International Migration Outlook 2025

4. Immigrant integration: The role of firms

César Barreto
OECD

Ana Damas de Matos
OECD

Alexander Hijzen
OECD

Abstract

This chapter sheds new light on the integration of immigrants in the labour market by focussing on the role of firms in shaping the immigrant earnings gap using linked employer-employee data for 15 OECD countries. The chapter starts by documenting the immigrant earnings gap at entry in the labour market and the extent to which it is driven by immigrants working disproportionately in lower-paying firms, industries and occupations. It then analyses how the earnings gap evolves as immigrants integrate in the labour market by moving to better-paying firms, industries and occupations. The implications for immigrant integration policy are discussed in the conclusion.

This project is part of the OECD LinkEED 2.0 project which mobilises linked employer-employee data for cross-country research and policy analysis (https://www.oecd.org/en/about/projects/linkeed-200.html). It relies on the contributions of a network of researchers with access to confidential linked employer-employee data for their country, including: Winnie Chan and Tahsin Mehdi (Canada), Lukas Delgado-Prieto (Colombia), Paolo Santini (Denmark), Stefano Lombardi (Finland), Yajna Govind (France), Gaetano Basso (Italy), Jordy Meekes (the Netherlands), Nazila Alinaghi and Lucas Chen (New Zealand), Patrick Bennett and Kjell Salvanes (Norway), Andrei Gorskov and Oskar Nordstrom-Skans (Sweden), Erik Vickstrom (United States).*

The Secretariat is also grateful to Jaime Arellano-Bover for his comments and suggestions.

*Any opinions and conclusions expressed herein are those of the author and do not represent the views of the US Census Bureau. The Census Bureau has ensured appropriate access and use of confidential data and has reviewed these results for disclosure avoidance protection (Project 7526852: CBDRB-FY25‑CES023‑009).

In Brief

• Immigrants fulfilling their earnings’ potential in the host country is the foundation for better living conditions and a safeguard against poverty. While there has been substantial progress in measuring integration outcomes, there is little cross-country evidence on the earnings of immigrants, mainly due to data availability.

• This chapter provides cross-country evidence on the size and sources of the immigrant earnings gap and its evolution over time in the host country labour market. To this end, it makes use of linked employer-employee panel data for the period 2000‑2019 for 15 OECD countries: Austria, Canada, Colombia, Denmark, Finland, France, Germany, Italy, Portugal, the Netherlands, New Zealand, Norway, Spain, Sweden and the United States.

• Immigrants at entry in the labour market earn 34% less than native‑born workers of the same age and sex, on average across countries. The immigrant earnings gap decreases by about one‑third (13 percentage points (p.p.)) in the first five years in the host country, and by about half in the first ten years.

• While immigrants with lower initial earnings are more likely to drop out of employment in the host country, this composition effect is small and does not drive the reduction in the earnings gap.

• Almost two‑thirds (63%) of the initial immigrant earnings gap can be explained by the concentration of immigrants in lower-paying firms and sectors, on average across countries. Work in lower-paying sectors accounts for over one‑third (36%) of the gap, and working in lower-paying firms, within sectors and regions, accounts for an additional quarter (27%).

• In the first years in the labour market, immigrants move to higher-paying sectors and firms. Moving to higher-paying sectors accounts for 18%, and to higher-paying firms for 24%, of the reduction of the gap in the first five years in the labour market. The firms immigrants move to over time are higher quality firms in a general sense: they are larger, more productive and less segregated.

• Working in lower-paying occupations accounts for a quarter of the initial immigrant earnings gap, on average for the six countries with available data. For this subset of countries, sectors and firms account nevertheless for about half of the earnings gap. The occupational effect is due partly to widespread immigrant overqualification in the host country labour markets. Strikingly, immigrants do not move to higher-paying occupations over time.

• Accounting for educational differences between immigrants and the native‑born for two countries with available data decreases the initial earnings gap by one‑quarter. However, it does not change the reduction of the earnings gap by much, nor does it change the role of firms and sectors in explaining the immigrant earnings gap and its reduction over time.

• The immigrant earnings gap at labour market entry is larger for immigrants from Africa, the Middle East and Asia, and smaller for immigrants from the EU15 and North America and Oceania. This is partly due to immigrants from Africa, the Middle East and Asia working in lower-paying sectors and firms. Immigrants of all origin groups move to higher paying firms over time.

• The earnings gap is slightly larger for men than women, and the reduction of the gap is larger for women. This reflects in part the greater employment concentration of male immigrants in lower-paying firms and its relative persistence over time.

• This chapter highlights the importance of policies related to the recognition of foreign qualifications and upskilling, given the persistence of the large share of the immigrant earnings gap that is due to immigrants working in lower-paying occupations.

• Nevertheless, it also shows that job mobility to better paying firms and sectors can play an important role in the earnings integration of immigrants. This suggests that there is scope for policies that target barriers to job mobility including providing information on job search and the host country labour market, career counselling and the development of professional networks, improving local transportation, combatting discrimination in the housing market and providing affordable housing, among others.

Introduction

A good understanding of the integration of immigrants in the host countries’ labour market and society is necessary to design efficient and effective integration policies. The outcomes of immigrants are well documented in an increasing number of areas: educational and labour market outcomes, living conditions, health outcomes, or civic participation (OECD/European Commission, 2023[1]).

When it comes to the integration of immigrants in the labour market, a key indicator of successful integration of employed immigrants is their earnings. Immigrants fulfilling their earnings’ potential in the host country labour market is the foundation for better living conditions and a safeguard against poverty. Beyond the benefits to immigrants themselves, higher earnings enhance immigrants’ contribution to the host country economy and are an important factor in the attraction and retention of immigrants.

While a rich and long-standing literature has documented the earnings of immigrants and their growth over time in selected host countries (e.g. Chiswick (1978[2]), Lubotsky (2007[3]), Abramitzky, Boustan and Eriksson (2014[4])), cross-country studies are scarce. This is primarily due to the lack of earnings data that are reliable and comparable across countries.

Most existing studies on immigrant earnings have taken a human-capital approach that focusses on immigrants’ characteristics as a source of the immigrant earnings gap. One explanation is that immigrants may be initially paid less because they are less productive than the native‑born due to lower educational attainment, low portability of skills acquired abroad, or lack of host country language fluency, but as they accumulate host-country specific human capital, their earnings increase.

A complementary explanation is that different firms pay equally productive workers differently. There is growing evidence that, in imperfectly competitive labour markets, earnings are not just determined by one’s human capital but also by one’s employer (Card, Heining and Kline, 2013[5]; Song et al., 2019[6]; OECD, 2022[7]). Consequently, immigrants may have lower earnings if they are more likely to be employed in lower-paying firms. This may be due to immigrants having smaller local networks, difficulties navigating the host country labour market, or hiring discrimination. In this case, job mobility between firms can play a potentially important role in advancing the integration of immigrants in the labour market.

The objective of this chapter is to provide comprehensive cross-country evidence on the sources of the immigrant earnings gap and its evolution over time since first entering the host country labour market. It focusses on three key questions. First, what is the importance of immigrant sorting across firms, industries, and occupations, in explaining the immigrant earnings gap at entry in the labour market? Second, to what extent do immigrants integrate in the labour market by moving to better paying firms, industries and occupations over time? Third, to what extent do differences in the characteristics of immigrants (e.g. country of origin, sex) matter for the earnings gaps at entry and their subsequent integration in the labour market?

To address these questions, the chapter uses linked employer-employee panel data for 15 OECD countries.1 For most countries, the data is based on administrative sources related to the social security or the tax system. The data are very comprehensive – often covering the universe of workers and their employers in the country – and provide highly reliable information, notably on earnings.

Studying the importance of firms for the immigrant earnings gap over time has important policy implications. Most policies targeted at the labour market integration of immigrants have focussed on the worker side, such as providing host country language classes, upskilling or facilitating the recognition of foreign qualifications. This chapter argues that policies that promote immigrant job mobility, which so far have received less attention from policymakers, should also be part of the immigrant integration policy toolkit.

The chapter is organised as follows. The first section lays out the data and the methodology for analysing the immigrant earnings gap and provides a description of the characteristics of immigrants included in the analysis. The second section provides cross-country estimates of the immigrant earnings gap at entry in the host labour market and over time. The third section estimates the contributions of firms, industries and occupations to the immigrant earnings gap. The fourth section presents results by region-of-origin and sex. The conclusion draws out the implications of the analysis for immigrant integration policy.

The framework

The empirical approach

This chapter studies the earnings of immigrants entering the labour market from the early 2000s to 2015 over five years, and up to ten years for the earlier cohorts.2 The earnings of immigrants at entry in the labour market, and in each subsequent year, are compared to those of the native‑born of the same age and sex, in the same calendar year. The immigrant earnings gap after five, and ten years, is compared with the initial gap to assess the reduction of the earnings gap as immigrants integrate in the host country labour market.

As a second step, the immigrant earnings gap in each year since entry is decomposed into differences due to immigrants working in lower-paying firms and sectors. For countries with information on occupations, the extent to which the earnings gap is due to immigrants working in lower-paying occupations is also analysed.3 The methodology is described in detail in Box 4.1.

Box 4.1. Methodology

Estimating the integration profiles of immigrants

The integration profiles of immigrants, i.e. the differences in immigrant outcomes relative to the native‑born with the same age and sex over time in the host country, are estimated using the following benchmark specification:

$\left(1\right) Y=X{\beta }_1+YSE{\delta }_1+u_1$

$Y$ represents the integration outcome for worker $i$ in year $t$ at its main employer in the year.1 Integration outcomes are worker wages, measured using the logarithm of real monthly earnings,2 or firm characteristics. The latter includes firm productivity (value added per worker or, if not available, sales per worker); firm wage premia (average firm wage conditional on worker composition); firm size (employment) and the degree of immigrant segregation (a dummy which equals one in firms where immigrants account for more than 50% of the workforce and zero otherwise). $X$ denotes a set of control variables which includes a quartic in age fully interacted with sex and year fixed effects. $YSE$ denotes a full set of years-since‑entry dummies or, more precisely, years since first entry in formal employment. The coefficient of interest, ${\delta }_1$, represents for each year since entry the average difference in the outcome of interest between immigrants and the native‑born of the same age and sex, controlling for common year effects. $u_1$ stands for the error term. The subscript 1 denotes the benchmark specification (Equation 1).

The benchmark specification is estimated by focussing on the first five years since first entry in formal employment to maximise the number of immigrant entry cohorts included in the analysis. The analysis tracks the evolution of immigrants’ outcomes following the first spell in formal employment, since in most countries the data do not provide information on the actual year of arrival in the host country.

The role of firms and occupations in the immigrant earnings gap: A Gelbach decomposition

To analyse the contribution of firms and occupations to the immigrant earnings gap at each year since entry estimated above, the benchmark specification is extended to include firm and (2‑digit) occupation fixed effects that control for differential sorting of immigrants relative to native born across firms and occupations, as follows:

$\left(2\right)\mathrm{ }Y=X{\beta }_2+YSE{\delta }_2+F{\psi }_2+O{\gamma }_2+u_2$

where ${\psi }_2$ and ${\gamma }_2$ denote the firm and occupation fixed effects respectively, while ${\delta }_2$ represents the unexplained earnings gap after controlling for firms and occupations.

The difference between ${\delta }_1$ and ${\delta }_2$ represents the part of the earnings gap that can be explained by differences in the sorting of immigrants and native‑born across firms and occupations. The differences in the explained earnings gap ${\delta }_1- {\delta }_2$ at each year since entry (YSE) is decomposed into parts attributable to firm and occupational sorting following (Gelbach, 2016[8]).3 The Gelbach-decomposition allows for the unambiguous determination of the contribution of firms (${\tau }_{\psi }$) and occupations $\left({\tau }_{\gamma }\right)$ to the explained part of the earnings gap at each year since entry:

${\left(3\right)\mathrm{ }\delta }_1-{\delta }_2={\tau }_{\psi }+{\tau }_{\gamma }$

In practice, ${\tau }_{\psi }$ is obtained by regressing the estimated firm-fixed effects from the full specification (Equation 2) on the covariates of the baseline specification (Equation 1), YSE and X. Analogously, ${\tau }_{\gamma }$ is obtained by regressing the estimated occupation fixed-effects in the full specification on the covariates of the baseline specification. The contribution of firms ${\tau }_{\psi }$ captures the impact of all time‑invariant characteristics of firms on the immigrant earnings gap. This includes the firm-specific characteristics as well as those of the broader industry and region groupings to which they belong (Card, Rothstein and Yi, 2024[9]; Card, Rothstein and Yi, 2025[10]).4

To understand the role of the firm net of that of regions and sectors, the contribution of firms (${\tau }_{\psi }$) is further decomposed into the contribution of regions (${\tau }_{\theta }$), the contribution of sectors (${\tau }_{\mathrm{\chi }}$) and the contribution of firms within regions and sectors (${\tau }_{\mathrm{v}}$). In practice, this is achieved by regressing the estimated firm-fixed effects from the full specification on regional, sectoral dummies and a residual term. The residual from this regression captures the contribution of firms within regions and sectors. Estimates of ${\tau }_{\theta },{\tau }_{\mathrm{\chi }}$ and ${\tau }_{\mathrm{v}}$ are obtained from separate regressions using the estimated sector, region fixed effects and residuals as dependent variables and regressing them on the covariates of the base model, YSE and X.

1. The main employer in a year typically corresponds to the employer with the highest reported earnings among all reported employment spells in the year or at the date in which the data is collected (e.g. October in Portugal). Except for Germany, the employer corresponds to the firm and not the establishment.

2. The focus is on monthly earnings to maximise country coverage and comparability. It has the caveat that some of the earnings catch-up can be attributed to improvements in hours worked and not necessarily hourly wages. To account for this, whenever possible, we also estimate the integration profiles based on hourly wages.

3. More details and a theoretical derivation of the Gelbach decomposition can be found in Annex 4.A.

4. The specification does not control for worker fixed effects; the firm component may partly capture unobserved worker heterogeneity. To address this, in addition to the Gelbach decomposition, we relate the wage gaps to AKM firm wage premia that controls for unobserved worker heterogeneity, with similar findings (Abowd, Kramarz and Margolis, 1999[11]).

Longitudinal linked employer-employee data

The framework is implemented using harmonised longitudinal linked employer-employee data for 15 OECD countries: Austria, Canada, Colombia, Denmark, Finland, France, Germany, Italy, Portugal, New Zealand, the Netherlands, Norway, Spain, Sweden and the United States.

The data are drawn from administrative records designed for tax or social security purposes or, in a few cases, mandatory employer surveys. As a result, these data are very comprehensive and of high quality, given the financial implications of reporting errors for tax and social security systems. In nine of the 15 countries, the data covers the universe of workers in the labour market; in the other six countries, a large representative random sample is used instead (Annex Table 4.A.1).4 Importantly, time‑invariant worker and firm identifiers allow the outcomes of workers and their employers to be followed over time.

Since the data are collected for national administrative purposes, the data are not necessarily comparable across countries. To enhance the cross-country comparability of the data, considerable efforts have been made to harmonise them in terms of data cleaning, sample selection (e.g. years and immigrant cohorts covered) and definition of variables (e.g. industry classification, earnings, regions of origin).

The main outcome of interest is individual monthly earnings. Monthly earnings capture both hours worked and hourly earnings. To disentangle these two margins of the immigrant earnings gap, hourly earnings are studied as an additional outcome for countries with available data on hours worked (Denmark, France, Italy, the Netherlands and Portugal). For Canada, Norway and the United States, data on yearly earnings are used instead, since there is no information on the number of months worked.

Immigrants are identified based on their nationality at entry in the labour market. The term immigrants in this chapter therefore refers to foreign nationals instead of foreign-born as is standard practice in OECD studies.5 Given that immigrant status is defined at entry in the labour market, citizenship acquisition after entry is not a confounding factor. This definition of immigrants may however include individuals born in the host country who do not have the host country’s nationality, such as the native‑born adult children of immigrants in some OECD countries.

The chapter focusses on immigrants’ first years in formal employment in the host country, that is, the first year they are in the data. To avoid including workers who are not entering the labour market but only returning from a spell out of work, the analysis includes only immigrants who are not in employment in the first three years of the data but enter employment in the subsequent years.6

A limitation of administrative data, relative to survey data for example, is the little information on individual characteristics. Information on educational attainment in particular, is not available for all countries. Hence, the immigrant earnings gap do not account for educational differences in the main results. However, additional results taking into account educational differences between immigrants and the native‑born are presented for countries with available data.

The characteristics of new immigrants in the labour market

The analysis covers all immigrants who first entered the labour market of the 15 OECD countries in the early 2000s to the mid‑2010s. Over this period, over 7 million new immigrants are observed in the data (Annex Table 4.B.1). Women represent 43% of all immigrants in the sample. The average age of immigrants at entry in the host country labour market is 32 years.

Immigrants’ regions-of-origin vary significantly between host countries. These differences partly reflect geography and historical ties between some origin and destination countries. In Portugal, Spain and the United States, a large share of immigrants come from Latin American countries (33%, 22% and 40%) (Figure 4.1). In Portugal, these mainly come from Brazil, while in Spain and the United States they mainly come from Spanish-speaking Latin-American countries. In France, over half of immigrants come from African countries (33% from North Africa and 21% from Sub-Saharan Africa), mostly from predominantly French-speaking former colonies. In Canada and New Zealand, close to half of immigrants come from Asia.

Figure 4.1. The region of origin of new immigrants in the labour market

Share of origin region in all immigrant labour market entries, percentage

Note: The figure shows the distribution of immigrants across regions of origin as measured in the first year in the host country labour market. Countries are sorted by the share of EU (EU‑15 + Other EU27) immigrants. OECD refers to the unweighted average across countries.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details. Authors’ calculations.

Statlink

The differences in regions of origin across OECD countries stem also from differences in the category of migration of the migration flows (e.g. labour, family humanitarian). Over half the immigrants in Austria, Denmark, the Netherlands, Norway come from another EU‑27 country. This reflects the importance of free movement migration within the EU-EFTA area to these countries. In Sweden, 30% of immigrants are from the Middle East, in part reflecting the importance of humanitarian migration. Information on immigrants’ category of migration is available in the administrative data in some countries. Box 4.5 presents results by category of migration for Canada.

The immigrant earnings gap at entry in the host country labour market and over time

Immigrants earn substantially less than the native‑born at entry in the labour market, but the gap reduces over time

At entry in the host country labour market, immigrants face large pay penalties relative to the native‑born of the same age and sex in all OECD countries included in the analysis. The earnings of immigrants are on average across the 15 OECD countries 34% lower than those of the native‑born (Figure 4.2).7 The earnings gap ranges from around 28% in Denmark, France and Portugal to 45% in Italy.

By the fifth year in the host country labour market, the earnings gap relative to the native‑born is significantly lower than the gap at entry. The earnings gap decreases by one‑third on average across countries, or 13 p.p., from 34% on entry to 21% after five years.8 Approximately half of the decrease of the immigrant earnings gap takes place from the first to the second year in the host country labour market, on average across countries.

For the earlier immigrant cohorts, for which it is possible to estimate the earnings gap after ten years, the gap more than halves, decreasing from 37% initially to 16% by the tenth year (Annex Figure 4.B.1).9 Hence, almost two‑thirds of the decrease in the immigrant earnings gap over ten years takes place in the first five years.

The decrease in the immigrant earnings gap over time reflects a significant increase in real earnings of immigrants. After five years in the labour market, the average earnings of immigrants are 24% higher than those of immigrants in the first year across the 15 host countries (Annex Table 4.B.2).

Figure 4.2. The immigrant earnings gap decreases with time spent in the host country labour market

Immigrant earnings gap in the first, second and fifth year since entry in the host country labour market, percentage

Notes: * The earnings measure corresponds to annual earnings. In countries with annual earnings information, the earnings gap in the first year corresponds to that recorded in the second year to correct for differences in days worked and mechanically low earnings induced by the day of entry in the labour market. ** For Colombia, the earnings gap is estimated after four years, instead of five, due to the short panel structure.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details. Authors’ calculations.

Statlink

Immigrant earnings gaps partly reflect differences in education and hours worked

Part of the immigrant earnings gap may be due to differences in skills if immigrants have lower skill levels than the native‑born. Alternative data sources show this is only partly the case. Immigrants are more likely, than the native‑born, to not have completed an upper secondary degree but also more likely to be tertiary educated in most host countries (Annex Table 4.B.3).

Controlling for education in countries where data is available (Germany, Portugal) reduces the initial earnings gap by about a quarter.10 However, taking differences in education into account only marginally changes the reduction of the earnings gap over time.11 The next sub-section shows that, in contrast, differences in the sectors, firms and occupations in which immigrants work account for a large share of the immigrant earnings gap and its reduction.

Differences in hours worked explain part of the initial immigrant earnings gap and its decrease over time. For the six countries with information on hours worked, Box 4.2 compares the monthly and hourly immigrant earnings gap at entry in the labour market and after five years. Working fewer hours than the native‑born accounts for around one‑quarter of the initial immigrant earnings gap on average across these countries. This is consistent with evidence that shows that recent immigrants are more likely to hold part-time jobs, and in particular involuntary part-time jobs (OECD/European Commission, 2023[1]). As immigrants spend more time in the host country labour market, they work more hours, contributing to higher earnings. Increased working hours account for one fifth of the decrease in the immigrant earnings gap in the first five years, on average across the six countries.

Box 4.2. Hours worked explain part of the initial immigrant earnings gap and its decrease over time

Part of the initial immigrant earnings gap is due to immigrants working fewer hours per month than the native‑born. The initial immigrant gap in hourly earnings is significantly lower than the gap in monthly earnings in all six countries for which data on hours worked are available. On average, the immigrant gap in hourly earnings is around one‑quarter lower than the monthly earnings gap (26% compared to 34%).

The role of hours worked is particularly large in Denmark and France. The initial hourly earnings gap represents only 42% and 53% of the monthly earnings gap in these countries. In contrast, in Portugal, hours worked are more similar for immigrants and the native‑born. The initial hourly earnings gap represents 90% of the monthly earnings gap.

By the fifth year in the labour market, hours worked acount for less of the monthly immigrant earnings gap than at entry. Immigrants work more hours as they integrate in the host country labour market, and this contributes to raising their monthly pay. One fifth of the decrease of the immigrant monthly earnings gap in the first five years is accounted for the increase in hours worked, on average across the six countries. The increase in hours worked is particularly important in Denmark and France, where the initial gaps in hours are larger.

Figure 4.3. Hours worked explain part of the initial immigrant earnings gap and its decrease over time

Immigrant monthly and hourly earnings gap in the first and fifth year since entry in the host country labour market, percentage

Note: For Germany, the sample is restricted to full-time workers for the calculation of hourly earnings gaps. Countries are ranked based on the monthly earnings gap in the fifth year.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details. Authors’ calculations.

Statlink

The reduction of the immigrant earnings gap is not driven by less successful immigrants dropping out of employment in the host country

A concern when studying the earnings of immigrants over time is that they may reflect changes in the composition of immigrants, due to outmigration or differences in labour market attachment. For example, the least successful immigrants in the labour market may be more likely to drop out of employment or leave the host country altogether (Lubotsky, 2007[3]). As a result, the composition of immigrants remaining in employment in the host country would shift towards more successful immigrants and reduce the earnings gap even if immigrants did not actually improve their labour market position. To address this concern, at least partly, the analysis is repeated by restricting the sample to immigrants who are still in employment in the host country after five years.

Across the OECD countries included in the analysis, immigrants who remain employed in the host country for five years or more had a smaller initial earnings gap, relative to the native‑born, than the initial gap estimated for all immigrants (irrespective of duration of stay) (Figure 4.4).

The selection effect is small on average across countries. The initial earnings gap of immigrants who are in employment at year five is 2 p.p. lower than the overall initial gap (32% compared with 34%). The selection effect tends to be larger in the host countries with the largest initial earnings gaps, such as Italy and Sweden but also in the United States. The selection effect is small in Germany, despite a large initial earnings gap, as well as in Portugal and the Netherlands.

Consequently, the positive selection of immigrants in employment does not drive the estimated decrease in the immigrant earnings gap. When focussing only on immigrants who remain employed in the host country for at least five years, the reduction of the earnings gap is estimated at 11 p.p., very close to the 13 p.p. when considering all immigrants entering the host country labour market.

Figure 4.4. Immigrants who remain employed in the host country after five years are positively selected

Immigrant earnings gap at entry for all immigrants and those observed in employment after five years or more in the country, percentage

Note: * The earnings measure corresponds to annual earnings. In countries with annual earnings information, the earnings gap in the first year corresponds to that recorded in the second year to correct for differences in days worked and mechanically low earnings induced by the day of entry in the labour market. ** For Colombia, the earnings gap is estimated after four years, instead of five, due to the short panel structure.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details. Authors’ calculations.

Statlink

The role of firms, sectors and occupations in the immigrant earnings gap

Almost two‑thirds of the initial immigrant earnings gap is due to immigrants working in lower-paying sectors and lower-paying firms within sectors

Almost two‑thirds (63%) of the initial immigrant earnings gap can be explained by the concentration of immigrants in lower-paying firms and sectors, on average across countries (Figure 4.5). These are sectors, and firms within sectors, where all workers, both foreign and native‑born, receive on average lower earnings. The combined contribution of firms and sectors to the immigrant earnings gap is substantial in all countries, ranging from 44% in Finland to 85% in Colombia. In contrast, immigrants tend to concentrate in higher-paying regions, such as urban areas. The positive impact of regions on the earnings gap is however very small in all countries.

Figure 4.5. Working in lower-paying firms and sectors accounts for over half of the initial immigrant earnings gap

Immigrant earnings gap at entry in the host country labour market, decomposed into the contribution of sectors, regions, firms (within sector and region) and a within-firm component, percentage

Note: The overall earnings gap corresponds to the average monthly earnings difference between immigrants upon entry in the host country and the native‑born workers of the same age and sex in the same year. The decomposition of the earnings gap follows Gelbach (2016[8]), see Box 4.1 for details. * The earnings measure corresponds to annual earnings. In countries with annual earnings information, the earnings gap in the first year corresponds to that recorded in the second year to correct for differences in days worked and mechanically low earnings induced by the day at entry in the labour market.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details. Authors’ calculations.

Statlink

The concentration of immigrants in lower-paying sectors alone accounts for over one‑third (36%) of the earnings gap on average across countries. The contribution of sectors to the earnings gap is large for all countries and ranges from 18% in Colombia to around 50% in Austria, Germany and Portugal. For example, in Portugal, 57% of all immigrants at labour market entry work in low-wage sectors such as accommodation and food, administrative service activities and construction.

Across OECD countries, immigrants are strongly concentrated in a few sectors, such as construction, domestic services, accommodation and food services, manufacturing, healthcare or information technology (OECD, 2020[12]). When entering the host country labour market, immigrants in this analysis are initially overrepresented in administrative service activities (which includes security and cleaning services to buildings), accommodation and food services and agriculture (Annex Table 4.B.4). These patterns are broadly similar across immigrant cohorts and countries.

Immigrants are not only concentrated in some sectors of the economy, but they also concentrate in specific firms within sectors. Across the 15 OECD countries considered in this chapter, the concentration of immigrants in lower-paying firms within industries accounts for about one‑quarter (27%) of the immigrant earnings gap at entry in the host country labour market.12 The contribution of firm-pay differences within industries to the immigrant earnings gap ranges from 13% to 40% in most countries. Two notable exceptions are the Netherlands where firm-pay differentials explain very little of the immigrant earnings gap, and Colombia, where firm effects account for two‑thirds of the earnings gap. The importance of firms in the immigrant earnings gap in Colombia may to some extent reflect the greater importance of differences in wage‑setting practices and performance across firms, as is the case in many emerging economies (Kline, 2024[13]).

Lower-paying firms are often also firms of low quality in a more general sense (Figure 4.6). Immigrants not only sort into lower-paying firms when first entering the host country labour market but also in firms that are smaller, less productive (measured by lower value added per worker) and substantially more segregated than firms that employ their native‑born counterparts in all countries.

The evidence presented above adds to a growing academic literature on immigrant firm segregation from across the OECD.13 Different mechanisms may be at play that explain immigrant workplace segregation. Many workers find employment through job referrals and their informal networks. If immigrants’ networks are disproportionately composed of other immigrants, then immigrants will tend to work together. Employer discrimination may also be at play, leading immigrants to concentrate in workplaces that do not discriminate against immigrants at hire.

Figure 4.6. Immigrants work in lower quality firms at entry in the labour market

Characteristics of the firms where immigrants work at entry in the labour market, relative to the native‑born of the same age and sex

Note: The firm fixed-effect is a firm earnings premium paid to all workers in a given firm net of worker composition following (Abowd, Kramarz and Margolis (1999[11]). Firm productivity is measured in terms of value‑added per worker, except for Portugal, where it is measured as sales per worker. Firm size refers to the number of employees in a given firm and year. Firm segregation is a dummy variable equal to one if immigrants comprise more than 50% of all workers in a given firm and year. Information on firm-level productivity is not available in Austria, Colombia, Germany, Spain, France, the Netherlands and the United States. The share of immigrants in the firm is not available in Spain, Italy and the United States. For Spain, Italy and France, the firm fixed-effects do not control for worker composition due to small sample sizes.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details. Authors’ calculations.

Statlink

Differences in pay between immigrants and the native‑born within the same firm account for the remaining initial immigrant earnings gap on average across countries (42%). This within-firm component ranges from 16% and 22% in Colombia and Spain to over 50% in countries such as Canada, Denmark, Finland, New Zealand or the United States. Within-firm earnings gaps may reflect differences in the type of work immigrants do within firms (i.e. differences in occupations), differences in their productivity within those occupations related to tenure, language fluency and educational attainment, or differences in pay for work of equal value (e.g. bargaining, discrimination).

Additional analysis for Germany and Portugal, for which data on education is available, suggests that differences in earnings within firms change only marginally when accounting for differences in education between the two groups (Annex Figure 4.B.2). Immigrants also do not sort into lower-paying sectors and firms due to lower education. In contrast, the section below shows, for countries with available data, that a large share of the immigrant earnings gap is due to immigrants working in occupations that pay lower wages.

Working in lower paying occupations accounts for one‑quarter of the immigrant initial earnings gap

On average across countries with available occupational information, the concentration of immigrants in lower-paying occupations explains over one‑quarter (27%) of the immigrant earnings gap (Figure 4.7). These are occupations, where all workers, both foreign and native‑born, are paid on average lower wages. The role of occupations in explaining the immigrant earnings gap is particularly pronounced in Portugal and Germany, where over 30% of the initial entry earnings gap can be explained by work in lower-paying occupations relative to the native‑born.14 In contrast, in Finland, occupations explain only 17% of the immigrant earnings gap at entry.

Accounting for occupations in which immigrants work reduces the magnitude of the immigrant earnings gap within firms, indicating that immigrants work in lower-paying occupations within the same firms. The remaining earnings gap within firms after controlling for occupations reflects differences in tasks and responsibilities within occupations, differences in productivity in a given job (e.g. tenure, educational attainment, language skills) or differences in pay for work of equal value.15

Figure 4.7. Working in lower paying occupations accounts for one quarter of the immigrant initial earnings gap

Immigrant earnings gap at entry in the host country labour market, decomposed into the contribution of sectors, regions, firms (within sector and region) and a within-firm component, percentage

Note: The overall earnings gap corresponds to the average monthly earnings difference between immigrants upon entry in the host country and the native‑born workers of the same age and sex in the same year. The decomposition of the earnings gap follows Gelbach (2016[8]), see Box 4.1 for details. Occupation is measured at the 2‑digit level across countries considered. For Finland, occupational information is only available from 2004 onwards. * The earnings measure corresponds to annual earnings. In countries with annual earnings information, the earnings gap in the first year corresponds to that recorded in the second year to correct for differences in days worked and mechanically low earnings induced by the day at entry in the labour market.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details. Authors’ calculations.

Statlink

The concentration of immigrants in lower-paying occupations reflects the lower educational attainment of part of the immigrant population but also the widespread overqualification of tertiary educated immigrants. The immigrant gap in overqualification has been well documented across the OECD (OECD/European Commission, 2023[1]; Dustmann, Frattini and Preston, 2012[14]). Box 4.3 presents overqualification rates for immigrants and the native‑born from alternative data sources for the countries covered in the analysis.

Box 4.3. Immigrant overqualification is widespread across OECD countries

On average across countries, the share of overqualified immigrants, measured as the proportion of tertiary educated immigrants working in medium- and low-skilled occupations, is 27 p.p. higher than that of the native‑born (Table 4.1). The overqualification rate of immigrants is particularly large in Italy and Norway, as well as Portugal and Sweden. It is relatively lower in Austria and Germany as well as close to zero in Canada and the United States.

Table 4.1. Overqualification rates among recent immigrants

Share of (recently arrived) tertiary-educated workers in medium- to low-skilled occupations by country

Country	Foreign-born	Native‑born	Difference (FB – NB)
ITA	64	14	50
NOR	49	12	37
PRT	41	12	29
SWE	41	12	28
DNK	34	13	22
ESP	51	33	18
FRA	38	20	18
FIN	33	18	15
OECD	44	17	27
NLD	23	14	9
AUT	31	23	8
DEU	25	18	8
CAN*	36	34	2
USA*	35	35	0

Note: For European countries, the overqualification rate for immigrants is based on all tertiary educated immigrants arriving to the host country between 2006‑2015 as recorded in the EU-LFS. For the United States and Canada, denoted with an asterisk, the overqualification rate is measured based on all tertiary educated immigrants in the country in 2015 as recorded in the Database on Immigrants in OECD and non-OECD Countries (DIOC). For New Zealand, information on occupations is not available in the Labour Force Survey and as such overqualification cannot be measured.

Source: EU-LFS, DIOC, authors’ calculations.

Statlink

Several mechanisms may be at play for explaining the role of occupations and overqualification. Immigrants may face barriers to human capital transferability acquired in the origin country, for example due to obstacles in the formal recognition of their qualifications in the host country. Language barriers and occupational licensing may also prevent reallocation towards higher-paying occupations. There may also be potential mismatches between the demand for skills in the host country labour market and immigrants’ skills acquired in the origin country.

Immigrants partially close the earnings gap by moving to better paying firms and sectors over time

As immigrants gain experience in the host country labour market, they partially close the earnings gap with respect to the native‑born through job mobility towards better-paying firms and sectors (Figure 4.8). On average across countries, the immigrant earnings gap is reduced by about one‑third, or 13 p.p., from the initial year in the labour market.

Moving to higher paying firms accounts for one‑quarter (24%) of the reduction in the gap, moving to higher-paying sectors accounts for 13%. The larger reduction in the immigrant earnings gap through job mobility to higher-paying firms than to higher-paying industries reflects that transitions between sectors are more difficult since different economic activities may require different skills. Moving to better paying regions is not relevant for the reduction in the gap, accounting for only 1 p.p. As immigrants tend to enter the labour market in higher-paying regions relative to the native‑born, there is little room for further improvement in this dimension.

Figure 4.8. Immigrants partially close the earnings gap by moving to better paying firms and sectors

Change in immigrant earnings gap between the fifth and the first year in the host country labour market, decomposed into changes in the contribution of sectors, regions, firms (within sector and region) and a within-firm component, p.p.

Note: The change in the earnings gap between the first and the fifth year in p.p. can be decomposed as the sum of the change in each component (firm, sector, region, within-firm) between the first and the fifth year. * The earnings measure corresponds to annual earnings. In countries with annual earnings information, the earnings gap in the first year corresponds to that recorded in the second year to correct for differences in days worked and mechanically low earnings induced by the day at entry in the labour market. ** For Colombia, the earnings gap is estimated after four years, instead of five, due to the short panel structure.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details. Authors’ calculations.

Statlink

The mobility of immigrants to higher-paying firms is coupled with improvements in firm quality relative to the native‑born (Figure 4.9). Across all the countries considered, immigrants move to firms that not only pay higher wages, but also are more productive, larger in size and less segregated.

Figure 4.9. Immigrants climb up the firm quality ladder towards larger, higher paying, more productive and less segregated firms

Change in firm characteristics of immigrants relative to the native‑born between the fifth and the first year in the host country labour market

Note: The firm fixed-effect is a firm earnings premium paid to all workers in a given firm net of worker composition following Abowd, Kramarz and Margolis (1999[11]). Firm productivity is measured in terms value‑added per worker, except for Portugal, where it is measured by sales per worker. Firm size refers to the number of employees in a given firm and year. Firm segregation is a dummy variable equal to one if immigrants comprise more than 50% of all workers in a given firm and year. Information on firm-level productivity is not available in Austria, Colombia, Germany, Spain, France, the Netherlands and the United States. The share of immigrants in the firm is not available in Spain, Italy and the United States. For Spain, Italy and France, the firm fixed-effects do not control for worker composition due to small sample sizes.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details. Authors’ calculations.

Statlink

While job mobility towards better firms is an important margin for wage progression among immigrants, it is important to keep in mind that most of the reduction in the immigrant earnings gap happens within firms. On average across countries, more than half (61%) of the reduction in the earnings gap by the fifth year can be attributed to progression within firms. Mobility within firms is particularly relevant in Finland, Norway and Sweden, as well as the Netherlands and Spain but less so in Austria and Colombia. This reflects increases in hours worked (as seen above), returns to tenure and potentially transitions to higher-paying occupations. The next sub-section extends the analysis to include occupational mobility for the countries with available data to assess to what extent the reduction in the within-firm earnings gap reflects transitions to higher-paying occupations.

Immigrants do not close the earnings gap by moving to higher paying occupations despite widespread overqualification

Job mobility towards higher-paying occupations does not contribute to the reduction in the immigrant earnings gap in the countries considered in the analysis with available data on occupations. On average across the seven countries, occupations have a close to zero contribution to the overall reduction in the immigrant earnings gap in the first years in the labour market (Figure 4.10). This pattern is similar across countries with the contribution of occupations to the reduction of the earnings gap being small in all countries, and in most cases negative. This means that immigrants move to higher-paying occupations at a lower rate than the native‑born of the same age and sex.

Despite the large contribution of working in lower-paying occupations to the immigrant earnings gap, immigrants do not move up the occupational ladder in the countries in the analysis. More immigrants than native‑born have a low educational attainment which partly explains they work in lower skill, and lower pay, occupations. However, it is also the case that tertiary educated immigrants face high rates of overqualification upon arrival in the host country: they work in lower skill, and lower paying occupations than predicted by their educational attainment. In the case of tertiary educated immigrants, upwards occupational mobility (gaining access to occupations that match their educational credentials) could play a role in decreasing the immigrant earnings gap over time.

Figure 4.10. Immigrants do not close the earnings gap by moving to higher-paying occupations

Change in the immigrant earnings gap between the fifth and the first year in the host country labour market decomposed into the contributions of firms, sectors, regions, occupations and the within-firm and within-occupation component, p.p.

Note: The change in the earnings gap between the first and the fifth year in p.p. can be decomposed as the sum of the change in each component (firm, sector, region, occupation, within-firm and occupation) between the first and the fifth year. * The earnings measure corresponds to annual earnings. In countries with annual earnings information, the earnings gap in the first year corresponds to that recorded in the second year to correct for differences in days worked and mechanically low earnings induced by the day at entry in the labour market. Occupation is measured at the 2‑digit level across countries considered. For Finland, occupational information is only available from 2004 onwards.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details. Authors’ calculations.

Statlink

Importantly, the contribution of firms and sectors remains substantial for the reduction in the immigrant earnings gap after accounting for occupational mobility. This reinforces the conclusion that job mobility to higher-paying firms and sectors, as well as earnings growth within firms, are key to the earnings integration of immigrants.

The results in this chapter show that the immigrant earnings gap is quite persistent over time, despite improvements in the first years in the labour market (Annex Table 4.B.1). This raises the question of whether the gap persists across generations. Box 4.4 presents recent cross-country evidence that shows that the adult native‑born children of immigrants experience a much lower earnings gap than their parents. They still work in lower paying sectors and jobs than the native‑born, but to a much lower extent than the foreign-born.

Box 4.4. The integration of the native‑born children of immigrants: The role of firms

The adult native‑born children of immigrants are a large and growing group in many OECD countries (OECD/European Commission, 2023[1]). While the adult native‑born children of immigrants experience an earnings gap relative to the native‑born without a migration background, recent evidence shows that it is a substantially lower gap than that experienced by recent immigrants (Boustan, 2025[15]).

Hermansen et al (2025[16]) uses administrative data for selected OECD countries to document the earnings gap of immigrants, and of the children of immigrants, within and across industries and jobs (a job is an occupation within an establishment or firm). The study shows that for the six OECD countries with available data (Canada, Denmark, Germany, the Netherlands, Norway and Sweden), the adult children of immigrants experience a much lower earnings gap than immigrants. Moreover, the earnings of the adult children of immigrants are virtually the same as those of the native‑born working in the same sectors and jobs. Hence, the small differences in earnings between the adult children of immigrants and their native‑born counterparts are due to them working in lower paying sectors and jobs.

Differences in the immigrant earnings gap by region-of-origin and sex

The immigrant earnings gap is larger for immigrants from Africa, the Middle East and Asia partly because they work in lower paying sectors and firms

The outcomes of immigrants in the labour market differ by immigrants’ region of origin (see Chapter 2). The differences in outcomes are driven by differences in immigrants’ reason for migration, individual characteristics (e.g. educational level, quality of the education, experience in host country), immigrant networks, among others. This section presents the earnings gap over time in the labour market for the main regions of origin of immigrants in the 15 OECD countries and assesses the contribution of sectors and firms in explaining the gaps.

The immigrant earnings gap at labour market entry is larger for immigrants from Africa, the Middle East and Asia, and smaller for immigrants from the EU15 and North America and Oceania, on average across host countries (Figure 4.11). The earnings gap relative to the native‑born of the same age and sex is 44% for immigrants from Africa and the Middle East and 40% for immigrants from Asia, compared with 19% for immigrants from EU‑15 countries and 15% for immigrants from North America and Oceania.

Immigrants from all regions of origin work in lower-paying sectors and firms than the native‑born. The contribution of work in lower-paying firms and sectors is higher for those origin groups with the largest earnings gaps. The contribution of working in lower-paying sectors and firms is 18 and 12 p.p. for immigrants from Africa and the Middle East. For immigrants from the EU15, the contributions are 8 p.p. and 4 p.p. on average across host countries. The small group of immigrants from North America and Oceania work in similarly paying firms than the native‑born (despite working in lower paying sectors on average).

Over time in the labour market, the earnings gap decreases for immigrants from all regions of origin. However, the reduction in the earnings gaps is similar across regions of origin over five years, such that initial level differences between immigrant groups are persistent. Immigrants of all origin groups move to higher-paying firms over time. Differences in firm effects between origin groups are also persistent over time. See Annex Figure 4.B.3, Annex Figure 4.B.4 and Annex Figure 4.B.5 for gaps by region of origin and host country.

Figure 4.11. The earnings gap by region of origin

Note: Average across countries (unweighted). The overall earnings gap corresponds to the average monthly earnings difference between immigrants upon entry in the host country and the native‑born workers of the same age and sex in the same year. The decomposition of the earnings gap follows Gelbach (2016[8]), see Box 4.1 for details.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details. Authors’ calculations.

Statlink

The composition of immigrants, including the countries they come from, is partly shaped by migration policies. Immigrants arriving though different pathways (e.g. labour, family or humanitarian) have different labour market attachment and different labour market outcomes. Box 4.5 presents estimations of the immigrant earnings gap, as well as the contributions of sectors and firms to the earnings gap, for immigrants in Canada by broad category of migration. Canada is one of the few countries for which detailed data on immigrants’ visa status is available.

Furthermore, several recent academic papers have focussed on humanitarian migrants, or large migration waves, to study the role of firms in their integration in specific OECD countries. The results of this literature are summarised in Box 4.6.

Box 4.5. Earnings gaps and the role of firms by category of immigration – Evidence from Canada

Immigrants arriving in the host country through different migration pathways often fare differently in the host country labour market and may require different integration support. While in many cases, labour immigrants arrive with a job offer, family migrants or refugees need to first choose whether to enter the labour market and if they do so, look for a job.

In Canada, the administrative data contains information on the residence or work permit the immigrant holds at each point in time. The data is very detailed but for the purpose of this analysis, immigrants were grouped into six categories at the time they entered the Canadian labour market.1

Immigrants from all six categories at entry earn less than the native‑born of the same age and sex (Figure 4.12). Among permanent immigrants at entry, economic immigrants experience the lowest earnings gap, and the largest decrease in the gap, followed by immigrants for family reasons and refugees.

The large initial earnings gap (mainly within sectors and firms) experienced by international students mainly reflects the importance of part-time jobs that can be combined with study. International students have the largest reduction in the earnings gap as many transition to full-time jobs after graduating.

Temporary migrants experience the lowest earnings gap. This is a heterogeneous group that includes temporary foreign workers of both high- and low-skill streams, as well as participants in Canada’s International Mobility Program.

Immigrants from all categories of migration work in lower-paying sectors and firms upon entry in the labour market. The firm effect accounts for between 10 and 20 p.p. of the initial earnings gap. Over time in the labour market, immigrants from all categories of migration move to better paying firms.

Figure 4.12. The immigrant earnings gap by category of migration in Canada

Immigrant earnings gap at entry in the host country labour market, and change between the fifth and first year, by category of migration, decomposed into the contribution of sectors, regions, firms (within sector and region) and a within-firm component, percentage and p.p.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details. Authors’ calculations.

Statlink

1. There are four categories of permanent migrants (principal applicant of the economic class, dependent of the economic class, immigrant for family reasons and refugee) and two categories of temporary migrants (international students and temporary economic migrants).

Box 4.6. The integration of refugees: The role of firms

The integration of refugees in the labour market poses distinct challenges compared to that of economic migrants (Brell, Dustmann and Preston, 2020[17]). Due to the forced nature of refugee immigration, refugees typically experience substantial and highly persistent gaps in integration outcomes, even when compared to other migrants (Fasani, Frattini and Minale, 2021[18]; Müller, Pannatier and Viarengo, 2023[19]).

A growing body of research aims at understanding the role of firms for the economic integration of refugees:

For the Netherlands, Yumoto et al. (2023[20]) document that refugees earn only a third of the average hourly pay of the native‑born and concentration of refugees in low-paying firms explains 17% of the overall hourly earnings gap between 2014 and 2021. Assortative matching (i.e. the extent by which high-ability workers work for high-paying firms) is estimated to be negative for refugees, indicating that high-ability refugees tend to work in low-paying firms. Within firms, the authors find no pay-setting differences between refugees and the native‑born.

For Denmark, Caiumi and Simonsen (2025[21]) leverage the quasi-random assignment of refugees to municipalities and find that on average refugees achieve a significant improvement in workplace quality only after ten years since arrival. However, refugees that are placed in municipalities with a higher share of co-nationals employed by high-quality employers experience significant increases in employment probability and annual earnings, which highlights the role of social connections for integration.

For Colombia, Delgado-Prieto (2025[22]) finds that, despite massive regularisation programmes targeted at Venezuelan migrants, only 10% of regularised migrants had formal jobs. Conditional on entering the formal sector, about half of the immigrant earnings gap can be attributed to the concentration in low-paying firms, with persistent gaps with respect to the native‑born in most job and firm dimensions.

For Israel, Arellano-Bover and San (2022[23]) study the mass migration of nearly 1 million former Soviet Union (FSU) Jews during the 1990s fleeing from the Soviet Union. This historical episode offers a unique setting to study unconstrained integration, as FSU immigrants received citizenship on arrival to Israel. At labour market entry, FSU immigrants earned 57% less than comparable natives and about 20% of the earnings gap can be explained by work in low-paying firms. Over time, FSU immigrants close the earnings gap relative to the native‑born after 29 years in the host country, and this process is substantially mediated by job mobility towards higher-paying, more desirable employers as opposed to improvements in differential pay-setting within firms.

The immigrant earnings gap at entry is slightly larger for men than women and declines more slowly over time

On average across OECD countries, the immigrant earnings gap at labour market entry is slightly higher among male immigrants, with male immigrants experiencing an average 35% gap relative to native‑born men, compared to an average 33% gap for female immigrants relative to native‑born women (Figure 4.13). There is a larger earnings gap among male immigrants in most countries, except for Colombia, Germany, France, Denmark and Norway.

Despite the lower immigrant earnings gap among women, it is important to keep in mind that immigrant women face a double pay penalty in the labour market. They experience an earnings gap relative to men and relative to the native‑born.

Figure 4.13. The immigrant earnings gap is slightly larger among male immigrants at labour market entry

Note: * The earnings measure corresponds to annual earnings. In countries with annual earnings information, the earnings gap in the first year corresponds to that recorded in the second year to correct for differences in days worked and mechanically low earnings induced by the day at entry in the labour market. ** For Colombia, the earnings gap is estimated after four years, instead of five, due to the short panel structure. Darker dots represent the value at YSE = 5, lighter dots at YSE = 1.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details.

Statlink

Between the first and the fifth year in the host country, immigrant men reduce earnings gaps relative to the native‑born at a slower pace than women. On average across OECD countries, immigrant men close 12 p.p. of the immigrant earnings gap after five years, compared to a 14 p.p. reduction among immigrant women. The slower pace of convergence to the native‑born, together with the higher earnings gaps at entry, results in immigrant earnings gaps being larger among male immigrants also after five years. A slower speed of convergence among immigrant men has been found to be present across countries in the literature (Lee, Peri and Viarengo, 2022[24]).

The contribution of work in lower-paying firms to the immigrant earnings gap is higher for men than women (Figure 4.14). On average across OECD countries, work in low-pay firms accounts for 12 p.p. of the immigrant earnings gap at entry for men, compared with 10 for women. The firm earnings gap is larger for immigrant men than women in all countries in the analysis, except for New Zealand, France and the Netherlands where firm earnings gaps are similar for men and women. There is no clear pattern on whether mobility towards higher-paying firms is more relevant for male than for female immigrants.

Figure 4.14. Work in lower-paying firms explains more of the immigrant earnings gap for male immigrants

Contribution of firms to the immigrant earnings gap at labour market entry, for immigrant men and women, p.p.

Note: Sample are all immigrants at labour market entry. Points above the red 45‑degree line imply a smaller contribution of firms to the immigrant-native earnings gap for female immigrants.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details.

Statlink

Conclusion

This chapter provides comprehensive cross-country evidence on the sources of the immigrant earnings gap at entry, and over time, in the host country labour market, based on linked employer-employee panel data for 15 OECD countries.

Its main findings can be summarised as follows. First, immigrants, when first entering the host country labour market, earn significantly less than native‑born workers of the same age and sex, but partially catch up as they become more integrated. On average across countries, the immigrant earnings gap declines from 34% when first entering the host country labour market to 21% after five years. Second, immigrants earn less because they are initially employed in lower-paying firms, industries and occupations. Third, over time, immigrants move to higher paying firms and industries, but they do not move to higher paying occupations. One‑quarter of the reduction in the earnings gap is due to moving to better paying firms within sectors, on average across host countries. Fourth, there are large differences in the immigrant earnings gaps at entry and their evolution over time by region-of-origin.

The main message for integration policy is that while measures that focus on the recognition of immigrant credentials or the remediation of skills gaps are crucial, they should be complemented with measures that support job search and job mobility. One of the most striking findings of the chapter is that occupational segregation is highly persistent. This highlights the importance of policies related to the recognition of prior qualifications and remediation of skill gaps related to language fluency and professional licensing requirements. However, the chapter also shows that job mobility to better paying firms and sectors can play an important role in the earnings integration of immigrants. This suggests that there is scope for policies that target barriers to job mobility including providing information on job search and the host country labour market, career counselling and the development of professional networks, improving local transportation, combatting discrimination in the housing market and providing affordable housing, among others.

References

[11] 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.

[4] Abramitzky, R., L. Boustan and K. Eriksson (2014), “A nation of immigrants: Assimilation and economic outcomes in the age of mass migration”, Journal of Political Economy, Vol. 122/3, https://doi.org/10.1086/675805.

[33] Andersson, F. et al. (2014), “Workplace Concentration of Immigrants”, Demography, Vol. 51/6, pp. 2281-2306, https://doi.org/10.1007/s13524-014-0352-3.

[28] Ansala, L., O. Åslund and M. Sarvimäki (2022), “Immigration history, entry jobs and the labor market integration of immigrants”, Journal of Economic Geography, https://academic.oup.com/joeg/article-abstract/22/3/581/6246156 (accessed on 6 June 2023).

[23] Arellano-Bover, J. and S. San (2022), “The role of firms in the assimilation of immigrants”, papers.ssrn.com, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3594778 (accessed on 6 March 2023).

[30] Aslund, O. et al. (2022), “Firm Productivity and Immigrant-Native Earnings Disparity”, SSRN Electronic Journal, https://doi.org/10.2139/ssrn.4114485.

[34] Åslund, O. and O. Skans (2010), “Will I See You at Work? Ethnic Workplace Segregation in Sweden, 1985–2002”, ILR Review, Vol. 63/3, pp. 471-493, https://doi.org/10.1177/001979391006300306.

[35] Aydemir, A. and M. Skuterud (2008), “The Immigrant Wage Differential within and across Establishments”, ILR Review, Vol. 61/3, pp. 334-352, https://doi.org/10.1177/001979390806100304.

[36] Barth, E., B. Bratsberg and O. Raaum (2012), “Immigrant wage profiles within and between establishments”, Labour Economics, Vol. 19/4, pp. 541-556, https://doi.org/10.1016/j.labeco.2012.05.009.

[15] Boustan, L. (2025), Intergenerational Mobility of Immigrants in 15 Destination Countries, National Bureau of Economic Research.

[17] Brell, C., C. Dustmann and I. Preston (2020), “The Labor Market Integration of Refugee Migrants in High-Income Countries”, Journal of Economic Perspectives, Vol. 34/1, pp. 94-121, https://doi.org/10.1257/jep.34.1.94.

[21] Caiumi, A. and E. Simonsen (2025), Employer Quality, Social Networks, and Economic Integration of Refugees.

[5] Card, D., J. Heining and P. Kline (2013), “Workplace Heterogeneity and the Rise of West German Wage Inequality*”, The Quarterly Journal of Economics, Vol. 128/3, pp. 967-1015, https://doi.org/10.1093/qje/qjt006.

[10] 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.

[9] 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.

[26] Carneiro, A., N. Fortuna and J. Varejão (2012), “Immigrants at new destinations: How they fare and why”, Journal of Population Economics, Vol. 25/3, pp. 1165-1185, https://doi.org/10.1007/S00148-011-0387-3.

[2] Chiswick, B. (1978), “The Effect of Americanization on the Earnings of Foreign-born Men”, Journal of Political Economy, Vol. 86/5, https://doi.org/10.1086/260717.

[37] Damas de Matos, A. (2016), “Firm heterogeneity and immigrant wage assimilation”, Applied Economics Letters, Vol. 24/9, pp. 653-657, https://doi.org/10.1080/13504851.2016.1218421.

[22] Delgado-Prieto, L. (2025), Firms, Regularization, and Immigrant Integration.

[14] Dustmann, C., T. Frattini and I. Preston (2012), “The Effect of Immigration along the Distribution of Wages”, The Review of Economic Studies, Vol. 80/1, pp. 145-173, https://doi.org/10.1093/restud/rds019.

[29] Eliasson, T. (2013), “Decomposing immigrant wage assimilation: The role of workplaces and occupations”, https://www.econstor.eu/handle/10419/82263 (accessed on 6 March 2023).

[18] Fasani, F., T. Frattini and L. Minale (2021), “(The Struggle for) Refugee integration into the labour market: evidence from Europe”, Journal of Economic Geography, Vol. 22/2, pp. 351-393, https://doi.org/10.1093/jeg/lbab011.

[8] Gelbach, J. (2016), “When do covariates matter? And which ones, and how much?”, Journal of Labor Economics, Vol. 34/2, pp. 509-543, https://doi.org/10.1086/683668.

[27] Glitz, A. (2014), “Ethnic segregation in Germany”, Labour Economics, Vol. 29, pp. 28-40, https://doi.org/10.1016/J.LABECO.2014.04.012.

[31] Gorshkov, A. (2023), Job Ladders and Labour Market Assimilation of Immigrants, Elsevier BV, https://doi.org/10.2139/ssrn.4342044.

[16] Hermansen, A. et al. (2025), “Immigrant–native pay gap driven by lack of access to high-paying jobs”, Nature, https://doi.org/10.1038/s41586-025-09259-6.

[13] Kline, P. (2024), “Firm wage effects”, in Handbook of Labor Economics, Elsevier, https://doi.org/10.1016/bs.heslab.2024.11.005.

[24] Lee, T., G. Peri and M. Viarengo (2022), “The gender aspect of migrants’ assimilation in Europe”, Labour Economics, Vol. 78, p. 102180, https://doi.org/10.1016/j.labeco.2022.102180.

[3] Lubotsky, D. (2007), “Chutes or Ladders? A Longitudinal Analysis of Immigrant Earnings”, Journal of Political Economy, Vol. 115/5, pp. 820-867, https://doi.org/10.1086/522871.

[19] Müller, T., P. Pannatier and M. Viarengo (2023), “Labor market integration, local conditions and inequalities: Evidence from refugees in Switzerland”, World Development, Vol. 170, p. 106288, https://doi.org/10.1016/j.worlddev.2023.106288.

[7] OECD (2022), OECD Employment Outlook 2022: Building Back More Inclusive Labour Markets, OECD Publishing, Paris, https://doi.org/10.1787/1bb305a6-en.

[12] OECD (2020), International Migration Outlook 2020, OECD Publishing, Paris, https://doi.org/10.1787/ec98f531-en.

[32] OECD (2018), International Migration Outlook 2018, OECD Publishing, Paris, https://doi.org/10.1787/migr_outlook-2018-en.

[1] OECD/European Commission (2023), Indicators of Immigrant Integration 2023: Settling In, OECD Publishing, Paris, https://doi.org/10.1787/1d5020a6-en.

[25] Raposo, P., P. Portugal and A. Carneiro (2019), “The Sources of the Wage Losses of Displaced Workers”, Journal of Human Resources, Vol. 56/3, pp. 786-820, https://doi.org/10.3368/jhr.56.3.0317-8667r3.

[6] Song, J. et al. (2019), “Firming up inequality”, Quarterly Journal of Economics, Vol. 134/1, pp. 1-50, https://doi.org/10.1093/QJE/QJY025.

[20] Yumoto, H. et al. (2023), “The role of firms in the economic assimilation of refugees”, International Journal of Population Data Science, Vol. 8/2, https://doi.org/10.23889/ijpds.v8i2.2335.

Annex 4.A. Data and methodology

Annex Table 4.A.1. Linked employer-employee datasets by country

Country	Name	Source	Sample	Period	Cohorts included
Austria	AMS-BMASK Arbeitsmarktdatenbank	Social security administration	Universe	2000‑2019	2003‑2015
Canada	Canadian Employer-Employee Dynamics Database	Tax administration	Universe	2001‑2019	2004‑2015
Colombia	Planilla Integrada de Liquidación de Aportes (PILA)	Social Security administration	Universe (formal sector)	2017‑2021	2018‑2020
Denmark	Integrerede Database for Arbejdsmarkedsforskning (IDA) and other data from Statistics Denmark	Tax administration	Universe	2000‑2019	2003‑2015
Finland	FOLK employment data from Statistics Finland, Employer Payroll Report from Tax Admin.	Tax administration	Universe	2000‑2019	2003‑2015
France	Panel DADS	Social security administration	8.5% random sample of workers	2002‑2019	2005‑2015
Germany	Sample of Integrated Employer-Employee Data (SIEED)	Social security administration	1.5% sample of establishments	2000‑2019	2003‑2015
Italy	Italian Social Security Dataset (INPS)	Social security administration	6.6% random sample of workers	2000‑2019	2003‑2015
Netherlands	CBS Microdata from Statistics Netherlands	Tax administration	Universe	2006‑2019	2009‑2015
New Zealand	Integrated Data Infrastructure (IDI)	Tax administration	Universe	2000‑2019	2003‑2015
Norway	Arbeidsgiver- og arbeidstakerregister (Aa-registeret), Lønns- og trekkoppgaveregisteret (LTO)	Tax administration	Universe	2000‑2019	2003‑2015
Portugal	Quadros de Pessoal	Mandatory employer survey	Universe	2002‑2019	2005‑2015
Spain	Muestra Continua de Vidas Laborales con Datos Fiscales (MCVL-CDF)	Social security and tax administration	4% random sample of workers	2006‑2019	2009‑2015
Sweden	Longitudinell integrationsdatabas för sjukförsäkrings- och arbetsmarknadsstudier (LISA), Företagens ekonomi (FEK), Jobbregistret (JOBB)	Social security administration	Universe	2002‑2018	2005‑2014
United States	U.S. Census Bureau Numerical Identification file (Numident), Internal Revenue Service (IRS) individual W‑2 forms, Longitudinal Business Database (LBD), American Community Survey (ACS)	Social security and tax administration, survey data	2.5% sample of workers (addresses)	2001‑2019	2004‑2015

Gelbach decomposition of the immigrant earnings gap

The following decomposition is an application of Gelbach (2016[8]) and similarly applied in (Raposo, Portugal and Carneiro (2019[25])A Gelbach decomposition allows to unambiguously determine the contribution of covariates for the change of a coefficient estimate. The benchmark integration specification can be written in matrix notation as follows:

$Y=X{\beta }_0+YSE{\delta }_0+u_0$

Where X is a matrix of control variables (sex dummy interacted with a quartic in age, calendar year effects). $YSM$ is a matrix collecting years-since‑migration dummies and ${\delta }_0$ a vector of corresponding regression coefficients for each ysm dummy. $u_0$ stands for the error term. The subscript 0 denotes the benchmark specification.

Based on the Frisch-Waugh-Lovell theorem, the least squares estimate of ${\delta }_0$ is given by running a regression of Y on YSE, after netting out the effect of the control variables in X from both Y and YSE. This is:16

$\widehat{{\delta }_0}={\left(YS{E}^{\text{'}}{R}_{X}YSE\right)}^{-1}YS{E}^{\text{'}}{R}_{X}Y=A_{X}Y$

Where $R_X=[I-X{\left(X^{\text{'}}X\right)}^{-1}X\text{'}]$ is the residual-maker that, when multiplied with any variable, generates the residuals from running a regression of that variable on X. In this case, they allow to net out the impact of X on both Y and YSE. For the sake of simplicity, define the matrix $A_X={\left(YS{E}^{\text{'}}{R}_{X}YSE\right)}^{-1}YS{E}^{\text{'}}{R}_X$ as matrix that, when multiplied with Y, generates coefficients for ${\delta }_0$ after netting out the effects from all control variables included in X as described above.

In a second step, we expand the benchmark specification to include occupation and firm dummies to account for the impact of sorting. This results in the full specification, expressed as:

$Y=X{\beta }_1+YSE{\delta }_1+F{\psi }_1+O{\gamma }_1+u_1$

Where ${\psi }_1$ and ${\gamma }_1$ denote the firm and occupation fixed-effects respectively. The difference between ${\delta }_0$ and ${\delta }_1$ is that the first is biased due to the omission of firms and occupations. Multiply both sides of the full model with $A_X$ and bring ${\delta }_1$ to the left-hand side to obtain:17

${\delta }_0-{\delta }_1=A_X F{\psi }_1+A_{X}O{\gamma }_1={\tau }_{\psi }+{\tau }_{\gamma }$

Where ${\tau }_{\psi }$ is the overall contribution of firm sorting to the integration of immigrants and ${\tau }_{\gamma }$ the corresponding contribution of occupational sorting. In practice, we can obtain the overall contribution of firms ${\tau }_{\psi }$ by a regression of estimated firm fixed-effects from the full specification on the covariates of the base model (YSM and X). Analogously, we can obtain the overall contribution of occupations ${\tau }_{\gamma }$.

Disentangling regional and sectoral sorting from firm sorting

In principle, the overall contribution of firms (${\tau }_{\psi })$ absorbs wage differentials between regions and between industries (Card, Rothstein and Yi, 2025[10]; Card, Rothstein and Yi, 2024[9]).This is because the sector and workplace region are time‑invariant firm characteristics, which are absorbed by the introduction of firm fixed-effects. Consequently, we decompose the overall contribution of firms (${\tau }_{\psi }$) into separate components capturing wage differentials (1) between regions (2) between sectors and (3) between firms within sectors and regions.

In order to disentangle regions and sectors from firms, we assume the firm fixed-effect to be orthogonal to sector and region effects, such that we can write firm fixed-effects as a function of region fixed-effects ($\theta )$, sector fixed-effects ($\mathrm{\chi })$ and a residual term containing the impact of firms after netting out the effects of regions and sectors ($v$):18

$F{\psi }_1=R\theta +S\mathrm{\chi }+v$

Multiply both sides of this expression with $A_X$ to obtain:

${\tau }_{\psi }={\tau }_{\theta }+{\tau }_{\mathrm{\chi }}+{\tau }_{\mathrm{v}}$

The gross contribution of firms to the integration of immigrants ${\tau }_{\psi }$ can be thus separated into the contribution of regions, sectors, and the net contribution of firms. In practice, we regress firm fixed effects from the full specification on 3‑digit industry dummies, NUTS‑2 region dummies and a residual term, which allows us to obtain estimates of these three terms. Finally, we obtain estimates of ${\tau }_{\theta },{\tau }_{\mathrm{\chi }}$ and ${\tau }_{\mathrm{v}}$ (i.e. the contribution of regions, sectors and firms) from separate regressions using the estimated sector, region fixed-effects and residuals as dependent variables and regressing them on the covariates of the base model YSE and X.

Annex 4.B. Additional tables and figures

Annex Table 4.B.1. Counts, age and share of women among new immigrants in the labour market

Average age, share of women and count of immigrants in the first year in the host country labour market

Country	Average age	Share of women	Number of immigrant workers
AUT	31.0	0.41	1 193 306
CAN	32.4	0.48	2 928 420
COL	31.6	0.68	118 488
DNK	29.8	0.54	224 744
ESP	31.8	0.45	20 646
FIN	32.3	0.42	123 325
FRA	33.0	0.47	118 302
ITA	31.6	0.34	265 270
NLD	30.6	0.47	553 720
NOR	31.6	0.41	210 266
NZL	33.9	0.34	90 663
PRT	32.0	0.43	228 540
SWE	34.4	0.35	82 332
DEU	30.7	0.44	176 856
USA		0.47	842 000
OECD	31.9	0.45	7 176 878

Note: Sample refers to all immigrants entering the labour market between 2006‑2015.

Source: National linked employer-employee data, see Table 4.A.1 for details. Authors’ calculations.

Statlink

Annex Figure 4.B.1. Immigrant earnings gaps after ten years

Immigrant earnings gap in the first, fifth and tenth year after entry in the host country labour market, percentage

Note: * The earnings measure corresponds to annual earnings. In countries with annual earnings information, the earnings gap in the first year corresponds to that recorded in the second year to correct for differences in days worked and mechanically low earnings induced by the day of entry in the labour market.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details. Authors’ calculations.

Statlink

Annex Table 4.B.2. Immigrant earnings over time in the labour market

Average real earnings (in USD) by years-since‑entry in the host country labour market

Country	Monthly earnings in USD
	YSE = 1	YSE = 5	Percentage change
AUT	885	1 068	21
COL	242	303	25
DEU	1 417	2 102	48
DNK	3 161	3 893	23
ESP	1 192	1 517	27
FIN	2 156	3 067	42
FRA	1 578	1 880	19
ITA	1 241	1 603	29
NLD	2 216	2 072	‑6
NZL	4 010	3 786	‑6
PRT	775	904	17
SWE	1 328	1 909	44
	Annual earnings in USD
CAN	29 557	37 618	27
NOR	59 752	67 877	14
USA	27 375	39 163	43

Note: Real monthly earnings in each currency are converted into USD using average annual exchange rates from OECD.Stat for the period 2003‑2019 (Euro area, Sweden, Denmark, New Zealand, Norway), 2004‑2019 (Canada), 2017‑2021 (Colombia).

Source: Linked employer-employee data, OECD.Stat, authors’ calculations.

Statlink

Annex Table 4.B.3. Educational attainment of recent immigrants

Share of high- and low-educated foreign-born and native‑born workers, percentage

Country	High-educated		Low-educated
	Foreign-born	Native‑born	Foreign-born	Native‑born
CAN*	64	54	15	19
DNK	52	28	21	30
NZL*	45	24	15	35
SWE	43	29	37	23
USA*	39	35	26	14
FRA	37	27	32	29
OECD	35	28	31	30
NOR	35	32	20	25
NLD	33	29	31	30
DEU	33	23	28	17
AUT	32	18	20	21
ESP	27	31	43	47
FIN	22	32	36	23
PRT	18	15	50	66
ITA	12	14	55	45

Note: High-educated individuals are those who completed a tertiary education degree. Low-educated individuals are those who did not complete upper secondary education. For European countries, immigrant education is measured in the year of arrival to the host country for the cohorts arriving in the years 2006‑2015. For Canada, New Zealand and the United States (denoted with an asterisk), the education shares are calculated based on DIOC, with reference year 2015 and based on immigrants entering the country within the past 10 years. For native‑born education shares are calculated as an average across all native‑born individuals. Countries are ranked based on share of highly educated immigrants.

Source: EU-LFS, DIOC, Authors’ calculations.

Statlink

Annex Table 4.B.4. Distribution of foreign- and native‑born across industries

Share of foreign-born and native‑born workers by 1‑digit industry, percentage

Industry (1‑digit)	Foreign-born	Native‑born	Difference (Foreign-born – Native‑born)
Accommodation and Food	16.5	5.2	11.3
Administrative Service Activities	17.2	6.5	10.7
Agriculture	4.6	2.9	1.7
Construction	8.6	7.6	1.1
Other Service Activities	2.9	2.7	0.2
Arts, Entertainment and Recreation	1.6	1.4	0.1
Real Estate Activities	0.8	0.9	‑0.2
Mining and Quarrying	0.3	0.5	‑0.2
Information and Communication	2.8	3.0	‑0.2
Water Supply	0.2	0.5	‑0.3
Electricity, Gas and Steam	0.1	0.5	‑0.4
Professional Activities	4.9	5.6	‑0.7
Transportation and Storage	4.3	5.3	‑1.0
Education	4.2	5.3	‑1.2
Financial and Insurance Activities	1.1	3.2	‑2.2
Wholesale and Retail Trade	11.5	15.0	‑3.5
Health	6.3	10.0	‑3.7
Public Administration and Defence	3.6	9.3	‑5.7
Manufacturing	8.6	14.4	‑5.8

Note: The foreign and native‑born columns refer to the share of native‑born (foreign-born) working in an industry relative to all native‑born (foreign-born) workers. For immigrant, the distribution is measured at labour market entry. Average across Austria, Canada, Denmark, Spain, Finland, France, Italy, the Netherlands, Norway, New Zealand, Portugal, Sweden and the United States.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details. Authors’ calculations.

Statlink

Annex Figure 4.B.2. Educational attainment does not explain immigrants working in lower-paying sectors and firms

Immigrant earnings gap in the year at entry in the host country labour market and change of the earnings gap between fifth and first year, decomposed into the contribution of sectors, regions, firms (within sector and region) and a within-firm component, adding and omitting educational controls, percentage and p.p.

Note: The overall earnings gap corresponds to the average monthly earnings difference between immigrants upon entry in the host country and the native‑born workers of the same age and sex in the same year. The decomposition of the earnings gap follows Gelbach (2016[8]), see Box 4.1 for details. The change in the earnings gap between the first and the fifth year in p.p. can be decomposed as the sum of the change in each component (firm, sector, region, within-firm) between the first and the fifth year. For Germany and Portugal, dummies for broad educational groups (no vocational training, vocational training and university degree) are introduced in all specifications to control for education, and the resulting estimates contrasted with the case when omitting educational controls.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details. Authors’ calculations.

Statlink

Annex Figure 4.B.3. Earnings gaps at entry in the labour market, by region of origin

Immigrant earnings gap at entry in the host country labour market by region of origin, percentage

Note: To report relevant group sizes, wage gaps for each country are reported if the region of origin amounts to at least 10% of the total migrant inflow at entry. The OECD average is an unweighted average by region of origin across all countries. Colombia is not included in the chart as the data identifies only Venezuelans.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details.

Statlink

Annex Figure 4.B.4. Immigrant earnings gaps and contribution of firms at entry in the host country labour market

Note: Each observation represents host country x region of origin. The earnings gap at entry, the contribution of sectors, and the contribution of firms (within of sectors and regions) are regressed on country dummies first such that each dot can be interpreted as the deviation of a given region of origin with respect to the average immigrant earnings gap at entry or average contribution of firms within a country. Dotted lines indicate areas above‑below the national average.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details.

Statlink

Annex Figure 4.B.5. Immigrant earnings gaps in the first and fifth year in the host country labour market, by region of origin

Note: Each observation represents host country x region of origin. Red line represents 45‑degree line.

Source: National linked employer-employee data, see Annex Table 4.A.1 for details.

Statlink

Notes

← 1.  The countries covered are Austria, Canada, Colombia, Denmark, Finland, France, Germany, Italy, Portugal, New Zealand, the Netherlands, Norway, Spain, Sweden and the United States.

← 2.  For the exact cohorts covered in each country, see Annex Table A.4.1. The analysis stops in 2019 to avoid the potential confounding effects of the COVID‑19 pandemic on labour markets.

← 3.  The countries with available information on occupations are Denmark, France, Germany, Italy, the Netherlands and Portugal.

← 4.  The data covers formal employment. Note that immigrants in formal employment may in some cases have an irregular migration status. This may be the case if their visa or work permit has expired (See on this issue Chapter 4. Addressing the illegal employment of foreign workers in (OECD, 2018[32])). Irregular migrants working in the informal sector are not covered in the data.

← 5.  For Colombia, immigrants are identified based on the regularisation documentation that was issued to Venezuelan refugees, who represent the vast majority of immigrants in the country.

← 6.  If an individual is absent from the labour market for more than three years, due to inactivity, unemployment, or parental leave for example, he/she will be considered as a new entrant in the labour market. The threshold of three years allows maximising the number of immigrant cohorts included in the analysis while minimising the cases in which immigrants are misclassified as new arrivals.

← 7.  Only annual earnings are available for Canada, Norway and the United States, instead of monthly earnings as in the other 12 countries in the analysis. Immigrants enter the host country labour market throughout the calendar year. Hence, their earnings in the first year in the labour market are mechanically low relative to the native‑born, as they have worked on average fewer months during the year. For these 3 countries, in the analysis, the earnings upon arrival are the earnings in the second calendar year.

← 8.  For Colombia, earnings are not available in the fifth year in the labour market, given that the migration wave of Venezuelans considered in the analysis is recent. The fourth year is used instead in the analysis.

← 9.  There is no data for the earnings of immigrants after 10 years in Colombia and the United States. The average takes into account the remaining 13 OECD countries.

← 10.  The initial immigrant earnings gap decreases from 43% to 33% in Germany and from 29% to 22% in Portugal.

← 11.  The decrease in the immigrant earnings gap after five years changes from 11.6 to 13.5 p.p. when controlling for education in Germany, and from 5.7 to 5.8 p.p. in Portugal.

← 12.  This is the effect of lower firm pay within 3‑digit industries.

← 13.  Several country studies provide evidence of immigrant workplace segregation across the OECD, such as Åslund and Skans (2010[34]) for Sweden, Carneiro, Fortuna and Varejão (2012[26]) for Portugal, Andersson et al. (2014[33]) for the United States, Glitz (2014[27]) for Germany, and (Ansala, Åslund and Sarvimäki (2022[28]) for Finland. There is further evidence that the firms immigrants concentrate in are disproportionately lower-wage firms, e.g. Ayemir and Skuterud (2008[35]) for Canada, Barth, Bratsberg and Raaum (2012[36]) for Norway, Eliasson (2013[29]), Aslund et al. (2022[30]) for Sweden, Damas de Matos (2016[37]) for Portugal, Arellano-Bover and San (2022[23]) for Israel and Gorshkov (2023[31]) for Denmark.

← 14.  For example, in Portugal, at labour market entry, over 70% of immigrants work in lower-paying occupations such as elementary occupations (34%), service and sales occupations (25%) and craft and related trades occupations (15%).

← 15.  The analysis controls for 2‑digit occupations, which is the most detailed level available across all countries. Results for Germany using 5‑digit occupations (which account for the hierarchical position within the firm) results in little changes in the contribution of occupations. The limited role of occupational mobility for earnings progression also remains unaltered (results available on request).

← 16.  Recall that the OLS estimator of $Y=X\beta +ϵ$in matrix form is given by $\beta ={\left(X^{\text{'}}X\right)}^{-1}X\text{'}Y$.

← 17.  Note that $A_{X}Y={\delta }_0$, $A_{X}X=A_X{u}_1=0$and $A_{X}YSM=I$

← 18.  Conceptually, this is similar to the approach in (Card, Rothstein and Yi, 2024[9]) and (Card, Rothstein and Yi, 2025[10]) to characterise sectoral and regional wage differentials. This is because the estimates for the regional and sectoral dummies in the regression capture (employment-weighted) average firm wage premia at the regional and sectoral level respectively.