure 3.5, Panel A and C). The role of education is likely to reflect in part the tendency of more highly educated workers to move to urban areas, which tend to be less exposed to trade shocks. Similarly, foreign-born workers are concentrated in regions more exposed to technology shocks in both manufacturing and non-manufacturing and less exposed to trade shocks in non-manufacturing. The importance of migrant status reflects the fact that migrants tend to settle in urban centres which are more exposed to technology shocks and less to trade shocks. Differences by age and sex are negligeable as regional differences in the age and sex structure of employment tend to be small.

Manufacturing workers in firms and industries with more generous wage‑setting practices (measured by wage premia, i.e. average wages conditional on workforce composition) as well as large firms are concentrated in regions more exposed to trade shocks, while the opposite is observed for non-manufacturing workers (Figure 3.5, Panel B). Manufacturing and non-manufacturing workers in large firms, high-wage firms and high-wage industries are concentrated in regions more exposed to technology shocks than their counterparts in smaller firms and lower wage firms and industries. Differences in exposure by firm type tend to be economically significant, suggesting that high-wage and high-productivity firms tend to locate in different locations from their low-wage, low-productivity counterparts. Moreover, the fact that they are somewhat more pronounced that those across groups of workers indicates that regional disparities primarily reflect the composition of firms rather than that of workers.

Figure 3.5. Exposure to structural change varies across firms and workers
Copy link to Figure 3.5. Exposure to structural change varies across firms and workers

Regional exposure to import competition from low wage countries and digital innovation by broad sector and worker/firm group in the base year, z-scores (standard deviations from the average)

3.2. The impact of structural change on places, people and firms

Copy link to 3.2. The impact of structural change on places, people and firms

The causal effects of structural change in local labour markets are analysed following the framework that was introduced by Dustmann et al. (2025[1]) and implemented by Autor et al. (2025[22]) for studying the local impact of the China shock across US labour markets (see Box 3.3 for details). The results are pooled across European Union countries plus Norway to allow including smaller countries with a limited number of regions in the analysis while doing statistical inference.8 The estimates for the United States are taken from Autor et al. (2025[22]) who use the same empirical model with slightly more disaggregated industry and regional data. Regions either refer to small regions (TL3) or commuting zones.

Box 3.3. Methodology II: The effects of trade and technology shocks on places, people and firms

Copy link to Box 3.3. Methodology II: The effects of trade and technology shocks on places, people and firms

The empirical approach for analysing the impact of structural change on local labour markets is based on Dustmann et al. (2025[1]) and Autor et al. (2025[2]). The effects of structural change are analysed separately for local employment, the flows of workers across regions, sectors and employment statuses as well as different types of firms. Formally, it is based on the decomposition of employment changes in each region r between year t and a base year $τ$ as a share of population in the base year (“places”) into net worker flows (“people”) and job reallocation between firms (“firms”):

$\begin{matrix} \underset{p l a c e s}{\underset{⏟}{\frac{E_{r t} - E_{r τ}}{p o p_{r τ}}}} = \underset{p e o p l e}{\underset{⏟}{\sum_{c} \frac{{i n f l o w}_{r, t - τ}^{c} - {o u t f l o w}_{r, t - τ}^{c}}{p o p_{r τ}}}} = \underset{f i r m s}{\underset{⏟}{\sum_{f} \frac{E_{r t}^{f} - E_{r τ}^{f}}{p o p_{r τ}}}} # (2) \end{matrix}$

where $E_{r t}$ is regional dependent employment and $p o p_{r τ}$ the initial population in the region. The change in regional employment between a given year t and the initial year $τ$ on the left-hand side is decomposed into the sum of net worker flows between different adjustment channels indexed by c. Workers may change sectors without changing regions (inter-sectoral mobility channel), change regions while remaining employed (geographic mobility channel),1 transition in and out of working-age employment (employment mobility channel), and enter or leave the working-age population (age‑related mobility channel channel).2 The place effect can also be decomposed into net employment changes across different firm types indexed by f. Firms are differentiated according to the wage‑setting practices of their industry, their firm wage practices relative to their industry average and their size. The decomposition is implemented using total regional employment as well as employment by broad sector (manufacturing, non-manufacturing).

To quantify the impact of structural change on places, workers and firms, long-difference models are separately estimated for each year t:

$\begin{matrix} Δ Y_{r t} = α + β_{t} Δ X_{r} + δ_{t} W_{r τ} + ϵ_{r t} # (3) \end{matrix}$

where $Δ Y_{r t}$ is the regional net employment change since the base year, or alternatively its worker or firm components, $Δ X_{r}$ refers to the local exposure to trade‑induced or technology-induced structural change. The coefficient of interest $β_{t}$ captures the effect of local exposure to structural change on places, workers or firms. $W_{r τ}$ is a set of region-level control variables measured in the base year that account for pre‑existing differences between regions such as the initial manufacturing share (trade) or the share of exposed industries and the regional employment-to-population changes between 2004 and 2012 (technology). Finally, $ϵ_{r t}$ is the error term clustered at the region level. All regressions are weighted by total regional employment in the base year. When pooling regions from different countries (as when estimating results for Europe), we additionally control for country fixed effects to account for between-country differences in time‑invariant unobserved heterogeneity. As the regressions are estimated by year, time fixed effects are absorbed in the constant.

1. For the present purposes, being out of work is defined as being out of dependent employment in the country and may include being self-employed or employed in another country.

2. The panel dimension of the data allows defining inflow and outflow concepts based on an individual’s age, region, employment status and sector in each year $t$ compared to the initial year $τ$ . Workers who are below the age of 18 in year $τ$ and above the age of 64 in year $t$ are always counted in the ageing channel.

3.2.1. The impact on trade and technology shocks on places

The estimated effects of trade on total regional employment and its contributions by manufacturing and non-manufacturing are presented in Figure 3.6 for Western Europe (9) (Austria, Denmark, France, Germany, Italy, Norway, Portugal, Spain and Sweden),9 Canada, Germany and the United States. The effects of technology on total regional employment and its contributions by non-routine and routine‑manual occupations are presented in Figure 3.7 for Europe (12) (Austria, Denmark, Finland, France, Germany, Hungary, Italy, the Netherlands, Norway, Portugal, Spain and Sweden), Canada Germany, and the United States. Regional employment is measured as the change in working-age employment over the period divided by the working-age population in the base year (to be consistent with the worker-level decomposition in Section 3.2.2).10 Unlike studies that focus only on firms or specific industries, the regional approach employed here captures how changes spread within an entire area through spillover and other general equilbrium effects. These include amongst others higher demand from manufacturing companies for local services (“multiplier effect”), links between firms in the local supply chain (“input-output linkages”), and workers moving between local firms and industries (“reallocation”). The overall net effect is the combined result of these different forces in addition to its direct effects on firms and industries.11

The effects of trade shocks on places

Regional exposure to import competition from low-wage countries is associated with a persistent decline in the share of manufacturing in employment, but with important differences in its size between Western Europe (9) and the United States. A one standard-deviation increase in regional trade exposure, which corresponds to moving from a region with average exposure to one in the 80th‑90th percentile, is associated with an average decline in manufacturing employment by 2018 of 0.4 p.p. in Canada, 0.5 p.p. in Western Europe (9), 1 p.p. in Germany and 1.4 p.p. in the United States. To put these effects in perspective, manufacturing employed 11% of working-age adults in the United States in 2000, 10% in Canada in 2001 and 5% in Western Europe in 2002. The documented declines per standard-deviation of trade exposure thus amount to the loss of about one in seven manufacturing jobs in the United States, one in ten in Western Europe and one in twenty-three in Canada.12

Growth in non-manufacturing employment eventually offsets job losses in manufacturing in most countries but with important differences in the timing of these effects. Cumulative job gains in non-manufacturing more than compensate any job losses in manufacturing from the start in Germany, but only from 2013 in the United States and 2015 in Western Europe (9). In Canada, where the adverse employment effect in manufacturing is small, there is no sign of offsetting job gains in non-manufacturing.13 A one standard-deviation increase in regional trade exposure is associated with an increase in non-manufacturing employment of 0.7 p.p. by 2018 in Western Europe (mainly in producer services, see Box 3.4), 2.1 p.p. in Germany and 2.7 p.p. in the United States.14 Although the empirical model does not establish a direct link between the rise in import competition and the growth in non-manufacturing employment,15 this may in part be due to the relative price effects from import competition that make tradeable services more competitive, potentially supported by lower wage pressures due to the increase in effective labour supply following the contraction in manufacturing employment. The growth of non-manufacturing employment may also result from supportive policies as discussed in Section 3.3. On net, regional trade exposure is associated with an increase in total employment to the initial population by 2018 of 0.2 p.p. in Western Europe, 1.1 p.p. for Germany and 1.4 p.p. in the United States. Canada is an exception in that local employment remains depressed.

While for the United States the effects of low-wage competition are largely driven by the rise of China in world trade since 2001, for Western Europe, eastern enlargement of the EU between 2004 and 2007 is also important and its effect is somewhat larger than that of increased import competition from China (Annex Figure 3.B.1). For the United Kingdom, both sources of import competition matter for the decline in manufacturing employment but import competition from China is the main force (Foliano and Riley, 2017[27]). In Germany, imports from China mainly replaced imports from other low-wage countries and hence did not have much of an impact on manufacturing employment (Dauth, Findeisen and Suedekum, 2014[28]). Similarly, in Japan, cheaper imports of intermediate inputs supported firm productivity and led to increased job creation (Taniguchi, 2019[29]). By contrast, in the United States and Canada, imports from China competed more directly with domestic firms and workers (Autor, Dorn and Hanson, 2013[20]; Murray, 2017[30]). In recent years, the effects of trade may be changing as exports from China are becoming more similar to those from advanced economies, notably in the euro area, and Central and Eastern European countries are increasingly integrated in the supply chain in the EU (Box 3.2).16

It is important to note that the local labour market effects of trade with China and Central and Eastern Europe are not only driven by increased import competition in low-wage sectors but are also be shaped by new opportunities for exports. In Germany, increased imports from low-wage countries went together with increased exports. Indeed, taking account of the effects of increased exports suggests that overall international trade led to the creation of more manufacturing jobs than it destroyed in Germany (Dauth, Findeisen and Suedekum, 2014[28]). This partly reflects the importance of specialised machinery and equipment in German manufacturing, sectors that were both less directly exposed to Chinese competition and well positioned to benefit from rising demand by Chinese firms. German exporters may also have benefited from a prolonged period of wage moderation, supporting competitiveness. By contrast, the local labour market effects were more negative in countries such as the United States and the United Kingdom where manufacturing firms were more directly exposed to import competition from China and the growth in imports was not matched with a similar rise in exports (Dorn and Levell, 2024[21]).

Figure 3.6. The effects of trade exposure on regional employment
Copy link to Figure 3.6. The effects of trade exposure on regional employment

p.p. effect of a one standard-deviation increase in regional exposure to import competition from low-wage countries on regional employment divided by the initial working-age population by broad sector

The effects of technology shocks on places

Regional exposure to emerging digital technologies is associated with a persistent increase in total employment. A one standard-deviation increase in a region’s technology exposure is associated with an average increase in total employment between 2012 and 2017 of 0.4 percentages points in Europe (12), Germany and the United States and 2.0 p.p. in Canada.17 These effects are largely driven by producer services (see Box 3.4) and industries intensive in non-routine occupations. Regions exposed to innovation in digital technologies that are most relevant for industries with a strong focus on routine‑manual occupations tend to see a decline in their employment share, except in Germany. In other words, exposure to digital innovation mainly augments labour, by boosting demand for workers in complementary tasks, but is labour-saving in routine‑manual occupations, where tasks are more easily automated. These findings are broadly similar to those in Prytkova et al. (2024[10]) for Europe and Prytkova and Petit (2025[9]) for the United States. One possible explanation for the divergent findings for Germany is that routine manual workers may have been better able to adapt to technological change by transitioning into different roles within firms, supported by vocational training systems, internal labour markets and social dialogue institutions (Battisti, Dustmann and Schönberg, 2023[31]; Dauth et al., 2021[32]).

The present analysis of the effects of digital innovation complements the empirical literature that examines how the diffusion of industrial robots affects local labour market outcomes. For the United States, Acemoglu and Restrepo (2020[33]) show that commuting zones that are more exposed to robot adoption between 1990 and 2007 experienced declines in both total employment and wages, with job losses concentrated in manufacturing and only limited job gains in services. Their framework highlights how robots substitute for routine production tasks while productivity gains may not fully translate into local job creation when adjustment frictions and demand leakages are present. For Germany, Dauth et al. (2021[32]) find that robot exposure reduced manufacturing employment in more exposed regions but did not lower total employment because service‑sector job growth offset losses in manufacturing. The burden of adjustment fell disproportionately on labour market entrants, while many incumbent workers transitioned to different roles within firms. The analysis does not capture recent developments in AI which may be particularly relevant for regions specialised in ICT and finance and may differ in its labour-saving and augmenting effects (Box 3.2).

Figure 3.7. The effects of technology exposure on regional employment
Copy link to Figure 3.7. The effects of technology exposure on regional employment

p.p. effect of a one standard-deviation increase in regional exposure to digital innovation on regional employment divided by the initial working-age population by technology type

Box 3.4. The effects of trade and technology shocks across sectors

Copy link to Box 3.4. The effects of trade and technology shocks across sectors

Trade and technology-driven structural change tend to be associated with job creation in non-manufacturing. This box takes a closer look at where these new jobs are emerging (Figure 3.8). It shows that job creation in non-manufacturing is concentrated in public services e.g. health and education but also producer services, e.g. ICT, financial and legal services. By contrast, consumer services show little to no significant response to trade or technology shocks. This pattern suggests that the jobs created in the wake of trade and technological change tend to be relatively skill intensive. Section 3.2.3 examines whether these jobs are also concentrated in firms and industries characterised by more generous wage‑setting practices.

Figure 3.8. The effects of exposure to trade and technology shocks on regional employment by sector
Copy link to Figure 3.8. The effects of exposure to trade and technology shocks on regional employment by sector

p.p. effect of a one standard-deviation increase in regional exposure to import competition from low-wage countries and digital innovation and regional employment by sector in (Western) Europe

3.2.2. The impact on people

Trade and technology-driven structural changes have important effects on regional employment and its composition across industries (Box 3.4). This sub-section examines how this transformation takes place by focussing on worker flows between i) working-age employment and non-working-population (“age-related mobility channel”), ii) working-age dependent employment and working-age dependent non-employment (“employment mobility channel”),18 iii) industries within regions (“inter-industry mobility channel”),19 iv) employment in different regions (“geographical mobility channel”). The analysis of these flows exploits the longitudinal dimension of the underlying worker-level data, which allows tracking transitions in and out of national employment and across firms and regions of individual workers since the reference year.

The effects of trade shocks on people

Changes in regional employment by broad sector in response to low wage import competition are largely brought about by flows in and out of work. This is reflected in the combined contributions of the employment and ageing channels in Figure 3.7. In all countries considered, net flows in and out of work account for the bulk of the change in regional employment by broad sector due to regional trade exposure.20 Importantly, this implies that workers that move out of manufacturing and move into non-manufacturing are by a large extent not the same individuals. However, the relative importance of the age‑related and employment mobility channels differs across countries. In the United States, age‑related mobility is the key channel for accommodating trade shocks, as older workers retire from manufacturing jobs and younger workers enter non-manufacturing jobs in more trade‑exposed regions, while the role of employment mobility is small. Age‑related mobility accounts for 65% of the decline in manufacturing employment between 2000 and 2019. By contrast, in Western Europe, Germany and Canada, both age‑related and employment mobility tend to be important. The importance of employment mobility in Western Europe for job gains in non-manufacturing reflects to an important extent the role of migration from Central and Eastern European countries to Western Europe following the removal of restrictions on labour mobility from the early 2010s onwards (see Annex Figure 3.B.2 for France and Germany).21^,22

Mobility between regions and industries plays only a minor role in adjusting to low-wage import competition. Net sectoral outflows account for 12% of the decline in manufacturing employment between 2002 and 2018 in Western Europe and 14% between 2000 and 2018 in the United States, while net inflows account for even less of the increase non-manufacturing employment. Consequently, direct worker flows between industries account for rather little of the reallocation of employment between sectors.23 There is evidence for the United States that manufacturing workers who earned higher wages before the China shock are more likely to move to jobs outside manufacturing and suffer smaller income losses. By contrast, lower-paid workers are less likely to change industry and suffer larger earnings losses as they remain exposed to the effects of import competition (Autor et al., 2014[34]). Net geographical mobility within countries accounts for a similarly small amount of the decline in employment in manufacturing (13% in Western Europe and 15% in the United States. Strikingly, in the United States, manufacturing employment declines in trade‑exposed regions as a result of fewer migration inflows from other regions rather than more migration outflows to other regions as often suggested (Autor et al., 2025[22]). Net internal migration tends to dampen the expansion of employment in non-manufacturing in Western Europe and the United States but supports it in Germany.24

Overall, the analysis shows that place‑based recovery and individual adjustment follow distinct trajectories. While total employment in local labour markets tends to recover from trade shocks after several years, manufacturing workers experience far more persistent effects. Adjustment is largely generational: older workers exit manufacturing, often permanently, while younger workers transition into jobs in non-manufacturing. This process can entail significant social costs. Evidence from the United States indicates that only a small share of lost earnings is offset by social transfers (Autor, Dorn and Hanson, 2013[20]) and that the consequences of trade shocks go well beyond the loss of income, potentially contributing to reduced marriage and cohabitation rates, as result of an increase in NEET rates among young men (Autor, Dorn and Hanson, 2019[35]), and increases in crime and mortality, associated with drug and alcohol abuse (Feler and Senses, 2017[36]; Autor, Dorn and Hanson, 2019[35]).

Figure 3.9. The effects of regional trade exposure on net worker flows
Copy link to Figure 3.9. The effects of regional trade exposure on net worker flows

p.p. effect of a one standard-deviation increase in regional exposure to import competition from low-wage countries on net worker flows by broad sector

The effects of technology shocks on people

As in the case of trade shocks, regional labour markets mainly adjust to digital innovation through transitions in and out of national employment, while transitions between regions play a minor role (Figure 3.10).25 The increase in employment between 2012 and 2017 due to regional exposure to digital technologies is overwhelmingly driven by net inflows from non-employment, particularly from education. Similarly, the decline in employment in routine‑manual occupations is entirely driven by net outflows to non-employment, and as in the case of trade shocks, resulting in persistent joblessness, with potentially important social and fiscal consequences. Empirical evidence further suggests that regions exposed to automation not only experience declines in routine manual employment but also increases in precarious employment, earnings in stability and labour force attachment (Acemoglu and Restrepo, 2020[33]; Graetz and Michaels, 2015[37]). In service‑sector contexts, algorithmic management and job restructuring can heighten job strain and reduce worker autonomy, potentially affecting well-being even when employment is retained (Rosenblat and Stark, 2015[38]; De Stefano, 2015[39]).

Figure 3.10. The effects of regional technology exposure on net worker flows
Copy link to Figure 3.10. The effects of regional technology exposure on net worker flows

p.p. effect of a one standard-deviation increase in regional exposure to digital innovation on net worker flows by type of occupation

3.2.3. The impact on firms

The impact of trade and technology shocks on local labour markets can also be analysed by examining how different groups of firms adjust. Three different firm splits are considered: i) across industries with wage premia above and below the median; ii) across firms with premia above and below the industry median; iii) across firms with more and less than 100 employees.

The effects of trade shocks on firms

Structural change differentially affects industries that differ in their wage‑setting practices (Figure 3.11).26 Regional exposure to import competition from low-wage countries reduces employment in manufacturing industries with more generous wage‑setting practices (conditional on worker composition), while the corresponding increase in employment in non-manufacturing also tends to be concentrated in high-wage industries but not quite to the same extent. (Panel A). This confirms the anecdotal evidence that the decline in manufacturing employment is associated with the destruction of quality jobs. The creation of new jobs in non-manufacturing, and in particular business services, is typically associated with more skilled employment but not always more generous wage setting-practices conditional on skills (Criscuolo et al., 2023[40]). In part, this is likely to reflect the lesser importance of collective bargaining beyond manufacturing.

Structural change also differentially firms that differ in their wage‑setting practices within industries (Panel B). When industry employment contracts, this tends to be concentrated in firms with wage premia below the industry average. These are typically firms that have below-average productivity and are relatively small (Panel C). By contrast, when industry employment expands, this tends to be concentrated in firms with wage premia above the average in their industry, which are more likely to be high productivity and larger firms.27 Consequently, employment shifts in both contracting and expanding industries from low-pay low-productivity firms towards higher paying and more productive firms. This process of job reallocation supports industry-level wage and productivity growth. These findings resonate well with those of previous firm-level studies that focus on the China shock. Using firm-level data for 12 European countries over the period 1996‑2007, Bloom et al. (2015[41]) show that import competition from China increases productivity growth within firms and job reallocation towards more productive firms, together accounting for approximately 14% of productivity growth in Europe between 2000 and 2007.28 Similarly, using firm-level data for France, Aghion et al. (2024[42]) show that employment contractions in response to import competition from China were concentrated in low productivity firms.

Figure 3.11. The effects of regional trade exposure across firms and industries
Copy link to Figure 3.11. The effects of regional trade exposure across firms and industries

p.p. effect of a one standard-deviation increase in regional exposure to import competition from low-wage countries on the composition of firms and industries, Western Europe (8)

The effects of technology shocks on firms

Regional exposure to digital innovation is associated with some important differences across firms and industries (Figure 3.12). Job creation in non-routine occupations and job destruction in routine‑manual occupations is evenly split across high and low industries. Within industries, job creation and job destruction does not show important differences across firms according to their wage‑setting practices. However, job destruction in routine‑manual occupations tends to be concentrated in small and medium-sized firms, while job creation in non-routine occupations is split evenly between small and medium enterprises (SMEs) and larger firms. This pattern is consistent with evidence that smaller firms often face greater adjustment costs when adopting new technologies, making workforce restructuring a more important margin of adaptation, while larger firms may be better positioned to absorb technological change through internal reallocation and the creation of new non-routine jobs.

Figure 3.12. The effects of regional technology exposure across firms and industries
Copy link to Figure 3.12. The effects of regional technology exposure across firms and industries

p.p. effect of a one standard-deviation increase in regional exposure to digital innovation on the composition of firms and industries, Europe (11)

3.3. Implications for place‑based and national policies

Copy link to 3.3. Implications for place‑based and national policies

The persistent and place‑specific effects of trade and technology shocks in many OECD countries highlight the importance of policies that help limiting local vulnerabilities and anticipating structural change, while alleviating its social costs by supporting transitions sectors and regions.

While labour mobility between regions and industries is often viewed as an important adjustment lever, the empirical evidence in this chapter shows that mobility responses play a modest role in attenuating the impacts of localised shocks.29 Most OECD countries exhibit low internal migration rates, often below 3% annually, indicating that many workers, particularly older or lower-skilled ones, do not relocate when local job prospects decline (Causa and Pichelmann, 2020[43]). Job mobility between industries also tends to be relatively low (Elliot, 2003; Fluchtmann et al., 2026). Low mobility may reflect a variety of barriers related to lack of opportunities, the cost of moving across firms, occupations or regions and the personal circumstances of workers (OECD, 2025[44]). These include social ties, housing constraints and a desire to remain embedded in local communities – a notion often referred to as the “right to stay”. Importantly, the less mobile workers are across regions and industries – that is, the steeper are local labour supply curves – the more adjustment to labour demand shocks occurs through changes in local wages and employment rates. Local labour markets exposed to adverse shocks therefore tend to experience persistently depressed earnings and employment rates, often lasting a decade or longer (Amior and Manning, 2018[45]; Rice and Venables, 2021[7]); Autor et al., 2022). Given the persistent effects of structural change and the resulting regional disparities, addressing them requires place‑based policies supported by national policies (Venables, 2024[46]; OECD, 2025[47]).

Place‑based policies are spatially targeted supply-side policies that are sensitive to local conditions and opportunities. They are designed to foster long-term local economic development and well-being, typically with the support of a higher-level government entity (OECD, 2025[47]). They usually aim at addressing market failures across different policy domains related to: (i) the local provision of public goods (“regional policies”); (ii) externalities of economic activities in particular locations (“place-based industrial policies”); and (iii) locational frictions to employment and labour mobility (“place-based employment and social policies”).30 While place‑based policies often have an economic efficiency rationale, they can also be undertaken for non-economic reasons, including political, social, environmental or security objectives. The challenge for policymakers is to integrate these strands into coherent strategies, co‑ordinated across different levels of government, that effectively respond to shocks and support people through transitions (Figure 3.1). Spatial targeting refers both to the allocation of resources – often requiring support from a higher-level government – and their place‑sensitive implementation, which relies on locally embedded institutions.31 The involvement of local and higher-level authorities underscores the importance of effective co‑ordination across levels of government.

Place‑based policies also face specific implementation and evaluation challenges. Multiple levels of government often share responsibilities, which can create co‑ordination gaps and dilute accountability. Spatial targeting may also lead to spatial substitution effects, where economic activity relocates from neighbouring areas rather than generating net gains. Relatedly, spatial sorting may occur when mobile firms or workers move to benefit from targeted support, complicating the attribution of policy impacts. Limited administrative and technical capacity at the local level can further constrain effective implementation. These challenges highlight the need for robust evaluation frameworks tailored to spatially targeted interventions (OECD, 2017[48]). Embedding evaluation into programme design and adjusting or discontinuing measures that do not demonstrate effectiveness, can improve policy learning and ensure the cost-efficient use of public resources.

Figure 3.13. A framework for place‑based policies
Copy link to Figure 3.13. A framework for place‑based policies

Box 3.5. Examples of place‑based policies

Copy link to Box 3.5. Examples of place‑based policies

Place‑based policies have experienced a renewed interest in recent years. Although place‑based policies are not new and have been an important part of government policy for many years, persistent and rising spatial economic disparities, together with the significant spatial impacts of structural change, have increased governments’ focus on their potential (OECD, 2025[49]). In recent years, many OECD countries have adopted or scaled up place‑based policies.

The EU Cohesion policy plays a crucial source of funding for place‑based policies in EU countries (OECD, 2025[47]). There are eight EU cohesion funds which are managed jointly by the European Commission and national and regional authorities in Member States, among which the European Regional Development Fund (ERDF), the European Social Fund Plus (ESF+) and the Cohesion Fund which are specifically designed to strengthen economic, social and territorial cohesion and correct imbalances between regions. Funds are primarily allocated based on the region’s level of development. While 70% of the ERDF and ESF+ are concentrated on the EU’s less developed regions, the Cohesion Fund is entirely allocated to low-income Member States (<90% of the EU average).
The United States has a long tradition of place‑based interventions, even in the absence of a formal regional policy framework. Federal programmes such as the Appalachian Regional Commission and the Delta Regional Authority target structurally disadvantaged regions through infrastructure, entrepreneurship and workforce development initiatives. In addition, the Economic Development Administration supports regional innovation strategies and cluster development in response to local economic shocks, including trade‑related adjustment.
In Canada, place‑based policies support AI development and diffusion by concentrating investments in regions with existing research, talent, and industrial ecosystems. The Pan-Canadian Artificial Intelligence Strategy exemplifies this approach, anchoring world-class AI research hubs in Montréal, Toronto – Waterloo, Edmonton, and Vancouver, while regional development agencies and provincial governments support local AI adoption by firms through innovation vouchers, applied research partnerships, and cluster initiatives (such as SCALE.AI in Québec – Ontario).
Japan has implemented comprehensive regional revitalisation policies to address population decline and economic concentration in Tokyo and surrounding areas. The “Comprehensive Strategy for Regional Revitalisation” promotes a stronger economy to improve well-being and make rural areas more attractive for young people and women. In addition, a “Strategy for the Future of Regions” is under consideration, aimed at promoting a growth-oriented regional development model centred on strategic industrial clusters.
Korea’s balanced national development strategy includes the designation of innovation cities and enterprise cities outside the Seoul metropolitan area. Public institutions have been relocated to these areas to stimulate local economies. In parallel, regional innovation systems are supported through targeted R&D investment and collaboration between universities and local industry. The objective is to reduce spatial concentration and foster regionally embedded technological upgrading.
Australia’s current approach to regional development is delivered through a national network of committees and a suite of regional policies co‑ordinated by the national government. These initiatives emphasise place‑based investment. In particular, the Regional Investment Framework highlights the importance of tailoring interventions to local contexts, valuing community input and priorities, and addressing diverse regional opportunities and challenges through co‑ordinated, whole‑of-government investment.

Source: OECD (2025[49]), “Place-based industrial policy: Lessons for place transformation”, https://doi.org/10.1787/43edc0df-en.

3.3.1. Regional policies

Regional policies address market failures in the local provision of public goods by ensuring that regions have the capacity, resources, and co‑ordination mechanisms needed to deliver essential services and infrastructure effectively (e.g. education, healthcare, and transport). These market failures arise when local authorities lack sufficient fiscal resources, when investment costs are high relative to local budgets, or when the benefits of public goods – such as transport networks, education and skills systems, digital infrastructure, health services, and environmental protection – extend beyond local boundaries. In practice, regional policies build on a region’s existing strengths for economic development and address specific weaknesses that form bottlenecks. Public investments are therefore primarily driven by a development objective. However, by ensuring minimum levels of public services they can also strengthen social cohesion and reduce the perception of being “left behind” (see Chapter 2).

Structural change fundamentally reshapes the spatial distribution of economic activity, population, and fiscal capacity within countries, with significant implications for the provision of public goods. Rapidly growing regions – often driven by high-tech and knowledge‑intensive services – benefit from expanding tax bases and innovation in service delivery, but also face acute pressures on housing, transport, education, and health systems, as well as risks of social and spatial exclusion. At the same time, deindustrialising regions experience job losses, population decline, and eroding local revenues, even as demand for public support, healthcare, and retraining increases. These contrasting trajectories strain -sub-national systems of public goods provision and can deepen territorial inequalities.

Designing effective regional policy responses is challenging. Fiscal transfers and place‑based investments must balance efficiency with equity, while avoiding long-term dependency or politically driven misallocation. Co‑ordination across levels of government is often weak, administrative capacity varies sharply across regions, and policies typically follow rather than anticipate economic change. Addressing these challenges requires forward-looking governance, sustained investment in human capital, and institutions capable of redistributing resources and opportunities across regions, ensuring that public goods continue to support social cohesion and inclusive development in the context of ongoing structural transformation. Evidence from the Tennessee Valley Authority investment program (Kline and Moretti, 2013[50]) and the EU Structural Funds (Becker, Egger and von Ehrlich, 2012[51]; Becker, Egger and von Ehrlich, 2010[52]) suggests that infrastructure investment can be cost-effective in delivering productivity growth in targeted regions and can act as a redistributive tool across locations, but also that questions remain about how long-lasting these effects are and their underlying mechanisms (Neumark and Simpson, 2015[53]).

The empirical evidence also suggests that investments in education make regions better equipped to adjust to negative shocks. Cities with a higher initial share of college‑educated workers were significantly more likely to recover from large deindustrialisation shocks, experiencing faster employment growth in the decades following manufacturing decline (Gagliardi, Moretti and Serafinelli, 2023[54]; Ahrend et al., 2026[55]). These results underscore that education is not only a long-term growth driver but also a key factor that enhances regional resilience, accelerates post‑shock recovery and reduces the risk of becoming locked into structural decline.

The provision of local public services can also help to alleviate structural barriers to geographical mobility and consequently support structural change and inclusiveness (Box 3.6). These may be combined with employment and social policies that address individual barriers to geographical mobility, by improving access to information on job vacancies and relocation, providing financial support towards relocation expenses, facilitating access to local services (e.g. housing) in the destination location, and enhancing the portability of benefits and pensions.

Box 3.6. Policies to support geographical mobility

Copy link to Box 3.6. Policies to support geographical mobility

Geographical mobility can support structural adjustment and reduce regional disparities. When workers move from low-wage, high-unemployment regions to high-wage, low-unemployment regions, labour supply adjusts and regional gaps narrow. However, internal migration rates remain low in most OECD countries, often below 3% annually (Causa and Pichelmann, 2020[43]) – see also Chapter 2. This suggests that multiple barriers limit mobility. These barriers can be grouped into individual barriers to mobility and structural barriers to mobility.

Policies addressing individual barriers to geographical mobility

Individual barriers relate to social ties, access to services, financial constraints, and information gaps about job opportunities (see Chapter 2). Social ties include the personal and professional networks that could be highly valuable to workers in a variety of activity, such as childcare, leisure, and job search. Furthermore, many workers lack savings to cover upfront relocation costs, such as transport, housing deposits or temporary double housing. Financial support measures, including relocation grants and mobility allowances, can reduce these constraints. Several OECD countries, including Australia, Austria, Germany, New Zealand and Romania, provide such support to unemployed workers who accept jobs in other regions. For example, Austria provides commuting and housing allowances such as up to EUR 260 per month for travel and EUR 400 per month for secondary housing. Limited knowledge about job opportunities in other regions reduces effective job search. Public employment services can improve matching by providing vacancy information, relocation counselling and job-search assistance across regions.

Policies addressing structural barriers to geographical mobility

Inter-regional mobility rates differ significantly across countries, reflecting structural barriers that arise from housing markets, local service provision and institutional settings (Causa, Abendschein and Cavalleri, 2021[56]; Cavalleri, Luu and Causa, 2021[57]). Housing supply constraints, including restrictive land-use and planning regulations, can raise prices in high-demand regions and deter internal migration. Reforms that increase housing supply responsiveness, including social housing provision, can lower these barriers. Limited availability of childcare and other local services can further constrain mobility, especially for dual-earner households. Differences in regional policies can also reduce portability of benefits and create regulatory obstacles, such as non-recognition of qualifications. Enhancing portability of social benefits and priority rankings in accessing social services (e.g. social housing) and expanding mutual recognition of occupational licences can reduce these institutional barriers.

Integrated policy approach

An effective mobility strategy combines measures that address both individual and structural barriers, targeting especially the most vulnerable (see Chapter 2). Financial and informational support can enable workers to respond to opportunities. Housing, childcare and institutional reforms can ensure that people are willing to relocate and that destination regions can absorb migrants from other regions. Addressing only one set of barriers is unlikely to generate large mobility responses. A comprehensive approach can strengthen adjustment to structural shocks while supporting inclusive regional development.

Source: OECD (2025[44]), Addressing Regional Labour Market Imbalances in Austria, https://doi.org/10.1787/0cf6186b-en; OECD (2026[58]), Reviving Productivity Growth in Canada, https://doi.org/10.1787/773dcd00-en.

3.3.2. Place‑based industrial policies

Industrial policies are typically justified based on the presence of externalities which drive a wedge between private returns and social benefits (Millot and Rawdanowicz, 2024[59]). These may be negative as in the case of carbon pricing policies to reduce environmental harm or positive as in the case of policies that seek to promote innovation through private‑public partnerships. The case for industrial policies does not depend only on the presence of externalities but also the benefits of policy interventions relative to their costs. When done well industrial policies can help aligning market outcomes with broader economic, social and environmental objectives.

Traditionally, industrial policy focussed on supporting a limited number of strategic firms or “national champions”, often in capital-intensive or high-technology sectors. Interventions typically took the form of direct subsidies, state ownership, or protection from competition, with the objective of achieving economies of scale, supporting international competitiveness, and boosting technological leadership at the national level. While this approach contributed to the development of key industries in some contexts, it often overlooked broader productivity dynamics, created risks of market distortion, and paid limited attention to territorial impacts.

More recently, industrial policy has evolved towards a broader, more forward-looking approach e.g. (Criscuolo et al., 2022[60]); OECD (2025[61]). Rather than targeting individual firms, the new generation of industrial policies focusses on strengthening entire sectors or value chains and enabling framework conditions such as skills development, innovation systems, access to finance, digital infrastructure, and regulatory environments. This approach places greater emphasis on anticipating global challenges, such as the renewable energy and digital transitions, and enhancing the adaptability of local economies to structural change. As a result, industrial policy is increasingly implemented through co‑ordinated packages rather than stand-alone interventions.

This evolving perspective on industrial policy has increasingly blurred the distinction with regional policy. Many industrial policy measures are now delivered at the subnational level and tailored to local capabilities, institutional contexts, and spatial development goals (OECD, 2025[49]). This reflects a growing recognition that national-level interventions often fail to address local specificities. At the same time, regional policies increasingly incorporate industrial policy objectives, such as fostering regional innovation and supporting strategic value chains, also in the light of the important role of spatial proximity in facilitating knowledge diffusion within technological clusters. By aligning national sectoral priorities with local capabilities, placed-based industrial policies mobilise underused assets across regions to create new sources of economic dynamism.

A good example are policies exploiting positive agglomeration effects by linking firms, universities, research centres and public authorities around shared regional specialisations. Examples include interventions in several OECD countries to support new or existing technological clusters, such as France’s Pôles de compétitivité, Canada’s Global Innovation Clusters and Germany’s Leading-Edge Cluster Competition. Specialisation Strategies (S3) in the European Union (EU) are also seen as a clear example of place‑based industrial policy because they explicitly tailor innovation and industrial priorities to the unique assets of each territory, using a participatory “entrepreneurial discovery” process and region‑specific governance to direct investment toward locally grounded strengths. Another common example are cluster policies, which aim to create the conditions under which technological clusters can emerge and grow. This includes providing shared public goods, such as research facilities, specialised skills and infrastructure, as well as strengthening collaboration and aligning regional strategies with industry needs. Moretti and Yi (2024[62]) provide evidence for the United States showing that market thickness or agglomeration effects enhance the adaptability of local labour markets to structural change.

Place‑based industrial policies should also promote sectoral diversification to limit local vulnerabilities and strengthen resilience to sector-specific shocks (OECD, 2025[49]). Diversification reduces dependence on a single industry and limits the impact of downturns in specific sectors. Governments can support this objective through targeted investment in education systems that build transferable skills, enabling workers to adapt across industries. Policies to attract foreign direct investment can introduce new technologies and expand emerging industries.

3.3.3. Employment and social policies

Employment and social policies play a crucial role in supporting the adaptability of regions to structural change by anticipating skills needs, alleviating the social consequences of job loss and supporting transitions between industries and regions.

In deindustrialising or resource‑dependent areas, active labour market policies – such as retraining, job-matching, and support for mobility – can complement regional and industrial policies that strengthen local demand, including support for small and medium-sized enterprises, social infrastructure, and public services. Active labour market policies provided by public (or private) employment services (PES) are central to this approach. By embedding employment support within the local context, providers of employment services help address specific local barriers, such as weak labour demand, specific skill needs and skill mismatches and limited transport connectivity). Evidence by Dorn and Lewell (2024[63]) suggests that active labour market policies can mitigate the impact of the China shock on employment in local labour markets.

Displaced workers in declining industries often face particular challenges, as industrial decline reduces local job opportunities. Chapter 3 of OECD (2024[64]) provides a detailed analysis of the challenges faced by workers in high-emission industries and discusses policies to support their transition. Well-designed out-of-work income support schemes, such as unemployment insurance and social assistance, can reduce earnings losses during periods of joblessness while supporting effective job search. Targeted measures, including in-work benefits and wage insurance schemes, can further facilitate transitions to new industries or regions where workers may initially face lower wages. Temporary wage supplements may also be linked to training investments to support longer-term upward mobility. Early intervention measures targeted at workers at risk of dismissal, alongside measures to manage collective redundancies, can further reduce the incidence and costs of displacement.

Beyond income support, structural transformation also increases the need for forward-looking upskilling and reskilling policies that prepare workers for emerging industries, occupations, and regions with stronger employment prospects (OECD, 2024[64]; 2025[65]). Strengthening skills anticipation systems is essential to identify evolving labour market needs early enough and ensure that training provision remains aligned with future demand. However, information about changing skills needs is often not communicated in a sufficiently accessible and actionable way, limiting participation in training. Career guidance plays a critical role in helping individuals understand both the opportunities and challenges associated with structural transformation. Flexible and modular learning programmes, combined with training leave and financial support, can help overcome barriers to adult learning (see Chapter 4). Training can also be made more accessible through work-based and on-site learning opportunities that provide practical experience alongside financial compensation. For example, initiatives in Australia and Canada provide targeted training pathways into growing sectors such as renewable energy and sustainable construction.

Structural change also affects the broader social fabric of local communities through rising joblessness, deteriorating health outcomes, social isolation, substance abuse, and family breakdown. Supporting declining regions therefore requires integrated support systems that combine employment services with mental health care, housing support, and childcare services (Fernandez et al., 2016[66]). Examples from Australia and South Korea, show how tailored well-being initiatives can help displaced workers rebuild purpose and reintegrate into the workforce (OECD, 2025[67]). Australia has implemented regional adjustment programmes in areas affected by industrial decline that combine training, peer support networks, and targeted mental health services with job placement assistance. In South Korea, comprehensive reemployment programmes for displaced workers include counselling, career coaching, and psychological support. Such integrated support systems may work better when the social partners are effectively engaged (OECD, 2019[68]).

3.4. Concluding remarks

Copy link to 3.4. Concluding remarks

This chapter examines how major forces of structural transformation – notably import competition from low-wage countries and innovation in digital technologies – can have a very heterogenous effect across local labour markets in OECD countries, depending on their initial sectoral specialisation and the composition of the workforce. Using linked employer-employee data for 14 OECD countries, it analyses how regions differ in their exposure to trade and technology shocks and how these shocks affect employment, workers, and firms. The central finding is that, while regional employment tends to prove resilient to structural shocks on average, adjustment involves profound shifts across industries in the most exposed places. This entails substantial and persistent social costs for incumbent workers whose jobs are destroyed, reflecting limited mobility across sectors and regions. Although new activities generate jobs over time, many displaced workers do not seize these opportunities and instead face prolonged unemployment or even withdraw from the labour force altogether. From a policy perspective the main challenge is how to limit local vulnerabilities and anticipate change by fostering industrial diversification, investing in education and infrastructure and anticipating skills needs, while alleviating the social costs by supporting displaced workers, removing barriers to job mobility sectors and regions and promoting job opportunities.

More attention should be paid to the social costs associated with structural change. Policy debates have long emphasised that structural transformation is welfare‑enhancing overall, with aggregate gains expected to outweigh losses and, in principle, allowing those adversely affected to be compensated. In practice, however, compensation and adjustment support have often fallen short. Moreover, models of structural change – particularly in the context of international trade – tend to adopt a long-run perspective in which workers are assumed to move smoothly across sectors and regions. This abstraction overlooks the reality that different activities require distinct skills and that many workers are strongly attached to their local communities, limiting their ability to consider a career change or willingness to relocate. As a result, adjustment costs are larger, more persistent, and more geographically concentrated than often anticipated. These gaps between theory and lived experience have contributed to a growing sense of discontent in some communities among those who felt “left behind” and may have undermined trust in government institutions and contributed to political polarisation (OECD, 2024[69]).

The work undertaken for this chapter represents an important investment in unlocking the regional dimension of linked employer-employee data. This progress opens multiple avenues for future research. One obvious possibility would be the development of a harmonised database on employment and wages – and, where feasible, productivity – at the level of commuting zones or small administrative units. Such a resource would enable secondary analysis of a wide range of policy-relevant questions, including the local labour market effects of other shocks such as natural disasters, the role of firms in shaping regional disparities, the dynamics and consequences of regional mobility, and the resilience of local economies to structural change. Strengthening the spatial granularity and comparability of microdata across countries would significantly enhance the evidence base needed to design policies that support workers and places through economic transformation.

References

[25] Abowd, J., F. Kramarz and D. Margolis (1999), “High Wage Workers and High Wage Firms”, Econometrica, Vol. 67/2, pp. 251-333, https://doi.org/10.1111/1468-0262.00020.

[33] Acemoglu, D. and P. Restrepo (2020), “Robots and Jobs: Evidence from US Labor Markets”, Journal of Political Economy, Vol. 128/6, pp. 2188-2244, https://doi.org/10.1086/705716.

[42] Aghion, P. et al. (2024), “Opposing Firm-Level Responses to the China Shock: Output Competition versus Input Supply”, American Economic Journal: Economic Policy, Vol. 16/2, pp. 249-269, https://doi.org/10.1257/pol.20210753.

[23] Ahrend, R. et al. (2026), “Measuring industrial restructuring pressure in regions: An employment-based index”, OECD Local Economic and Employment Development (LEED) Papers, No. 2026/01, OECD Publishing, Paris, https://doi.org/10.1787/14baf9f3-en.

[55] Ahrend, R. et al. (2026), “Resilient, bouncing back or trapped?: Mapping responses to industrial restructuring pressure”, OECD Local Economic and Employment Development (LEED) Papers, No. 2026/02, OECD Publishing, Paris, https://doi.org/10.1787/1817c1cc-en.

[45] Amior, M. and A. Manning (2018), “The Persistence of Local Joblessness”, American Economic Review, Vol. 108/7, pp. 1942-1970, https://doi.org/10.1257/aer.20160575.

[22] Autor, D. et al. (2025), “Places versus people: the ins and outs of labor market adjustment to globalization”, in Handbook of Labor Economics, Elsevier, https://doi.org/10.1016/bs.heslab.2025.07.004.

[76] Autor, D., D. Dorn and G. Hanson (2021), On the Persistence of the China Shock, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w29401.

[35] Autor, D., D. Dorn and G. Hanson (2019), “When Work Disappears: Manufacturing Decline and the Falling Marriage Market Value of Young Men”, American Economic Review: Insights, Vol. 1/2, pp. 161-178, https://doi.org/10.1257/aeri.20180010.

[20] Autor, D., D. Dorn and G. Hanson (2013), “The China Syndrome: Local Labor Market Effects of Import Competition in the United States”, American Economic Review, Vol. 103/6, pp. 2121-2168, https://doi.org/10.1257/aer.103.6.2121.

[2] Autor, D. et al. (2025), Places versus People: The Ins and Outs of Labor Market Adjustment to Globalization, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w33424.

[34] Autor, D. et al. (2014), “Trade Adjustment: Worker-Level Evidence *”, The Quarterly Journal of Economics, Vol. 129/4, pp. 1799-1860, https://doi.org/10.1093/qje/qju026.

[31] Battisti, M., C. Dustmann and U. Schönberg (2023), “Technological and Organizational Change and the Careers of Workers”, Journal of the European Economic Association, Vol. 21/4, pp. 1551-1594, https://doi.org/10.1093/jeea/jvad014.

[51] Becker, S., P. Egger and M. von Ehrlich (2012), “Too much of a good thing? On the growth effects of the EU’s regional policy”, European Economic Review, Vol. 56/4, pp. 648-668, https://doi.org/10.1016/j.euroecorev.2012.03.001.

[52] Becker, S., P. Egger and M. von Ehrlich (2010), “Going NUTS: The effect of EU Structural Funds on regional performance”, Journal of Public Economics, Vol. 94/9-10, pp. 578-590, https://doi.org/10.1016/j.jpubeco.2010.06.006.

[41] Bloom, N., M. Draca and J. Van Reenen (2015), “Trade Induced Technical Change? The Impact of Chinese Imports on Innovation, IT and Productivity”, The Review of Economic Studies, Vol. 83/1, pp. 87-117, https://doi.org/10.1093/restud/rdv039.

[16] Brynjolfsson, E., B. Chandar and R. Chen (2025), Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence.

[26] Card, D., J. Rothstein and M. Yi (2025), “Location, Location, Location”, American Economic Journal: Applied Economics, Vol. 17/1, pp. 297-336, https://doi.org/10.1257/app.20220427.

[77] Card, D., J. Rothstein and M. Yi (2024), “Industry Wage Differentials: A Firm-Based Approach”, Journal of Labor Economics, Vol. 42/S1, pp. S11-S59, https://doi.org/10.1086/728803.

[56] Causa, O., M. Abendschein and M. Cavalleri (2021), “The laws of attraction: Economic drivers of inter-regional migration, housing costs and the role of policies”, OECD Economics Department Working Papers, No. 1679, OECD Publishing, Paris, https://doi.org/10.1787/da8e368a-en.

[43] Causa, O. and J. Pichelmann (2020), “Should I stay or should I go? Housing and residential mobility across OECD countries”, OECD Economics Department Working Papers, No. 1626, OECD Publishing, Paris, https://doi.org/10.1787/d91329c2-en.

[57] Cavalleri, M., N. Luu and O. Causa (2021), “Migration, housing and regional disparities: A gravity model of inter-regional migration with an application to selected OECD countries”, OECD Economics Department Working Papers, No. 1691, OECD Publishing, Paris, https://doi.org/10.1787/421bf4aa-en.

[75] Citino, L. and A. Linarello (2021), “The impact of Chinese import competition on Italian manufacturing”, Review of International Economics, Vol. 30/3, pp. 702-731, https://doi.org/10.1111/roie.12587.

[60] Criscuolo, C. et al. (2022), “An industrial policy framework for OECD countries: Old debates, new perspectives”, OECD Science, Technology and Industry Policy Papers, No. 127, OECD Publishing, Paris, https://doi.org/10.1787/0002217c-en.

[40] Criscuolo, C. et al. (2023), “Worker Skills or Firm Wage-Setting Practices? Decomposing Wage Inequality Across 20 OECD Countries”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.4324842.

[28] 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, Vol. 12/6, pp. 1643-1675, https://doi.org/10.1111/jeea.12092.

[32] Dauth, W. et al. (2021), “The Adjustment of Labor Markets to Robots”, Journal of the European Economic Association, Vol. 19/6, pp. 3104-3153, https://doi.org/10.1093/jeea/jvab012.

[73] De Lyon, J. and J. Pessoa (2021), “Worker and firm responses to trade shocks: The UK-China case”, European Economic Review, Vol. 133, p. 103678, https://doi.org/10.1016/j.euroecorev.2021.103678.

[11] de Soyres, F. et al. (forthcoming), “From Partners to Rivals: The Global Trade Dynamic”.

[39] De Stefano, V. (2015), “The Rise of the ’Just-in-Time Workforce’: On-Demand Work, Crowd Work and Labour Protection in the ’Gig-Economy’”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.2682602.

[70] Dijkstra, L., H. Poelman and P. Veneri (2019), “The EU-OECD definition of a functional urban area”, OECD Regional Development Working Papers, No. 2019/11, OECD Publishing, Paris, https://doi.org/10.1787/d58cb34d-en.

[74] Donoso, V., V. Martín and A. Minondo (2014), “Do Differences in the Exposure to Chinese Imports Lead to Differences in Local Labour Market Outcomes? An Analysis for Spanish Provinces”, Regional Studies, Vol. 49/10, pp. 1746-1764, https://doi.org/10.1080/00343404.2013.879982.

[21] Dorn, D. and P. Levell (2024), “Labour market impacts of the China shock: Why the tide of Globalisation did not lift all boats”, Labour Economics, Vol. 91, p. 102629, https://doi.org/10.1016/j.labeco.2024.102629.

[63] Dorn, D. and P. Levell (2024), “Trade and inequality in Europe and the US”, Oxford Open Economics, Vol. 3/Supplement_1, pp. i1042-i1068, https://doi.org/10.1093/ooec/odad046.

[1] Dustmann, C. et al. (2025), “The Effects of Immigration on Places and People – Identification and Interpretation”, Journal of Labor Economics.

[5] Ellison, G. and E. Glaeser (1997), “Geographic Concentration in U.S. Manufacturing Industries: A Dartboard Approach”, Journal of Political Economy, Vol. 105/51, pp. 889-927, https://doi.org/10.1086/262098.

[14] Ernst, C., G. Michelena and P. Bertin (2026), Tariffs and Labor Markets: The Employment Impact of the Recent Trade Conflict. A Multiregional Input-Output Analysis, Elsevier BV, https://doi.org/10.2139/ssrn.5849582.

[3] Fadic, M. et al. (2019), “Classifying small (TL3) regions based on metropolitan population, low density and remoteness”, OECD Regional Development Working Papers, No. 2019/06, OECD Publishing, Paris, https://doi.org/10.1787/b902cc00-en.

[36] Feler, L. and M. Senses (2017), “Trade Shocks and the Provision of Local Public Goods”, American Economic Journal: Economic Policy, Vol. 9/4, pp. 101-143, https://doi.org/10.1257/pol.20150578.

[66] Fernandez, R. et al. (2016), “Faces of joblessness: Characterising employment barriers to inform policy”, OECD Social, Employment and Migration Working Papers, No. 192, OECD Publishing, Paris, https://doi.org/10.1787/5jlwvz47xptj-en.

[18] Filippucci, F., P. Gal and M. Schief (2024), “Miracle or Myth? Assessing the macroeconomic productivity gains from Artificial Intelligence”, OECD Artificial Intelligence Papers, No. 29, OECD Publishing, Paris, https://doi.org/10.1787/b524a072-en.

[27] Foliano, F. and R. Riley (2017), “International Trade and UK De-Industrialisation”, National Institute Economic Review, Vol. 242, pp. R3-R13, https://doi.org/10.1177/002795011724200110.

[82] Friesenbichler, K., A. Kügler and A. Reinstaller (2023), “The impact of import competition from China on firm‐level productivity growth in the European Union*”, Oxford Bulletin of Economics and Statistics, Vol. 86/2, pp. 236-256, https://doi.org/10.1111/obes.12574.

[15] Furceri, D. et al. (2019), “Macroeconomic Consequences of Tariffs”, IMF Working Papers, Vol. 19/9, p. 1, https://doi.org/10.5089/9781484390061.001.

[54] Gagliardi, L., E. Moretti and M. Serafinelli (2023), “The World’s Rust Belts: The Heterogeneous Effects of Deindustrialization on 1,993 Cities in Six Countries”, https://doi.org/10.3386/w31948.

[37] Graetz, G. and G. Michaels (2015), Robots at Work, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2575781 (accessed on 19 November 2018).

[17] Hampole, M. et al. (2025), Artificial Intelligence and the Labor Market, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w33509.

[80] Herrendorf, B., R. Rogerson and A. Valentinyi (2014), “Growth and Structural Transformation”, in Handbook of Economic Growth, Elsevier, https://doi.org/10.1016/b978-0-444-53540-5.00006-9.

[83] Kaplan, G. and S. Schulhofer‐Wohl (2017), “Understanding the long‐run decline in interstate migration”, International Economic Review, Vol. 58/1, pp. 57-94, https://doi.org/10.1111/iere.12209.

[50] Kline, P. and E. Moretti (2013), “Local Economic Development, Agglomeration Economies, and the Big Push: 100 Years of Evidence from the Tennessee Valley Authority *”, The Quarterly Journal of Economics, Vol. 129/1, pp. 275-331, https://doi.org/10.1093/qje/qjt034.

[8] Krugman, P. (1992), Geography and Trade, MIT Press.

[78] Malgouyres, C. (2016), “The impact of Chinese import competition on the local structure of employment and wages: Evidence from France”, Journal of Regional Science, Vol. 57/3, pp. 411-441, https://doi.org/10.1111/jors.12303.

[71] Melitz, M. (2003), “The Impact of Trade on Intra-Industry Reallocations and Aggregate Industry Productivity”, Econometrica, Vol. 71/6, pp. 1695-1725, https://doi.org/10.1111/1468-0262.00467.

[59] Millot, V. and Ł. Rawdanowicz (2024), “The return of industrial policies: Policy considerations in the current context”, OECD Economic Policy Papers, No. 34, OECD Publishing, Paris, https://doi.org/10.1787/051ce36d-en.

[81] Molloy, R., C. Smith and A. Wozniak (2011), “Internal Migration in the United States”, Journal of Economic Perspectives, Vol. 25/3, pp. 173-196, https://doi.org/10.1257/jep.25.3.173.

[62] Moretti, E. and M. Yi (2024), Size Matters: Matching Externalities and the Advantages of Large Labor Markets, National Bureau of Economic Research, Cambridge, MA, https://doi.org/10.3386/w32250.

[30] Murray, A. (2017), The Effect of Import Competition on Employment in Canada: Evidence from the ’China Shock’.

[53] Neumark, D. and H. Simpson (2015), “Place-Based Policies”, in Handbook of Regional and Urban Economics, Elsevier, https://doi.org/10.1016/b978-0-444-59531-7.00018-1.

[79] Nilsson Hakkala, K. and K. Huttunen (2016), Worker-Level Consequences of Import Shocks.

[58] OECD (2026), Reviving Productivity Growth in Canada, OECD Publishing, Paris, https://doi.org/10.1787/773dcd00-en.

[44] OECD (2025), Addressing Regional Labour Market Imbalances in Austria, OECD Reviews on Local Job Creation, OECD Publishing, Paris, https://doi.org/10.1787/0cf6186b-en.

[61] OECD (2025), “An institutional framework for industrial policies”, OECD Science, Technology and Industry Policy Papers, No. 180, OECD Publishing, Paris, https://doi.org/10.1787/0135eaba-en.

[13] OECD (2025), OECD Economic Outlook, Volume 2025 Issue 2: Resilient Growth but with Increasing Fragilities, OECD Publishing, Paris, https://doi.org/10.1787/9f653ca1-en.

[65] OECD (2025), OECD Skills Outlook 2025: Building the Skills of the 21st Century for All, OECD Publishing, Paris, https://doi.org/10.1787/26163cd3-en.

[49] OECD (2025), “Place-based industrial policy: Lessons for place transformation”, OECD Local Economic and Employment Development (LEED) Papers, No. 2025/03, OECD Publishing, Paris, https://doi.org/10.1787/43edc0df-en.

[67] OECD (2025), “Place-based policies for displaced workers: Lessons from past economic disruptions”, OECD Local Economic and Employment Development (LEED) Papers, No. 2025/04, OECD Publishing, Paris, https://doi.org/10.1787/69b86784-en.

[47] OECD (2025), Place-Based Policies for the Future, OECD Regional Development Studies, OECD Publishing, Paris, https://doi.org/10.1787/e5ff6716-en.

[19] OECD (2024), Job Creation and Local Economic Development 2024: The Geography of Generative AI, OECD Publishing, Paris, https://doi.org/10.1787/83325127-en.

[64] OECD (2024), OECD Employment Outlook 2024: The Net-Zero Transition and the Labour Market, OECD Publishing, Paris, https://doi.org/10.1787/ac8b3538-en.

[69] OECD (2024), OECD Survey on Drivers of Trust in Public Institutions – 2024 Results: Building Trust in a Complex Policy Environment, OECD Publishing, Paris, https://doi.org/10.1787/9a20554b-en.

[24] OECD (2023), “Grow and Go? Retaining Scale-ups in the Nordic Countries”, OECD Regional Development Papers, No. 51, https://doi.org/10.1787/9be5339d-en.

[68] OECD (2019), Negotiating Our Way Up: Collective Bargaining in a Changing World of Work, OECD Publishing, Paris, https://doi.org/10.1787/1fd2da34-en.

[48] OECD (2017), “Making policy evaluation work: The case of regional development policy”, OECD Science, Technology and Industry Policy Papers, No. 38, OECD Publishing, Paris, https://doi.org/10.1787/c9bb055f-en.

[12] OECD (forthcoming), “Adjusting to structural changes: A case study of the automotive sectors”, OECD Publishing, Paris.

[9] Prytkova, E. and F. Petit (2025), “The Employment Impact of Emerging Digital Technologies: Evidence from US Labor Markets”, in The Changing Nature of Work, National Bureau of Economic Research (NBER), University of Chicago Press, https://www.nber.org/books-and-chapters/changing-nature-work/employment-impact-emerging-digital-technologies-evidence-us-labor-markets.

[10] Prytkova, E. et al. (2024), “The Employment Impact of Emerging Digital Technologies”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.4739904.

[7] Rice, P. and A. Venables (2021), “The persistent consequences of adverse shocks: how the 1970s shaped UK regional inequality”, Oxford Review of Economic Policy, Vol. 37/1, pp. 132-151, https://doi.org/10.1093/OXREP/GRAA057.

[38] Rosenblat, A. and L. Stark (2015), “Uber’s Drivers: Information Asymmetries and Control in Dynamic Work”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.2686227.

[6] Rosenthal, S. and W. Strange (2004), “Chapter 49 Evidence on the nature and sources of agglomeration economies”, in Handbook of Regional and Urban Economics, Cities and Geography, Elsevier, https://doi.org/10.1016/s1574-0080(04)80006-3.

[4] Setzler et al. (forthcoming), “The local impact of structural change (Preliminary title)”.

[29] Taniguchi, M. (2019), “The effect of an increase in imports from China on local labor markets in Japan”, Journal of the Japanese and International Economies, Vol. 51, pp. 1-18, https://doi.org/10.1016/j.jjie.2018.09.001.

[72] Utar, H. (2018), “Workers beneath the Floodgates: Low-Wage Import Competition and Workers’ Adjustment”, The Review of Economics and Statistics, Vol. 100/4, pp. 631-647, https://doi.org/10.1162/rest_a_00727.

[46] Venables, A. (2024), “The case for place-based policy”, CEPR Policy Insight 128, https://cepr.org/publications/policy-insight-128-case-place-based-policy.

Annex 3.A. Data sources

Copy link to Annex 3.A. Data sources

Country	Name	Source	Sample	Period
Australia	ALIFE L-LEED	Tax administration	Universe	2001-2023
Austria	AMS-BMASK Arbeitsmarktdatenbank	Social security administration	Universe	1997-2024
Canada	Canadian Employer-Employee Dynamics Database	Tax administration	Universe	2001-2019
Denmark	Integrerede Database for Arbejdsmarkedsforskning (IDA) and other data from Statistics Denmark	Tax administration	Universe	1995-2022
Finland	FOLK employment data from Statistics Finland, Employer Payroll Report from Tax Admin.	Tax administration	Universe	2004-2022
France	Panel DADS	Social security administration	8.5% random sample of workers	1994-2021
Germany	Integrierte Erwerbsbiographien (IEB)	Social security administration	10% random sample of workers	2000-2019
Hungary	ADMIN – I – Panel of administrative data (OEP, ONYF, NAV, NMH, OH)	Social security administration	50% random sample of workers	2003-2017
Italy	Italian Social Security (INPS)	Social security administration	6.7% random sample of workers	1995-2022
Netherlands	CBS Microdata from Statistics Netherlands	Tax administration	Universe	2006-2019
Norway	Arbeidsgiver- og arbeidstakerregister (Aa-registeret), Lønns- og trekkoppgaveregisteret (LTO)	Tax administration	Universe	2001-2019
Portugal	Quadros de Pessoal	Mandatory employer survey	Universe	2002-2023
Spain	Muestra Continua de Vidas Laborales con Datos Fiscales (MCVL-CDF)	Social security and tax administration	4% random sample of workers	1995-2023
Sweden	Longitudinell integrationsdatabas för sjukförsäkrings- och arbetsmarknadsstudier (LISA), Företagens ekonomi (FEK), Jobbregistret (JOBB)	Social security administration	Universe	2002-2018
United States	Longitudinal Employer Household Dynamics (LEHD)	Social security administration (UI administrative records)	Universe	2000-2019

Annex 3.B. Additional results

Copy link to Annex 3.B. Additional results

Annex Figure 3.B.1. The effects of trade exposure by import region-of-origin and broad sector
Copy link to Annex Figure 3.B.1. The effects of trade exposure by import region-of-origin and broad sector

p.p. effect of a one standard-deviation increase in regional exposure to import competition from low-wage countries on regional employment divided by the initial working-age population, by broad sector, Western Europe (9)

Annex Figure 3.B.2. The effects of trade exposure by migration status and broad sector
Copy link to Annex Figure 3.B.2. The effects of trade exposure by migration status and broad sector

p.p. effect of a one standard-deviation increase in regional exposure to import competition from low-wage countries on regional employment divided by of the initial working-age population, by migration status and broad sector, France and Germany

Notes

Copy link to Notes

← 1. This chapter is the outcome of a collaboration between the Directorate of Employment, Labour and Social Affairs (ELS) and the Centre for Entrepreneurship, SMEs, Regions and Cities (CFE). The analysis using linked employer-employee builds on contributions of national experts from the OECD LinkEED 2.0 network (www.oecd.org/en/about/projects/linkeed-200.html): Patrick Bennett (University of Liverpool and IZA), Filipe Bento Caires (European University Institute), Dogan Gülümser (RF Berlin), Marie Kjeldsen (Aarhus University), Benjamin Lochner (FAU, IAB and IZA), Stefano Lombardi (VATT, IFAU, IZA and UCLS), Claudio Luccioletti (Bank of Italy), Tahsin Mehdi (StatCan), Jordy Meekes (Leiden University and IZA), Balázs Muraközy (University of Liverpool), Marco G. Palladino (Banque de France), Kjell Salvanes (NHH and IZA), Oskar Nordström Skans (Uppsala University, UCLS, IZA and IFAU), Rune Vejlin (Aarhus University and IZA) and Lennart Ziegler (University of Vienna). Bradley Setzler (Penn State) provided scientific guidance. The project is also part of the OECD Transforming Places project, funded by UKRI (https://www.oecd.org/en/about/projects/transforming-places.html). For additional information on the empirical analysis, please see the companion paper to this chapter by Setzler et al. (forthcoming[4]).

← 2. Structural changes in the composition of employment between industries in advanced economies since the early 2000s are typically attributed to a combination of technological change, growing trade integration and changes in consumer demand (partly as a result of population ageing) (Herrendorf, Rogerson and Valentinyi, 2014[80]). The present chapter focusses on key supply-side factors related to trade and technology. While demand-side factors also played an important role, these cannot be readily analysed in the analytical framework adopted in this chapter.

← 3. In a companion paper to this chapter by Setzler et al. (2026) differences in the adjustment response of regions to local exposure are examined as well as the factors driving them.

← 4. In addition to having detailed information on the location of the employer and the employee, the use of commuting zones also requires corresponding information on population size in the base year, which is not always available, especially when user-based definitions of commuting zones are employed.

← 5. A possible alternative to small administrative regions or commuting zones would be to make use of functional urban areas that connect cities with their surrounding commuting zones (Dijkstra, Poelman and Veneri, 2019[70]). However, functional urban areas do not cover rural areas.

← 6. Differences in specialisation across regions are notably more pronounced in the United States. In part, this may be a data issue since for the US commuting zones and 4‑digit NAICS industries are used, and sparsely populated regions may further contribute to this dispersion by amplifying concentration in a narrow set of industries. This also generates larger differences in regional exposure to trade and technology shocks, as shown in Figure 3.4.

← 7. Following the enlargement of the European Union to Central and Eastern Union in 2004 and 2007, most countries maintained significant restrictions on the movement of labour until 2011 and 2014. Sweden and the United Kingdom are notable exceptions in that they all allowed for free movement immediately after enlargement. Since the trade shock is measured over the period 2000-2007, it largely preceded the wave of immigration from Central and Eastern Europe.

← 8. The pooled regressions include country-fixed effects but are otherwise identical to the country-specific regressions.

← 9. Finland, Hungary, and the Netherlands are excluded from the baseline analysis of the trade shock as the data do not allow starting from 2002. These countries are included when considering the technology shock.

← 10. This is different from the change in the working-age employment rate. The latter takes account of both changes in working-age employment and changes in working-age population. Changes in working-age population may be important when trade exposure affects net migration between regions or countries. The effects on net migration are considered in Section 3.2.2 as part of the worker-level analysis. Results for the change in the working-age population are reported in Setzler et al. (forthcoming[4]).

← 11. For studies taking a firm-level or industry-level, see for example De Lyons and Pessoa (2021[73]) for the United Kingdom, Utar (2018[72]) for Denmark and Nilsson, Hakkala and Huttunen (2016[79]) for Finland.

← 12. While the magnitude of the effect in Canada may seem small, when expressing jobs lost as a share of the employment change in manufacturing between 2001‑2018, our estimate implies that about 20% of the decline in manufacturing employment can be attributed to import competition from low-wage countries. This is very similar to the estimate reported by Murray (2017[30]).

← 13. In fact, the effect on non-manufacturing is slightly negative. This may reflect spillovers effects from manufacturing to non-manufacturing due to input-output linkages, consistent with Murray (2017[30]).

← 14. As shown in Annex Figure 3.B.2 for France and Germany, job gains in non-manufacturing reflect to an important extent job gains among migrants. This suggests that the workers who take up employment in expanding non-manufacturing sectors are not the same as those who lose their job in manufacturing.

← 15. The empirical model relies on differences in industry shares within manufacturing across regions for identification.

← 16. For other country studies, please see amongst others Malgouyres (2016[78]) for France, Donoso (2014[74]) for Spain and Citino and Linarello (2021[75]) for Italy.

← 17. While the qualitative results for Canada are similar to those for other countries, the effect on employment in non-routine relevant technologies is much more pronounced. In part, this is likely to reflect data issues (e.g. regional definitions, missing values) and, in part, the higher regional concentration of ICT activities in places such as Monreal, Toronto and Vancouver with above-average employment growth. All in all, the quantitative estimates deserve to be interpreted with considerable caution.

← 18. Non-dependent employment includes self-employment, joblessness or being out of the country (international migration).

← 19. In the case of trade shocks, this is manufacturing versus non-manufacturing. In the case of the technology shock, we consider all industries in a region and, as a result, inter-industry flows do not contribute to net changes in regional employment. Inter-industry mobility is therefore not considered for the analysis of the technology shock.

← 20. Autor et al. (2013[20]; 2021[76]) further find for the United States that the China shock significantly increased long-term receipt of disability and retirement benefits, and to a lesser extent unemployment benefits.

← 21. The role of international migration for the decline in manufacturing employment is very small.

← 22. For the purposes of the decomposition, it assumed that workers retire at age 65 everywhere. This may be more appropriate for the United States than for Western Europe where the effective retirement age tends to be lower and early retirement is more common. This may explain in part why the contribution of the employment channel is more important than the ageing channel in Western Europe, notably in manufacturing.

← 23. By construction, net inter-industry outflows from manufacturing employment and net inflows into non-manufacturing employment are identical in absolute terms. Since non-manufacturing accounts for a large share of employment, inter-industry flows are even less important there.

← 24. Evidence from the United States points to a broad-based decline in internal migration since the 1980s, including a sharp fall in interstate mobility, while evidence for other OECD countries suggests that inter-regional mobility is often too limited to offset regional labour market shocks, partly reflecting housing costs, institutional barriers and uneven access to high-productivity labour markets, see e.g. Causa et al. (2021[56]), Kaplan and Schulhifer-Wohl (2017[83]) and Molloy et al (2011[81]).

← 25. The role of inter-industry mobility is not considered since the analysis of digital innovation differentiates between occupations rather than industries.

← 26. The wage‑setting practices of industries are measured by averaging firm wage premia – average firm wages conditional on workforce composition – across firms in the same industry following Card et al. (2024[77]).

← 27. Firm wage premia are sometimes used as a proxy for productivity when information on productivity is not. The seminal model of trade with heterogenous firms suggests that trade liberalisation increases the size of the most productive firms by creating new export opportunities while increased import competition reduces the size of the least productive or may even force them out of the market (Melitz, 2003[71]). By reallocating resources toward more productive firms, this is expected to contribute to aggregate productivity and economic welfare.

← 28. Using more recent data, Friesenbichler et al. (2023[82]) find these productivity effects may have become smaller or even negative as Chinese exports have gained in sophistication.

← 29. The lack of mobility between industries and regions may reflect price rigidities that prevent prices of non-tradable goods and wages from fully reflecting local economic conditions but also barriers to job mobility that prevent labour demand and supply from adjusting to relative prices. To the extent that older workers tend to be less mobile, population ageing is likely to have depressed mobility between sectors and regions, with potentially important implications for the adaptability of regions to structural change.

← 30. This characterisation of policies used in this section allow for separate discussions on regional, industrial and employment and social policies, consistent with the framework in Figure 3.1. Related policy discussions often make use of alternative definitions which either do not provide a neat delineation across policy domains, as in the case of more general discussions of place‑based policies or tend to emphasise one of policy domain using a broad definition that also captures elements of other policy fields.

← 31. Employment and social policies that are allocated according to local needs and delivered by local public employment services that tailor their policies to local conditions can be considered spatially targeted. By contrast, industrial policies aimed at specific industries may have a local dimension due to the existing industrial mix but are not considered spatially targeted unless they are designed and implemented in a place‑sensitive manner.

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment

Governance

Health

Industry, business and entrepreneurship

Regional, rural and urban development

Science, technology and innovation

Society

Taxation

Trade

Energy

Nuclear energy

Transport

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment

Governance

Health

Industry, business and entrepreneurship

Regional, rural and urban development

Science, technology and innovation

Society

Taxation

Trade

Energy

Nuclear energy

Transport

Countries A - C

Countries D - I

Countries J - M

Countries N - R

Countries S - T

Countries U - Z

Regional and global engagement

Countries

Countries A - C

Countries D - I

Countries J - M

Countries N - R

Countries S - T

Countries U - Z

Regional and global engagement

Publications

Publications

Featured publications

Data

Data

Featured data

News & events

News & events

Featured events

About OECD

About

Engage with us

Work with us

Featured topics

Agriculture and fisheries

Climate change

Development

Digital

Economy

Education and skills

Employment

Environment

Finance and investment