What else? Immigrant–native cohorts entering the labour market under a context of adverse shocks: the great recession and COVID-19

Abstract

Purpose

This paper aims to analyse the short- and long-term effects of entering the Spanish labour market under tough economic conditions on young immigrant–native earnings and employment outcomes.

Design/methodology/approach

The authors use cohorts, where the entry cohort into the labour market is the unit of observation. As a database, the authors use the continuous sample of working histories covering the period 2007–2021. Then, the authors estimate the model using weighted least squares.

Findings

The results show that the great recession and COVID-19 led to a blockage at the entrance of the labour market, reducing the number of workers. Additionally, the authors observe an adverse impact in terms of employment and earnings on those entering the labour market. Besides, this effect varies in intensity and persistence for natives and immigrants, as well as by country of birth, age of entry, gender and educational level.

Originality/value

A contribution to the literature is the analysis of the earnings and employment trajectories of young people entering the Spanish labour market for the first time during an adverse shock, such as the 2008 economic crisis or the COVID-19 crisis, and the possible differences that exist between native and immigrant workers. So, the authors analyse the labour market trajectories of workers covering the most recent years. Likewise, the authors carry out an extensive heterogeneity analysis in which they distinguish workers by educational level, gender, age of entry into the labour market and immigrants by their country of birth. This represents an additional contribution. The use of a cohort approach also contributes to the existing literature.

Citation

Escalonilla, M., Cueto, B. and Perez-Villadoniga, M.J. (2024), "What else? Immigrant–native cohorts entering the labour market under a context of adverse shocks: the great recession and COVID-19", Applied Economic Analysis, Vol. 32 No. 96, pp. 207-231. https://doi.org/10.1108/AEA-07-2023-0275

1. Introduction

The labour market integration of immigrants is crucial for economic and social cohesion in host countries. Research extensively examines immigrant assimilation, its impact on native populations' earnings and employment, and differences between groups (Izquierdo et al., 2009; Peri, 2014; Edo, 2016). Economic conditions disproportionately affect immigrants, particularly during downturns (Xu, 2018; Orrenius and Zavodny, 2010; Gálvez-Iniesta, 2022). Immigrants and young workers, often in low-paying, temporary jobs, suffer most during economic crises, facing job losses and skills depreciation (OECD, 2009; Marcus and Gavrilovic, 2010). The great recession exacerbated these challenges, significantly impacting immigrant labour market insertion (Tilly, 2011; Fellini, 2018). In Spain, once a top destination for immigrants (Domínguez-Mujica et al., 2014; Eurostat), the recession led to a sharp decline in job opportunities, particularly in construction and services sectors, resulting in a surge in immigrant unemployment (Sanromá et al., 2009; Gonzalez and Ortega, 2010; Stanek and Veira, 2012; Carrasco et al., 2019; Bentolila et al., 2012a, 2012b, Bernardi et al., 2011; Papademetriou and Terrazas, 2009).

The Spanish case also turns out to be of great interest because the effect of the great recession has been particularly significant among the youngest workers (Bentolila et al., 2012a, 2012b; Dolado et al., 2013). Youth unemployment rates were above 40%–50% during the crisis and continued to be high when the economy started to recover. Together with a greater vulnerability to unemployment, this group is characterized by having precarious working conditions, i.e. a high temporary employment rate and part-time employment, which translates into job turnover and lower earnings.

Although there is evidence of the short-term impact of an adverse shock on the immigrant population and its differences with natives, little is known about the long-term effects. Recently, the global economic downturn because of the COVID-19 pandemic has reignited a discussion on the aftermath of a deep recession on first-time labour market entrants because of its huge shocks to the labour market. Therefore, the objective of this paper is to analyse to what extent different adverse shocks, such as the great recession and COVID-19 affected the earnings and employment of young immigrant workers who entered the Spanish labour market during the period 2007–2021 and their persistence over time.

This paper contributes to the growing literature covering an extensive period, 2007–2021, in which two different crises emerged in Spain: a financial crisis in 2008 and the COVID-19 crisis. Thus, we do not only examine the impact on earnings and employment of natives and immigrants of the 2008 economic crisis, but also the impact of the COVID-19 health crisis. Likewise, another contribution to the literature is that we include an extensive heterogeneity analysis. Specifically, we carry out the analysis strategy by distinguishing immigrants based on their country of birth or, more specifically, by geographical areas. As shown by the “Padron Continuo” prepared by the National Institute of Statistics of Spain, on average 33.1% of immigrants residing in Spain between 2017 and 2022 were born in South America, followed by 22.6% and 18.1% who come from the European Union and Africa, respectively. In this way, we may find differences within the immigrant group. We extend our findings by further splitting our sample of natives and immigrants by educational level, gender and age range of entry into the labour market.

The rest of this paper is organized as follows. Section 2 summarizes the literature review. Section 3 describes the data. Section 4 describes the empirical strategy. Section 5 presents the results, and Section 6 presents the heterogeneity analysis. Section 7 presents a discussion of results, and Section 8 briefly concludes.

2. Literature review

Economic conditions can affect the labour market outcomes of natives and non-natives in different ways. In terms of employment, evidence shows that immigrants are particularly affected by job loss during a recession because of their great sensitiveness to business cycles (see Bratsberg et al., 2018; Dustmann et al., 2010; Carrasco and García‐Pérez, 2015, Arcarons and Muñoz-Comet, 2018).

The analysis by Orrenius and Zavodny (2010) about employment and unemployment rates over the past 15 years in the USA reveals that immigrants’ labour market outcomes are more cyclical than those of natives. Also, for the USA, Hoynes et al. (2012) find that the impacts of the great recession have been felt most strongly by immigrants, young and low-educated workers. In the case of Spain, Motellón and López-Bazo (2015) show that the rate of job loss for immigrant males in the first quarter of 2012 was around 23.2%, while that for natives was 9.5%. Following the Spanish case, Mooi-Reci and Muñoz-Comet (2016) find that during the great recession, the likelihood of job loss was disproportionally higher among immigrants than natives. Factors such as occupational level and temporary employment contracts account for a major part of the difference in job losses between them.

In terms of earnings, adverse shocks affect labour markets, denying employment to job seekers and reducing earnings for the employed (Fornaro and Wolf, 2020; Guerrieri et al., 2022), especially for young workers, who have little job-specific human capital and thus may be relatively less valuable to their firms (Forsythe, 2022). There is also evidence of differences in the sensitivity of income to the business cycle between immigrant and native workers. Barth et al. (2006) find that immigrants’ hourly earnings in the USA are more sensitive than natives’ earnings to changes in state-level unemployment rates. A similar pattern has been found in Norway (Barth et al., 2004). On the other hand, Murillo-Huertas and Simón (2017) examine the relative wages of immigrants in Spain, with a particular focus on the impact of the great recession. Their results show that the great recession had a noticeable impact on the relative wages of immigrants, given that the significant increase in the native–immigrant wage gap observed during the previous expansionary period was mitigated during the economic downturn because of composition effects arising from the severe employment destruction pattern.

This greater vulnerability of immigrants during an economic downturn is explained by the differences between immigrant and native-born populations in terms of human capital, age, the sectors where they are employed, and the types of jobs they hold (Xu, 2018).

First, extensive literature shows that immigrants tend to accumulate less human capital than natives, pushing them to the bottom of the job ladder (Hoynes et al., 2012; Orrenius and Zavodny, 2010; Fromentin, 2016; Rodríguez-Planas and Nollenberger, 2016). Employers are more willing to shed employees with low marginal productivity or with little specialized training. Workers with higher educational attainment, instead, may move down the skill ladder, displacing less skilled workers (Papademetriou and Terrazas, 2009; Devereux, 2004). Therefore, low-educated workers are more vulnerable to structural changes than more educated ones (Bratsberg et al., 2018).

Second, immigrants are overrepresented in sectors vulnerable to the business cycle, such as construction, manufacturing and services (OECD, 2009, McKenzie, 2008). These sectors have been severely affected by the economic crisis, causing great job losses [1].

Third, there is a high concentration of immigrants in atypical jobs, understood as temporary or part-time jobs (Fernández and Ortega, 2008; European Commission, 2011), and higher shares of fixed-term contracts are typically associated with larger inflows into unemployment (see Dolado et al., 2002).

Fourth, there is an overrepresentation of young immigrants in the labour market relative to total immigrants (McKenzie, 2008). Young workers constitute one of the most vulnerable groups to an economic slowdown, suffering a greater probability of losing their jobs (Marcus and Gavrilovic, 2010; European Commission, 2011). The results of Hoynes et al. (2012) for the USA show that cyclicality declines with age.

There is evidence that the initial impact of a crisis in terms of earnings and employment of young workers persists over time. In the case of Belgium, Cockx and Ghirelli (2016) find negative effects in terms of the number of hours worked and earnings that persist up to 12 years. Choi et al. (2020), analyse the long-term effects of initial labour market conditions by comparing cohorts who graduated from college before, during and after the 1997–1998 Asian financial crisis in South Korea. The authors find a substantial and persistent reduction in employment and earnings, among other variables. Likewise, they suggest that labour market entry in a large-scale recession has prolonged effects on a young worker’s life course even after the penalties in the labour market have disappeared. Escalonilla et al. (2021), on the other hand, examine the Spanish case and find that for low-educated individuals, the main effect is via employment, even blocking their entry into the labour market. For high-skilled individuals, they obtain a long-term penalty in terms of wages as well as a reduction in the days worked.

In contrast, little is known about the persistence of the effects of a crisis on the immigrant population and the difference with native peers. Aslund and Rooth (2007) analyse the Swedish immigrant cohorts who started their professional career between 1987 and 1991 and estimate the long-term effects on immigrant earnings and employment from adverse labour market conditions faced upon arrival. Based on their results, one percentage point higher unemployment decreases the chances of employment by 3–5 percentage points and lowers earnings by 13–17 percentage points during the five to seven years after immigration. Therefore, facing tough economic conditions has a clear impact on immigrant earnings and employment for at least 10 years.

Recent literature addresses the macroeconomic implications of COVID-19 (Fornaro and Wolf, 2020; Guerrieri et al., 2022). Thus, Favara et al. (2023) analyse the impact of the negative shock created by the COVID-19 crisis on the labour market outcomes of young adults in India, Peru and Vietnam. The authors reveal that the shock induced by the COVID-19 crisis has denied employment opportunities and reduced earnings for young people. On average, their results show that this crisis reduced monthly earnings by 19.4% and employment levels by 17.5% in the three countries examined.

Regarding immigration issues, Borjas and Cassidy (2020) address the adverse effect of the COVID-19 labour market shock on immigrant employment in the USA, showing that the employment decline was particularly severe for immigrants. Although immigrant men were more likely to be employed than native men, the COVID-19 shock eliminated the immigrant employment advantage. Thus, immigrant men had lower employment rates than native men in 2020. The reversal occurred both because the rate of job loss for at-work immigrant men rose relative to that of natives and because the rate at which out-of-work immigrants could find jobs fell relative to the native job-finding rate.

In summary, facing poor labour trajectories increases the unemployment likelihood, which involves negative work and earnings scars in the long run (Gregg, 2001; Gregory and Jukes, 2001; Hansen and Lofstrom, 2009).

3. Data set and variables: descriptive statistics

The data source we use in this analysis is the Continuous Sample of Working Histories (CSWH) [2], conducted by the Spanish Ministry of Labour, Migration and Social Security. It provides a representative sample of all persons who have been enrolled in affiliation or received a Social Security contributory pension at some point during the reference year. The data contains information on working conditions as well as the socioeconomic characteristics of individual workers. This data set has been released annually since 2004, including all relationships with the Social Security Administration each year. Thus, it is possible to follow the labour market trajectories of individuals, allowing longitudinal analyses [3].

3.1 Sample and main variables

We consider all immigrants and natives aged between 16 and 30 who enrolled in Spanish Social Security for the first time between 2007 and 2021 [4]. We define immigrants as those individuals who were not born in Spain. The sample is selected on the basis of its entry year into the labour market. Using all editions of the CSWH available to date, we reconstruct the labour trajectories of workers to their last observed position in 2021 [5].

Our sample encompasses 1,519,097 employment spells. In terms of labour market participants, it includes 259,180 workers (natives and immigrants altogether), of whom about half are women (Table 1). Nearly 70% of young workers are aged between 16 and 22. By educational level, we observe that a significant proportion of workers are low educated. In fact, 60% of young workers in our sample have compulsory studies or lower, followed by 29% with vocational training or post-compulsory education. Only 11% of individuals have completed a university degree.

About 25% of the workers in our sample were born in a foreign country. We observe differences in personal characteristics between those who were born in Spain and abroad. First, immigrants are older than native-borns. Second, although the majority of natives and immigrants are low educated, the percentage is greater in the latter (59% and 63%, respectively). Nearly 31% of natives have post-compulsory studies, compared to 24% of immigrants. However, the percentage of native-born and foreign-born people with higher education is similar, reaching 10%.

The CSWH allows us to distinguish immigrants by their country of birth. In this case, there are many countries, so we group immigrants by country groups or geographic areas. For example, we group those immigrants born in Latin America into one group. In this way, we have six groups: those born in the EU15, excluding Spain (henceforward EU15), born in the rest of Europe (henceforward rEU), Latin America, Asia, North America (henceforward USA and Canada) and Africa.

Tables 2 and 3 present descriptive statistics of immigrants in our sample based on their country of birth. Latin American immigrants account for nearly 51% of the total, followed by those from non-EU15 European countries (20.3%) and African countries (14.2%). Immigrants from EU15 and Asian countries comprise a smaller portion, around 7%. However, the percentage of immigrants from North America is negligible, leading us to exclude this group from our analysis to ensure consistency and accuracy.

Our analysis focuses on seven labour market outcomes related to earnings and employment. As a measure of workers’ earnings, we use the logarithm of total annual earnings and the logarithm of average monthly earnings. Additionally, we analyse the logarithm of October-gained earnings (in line with the Wage Structure Survey, provided by the Spanish National Institute of Statistics) [6]. Regarding employment outcomes, we use the total number of months employed each year in logarithm terms and the employment probability, defined as the probability of being working at any month, the probability of being working at least six months [7] and the probability of being working in October.

3.2 Entry cohort as a unit of observation

Our analysis is carried out using cohorts, understood as delimited groups whose members share the same characteristic across different points in time within the labour market, thus dynamically tracking their earnings and employment trajectories (longitudinally). We define as cohort all individuals who share the same entry period into the labour market. Thus, we use the entry cohort as the unit of observation. Taking into account our time span, we examine cohorts entering the labour market between 2007 and 2021, which amounts to 15 different cohorts. This approach represents an alternative, and, therefore, a contribution to the existing literature, to previous studies where the unit of observation was the individual, because it analyses aggregate population groups [8].

4. Empirical strategy

Following Oreopoulos et al. (2012), we exploit the variation in the rate of unemployment at the regional level over the period 2007–2021 as a proxy for economic conditions.

Yet, workers’ labour market trajectory is affected not only by initial economic conditions but also by economic conditions at each moment. Then, as in Oreopoulos et al. (2012), we estimate a dynamic model that controls for the initial regional unemployment rate as well as for regional unemployment rates faced by the cohort throughout their labour trajectory. This allows us to distinguish between the effect of entry conditions and the impact of labour market conditions during their labour trajectory.

Because the rate of unemployment varies across regions and cohorts, individual level data were collapsed by entry cohort (c), entry region (r) and calendar year (t) and we worked with group-specific means of the variables, weighted by the corresponding cell sizes [9] [10]. The dynamic model can be written as follows (Oreopoulos et al., 2012):

(1) y¯crt=α+β0 * Exp0 * URre=0+β1 * Exp1 * URre=1+…+β14 * Exp14 * URre=14+δe+ρc+θr+γt+ecrt

where y¯crt is the group-specific mean of our outcome variables for entry cohort (c) in entry region (r) in calendar year (t); Exp_e is a dummy variable that takes value 1 if the cohort experience in a given year is equal to e; URre denotes the regional unemployment rate to which a cohort was exposed in each year of experience (e) in the corresponding region (r_e).

Fixed effects related to potential experience δ_e, entry cohort ρ_c, entry region θ_r and calendar year γ_t are also included in the empirical specification. Potential experience fixed effects δ_e capture time-invariant differences among all cohorts with the same amount of (potential) work experience. Entry cohort fixed effects ρ_c capture time-invariant differences between the characteristics of the different entry cohorts. Calendar-year fixed effects γ_t capture the component of regional business cycle variation that is common to all regions. Entry region fixed effects θ_r capture time-invariant differences across regions. Finally, α is the constant term and e_crt is a random error term.

Our coefficients of interest are β₀, β₂, β₃,…, β₁₄, which capture changes in experience profiles on earnings and employment that are attributable to entering the labour market during an adverse shock or during periods of economic growth, net of the effect of the future sequence of unemployment rates (that are correlated with the initial conditions). The model is estimated using weighted least squares, where weights are group sizes. To account for group-specific error components, we cluster standard errors at the entry cohort-region (cr) level [11].

5. Results

Before presenting and discussing the main results, we show graphical evidence on the evolution of our dependent variables related to earnings by entry cohort and potential experience over time. Because our observation unit is the cohort, each line corresponds to an entry cohort and its potential experience.

5.1 Descriptive evidence

Figures 1, 2 and 3 show the average evolution of three native-and-immigrant-earnings measures (the logarithm of total annual earnings, the logarithm of average monthly earnings and the logarithm of October-gained earnings, respectively) by entry cohorts over time. The dotted line represents the evolution of earnings by entry cohorts when they just enter the labour market, that is, they have no previous experience in the Spanish labour market. The coloured solid lines capture the evolution of earnings of each entry cohort as they gain professional experience in the labour market. The black dashed line represents the average earnings of entry cohorts when they accumulate 6 years of experience.

In a comparative way, immigrants seem to gain higher earnings at the time they enter the labour market than natives, regardless of their entry year (see the dotted line). The trend changes as cohorts accumulate experience. While earnings by immigrants remain stable or even tend to fall once the crisis begins, the opposite happens in the case of natives, for whom earnings increase with professional experience, although at a lower rate for those entering pre-crisis years and facing the crisis inside the labour market (see solid lines). When cohorts accumulate six years of experience, the difference in earnings between natives and non-natives is considerably reduced. All this evidence is similar, using any of the three earnings variables included in this paper and distinguishing by the immigrants’ country of birth [12].

Hence, we observe that the labour market entry earnings of both natives and immigrants are negatively affected by adverse shocks such as the 2008 economic crisis and the COVID-19 crisis. Once young natives enter the labour market, they progress in terms of earnings as they accumulate experience. Immigrants, however, in spite of gaining higher entry earnings, do not progress over time as natives but slow down in terms of earnings.

5.2 The impact of adverse shocks on earnings

Table 4 displays the estimated average marginal effects of the unemployment rate on earnings outcomes, with Figure 4 providing a graphical representation. The regressions are conducted separately for natives and non-natives, with the coefficients representing estimates of β0, β2,…, β14 from equation (1).

Natives experience a lasting negative impact on annual earnings, with a 1% increase in the unemployment rate leading to a 1.3% decrease at entry, escalating to a 3.2% drop after 12 years. Similar trends are observed for average monthly and October earnings. In contrast, immigrants face a more immediate negative impact on earnings, with a 1% rise in the unemployment rate resulting in a 1.8% decrease at entry. However, this effect diminishes over time, disappearing after around 11 years.

Further analysis, plotted in Figure 5, reveals variations in the magnitude and duration of negative impacts on annual earnings among immigrant groups based on their country of birth (see tables with estimates in the Supplementary File for our earnings outcomes). Immigrants from EU15 countries experience a short-lived negative impact, while those from rEU countries face a more persistent but stable effect. Greater and longer lasting negative impacts are observed for immigrants from Latin America and Asia. However, immigrants from African countries experience a smaller initial impact that persists and increases over time [13].

These findings are consistent across all earnings measures used, indicating that natives are more negatively affected in the long run by deteriorating labour market conditions compared to immigrants. However, immigrants face more immediate but diminishing negative effects on earnings. Additionally, variations in the earnings penalty are observed within immigrant groups.

5.3 The impact of adverse shocks on employment probability

Let us now focus on the impact of adverse shocks on employment probability. Table 5 presents the estimated average marginal effects of the unemployment rate on employment probabilities (see Figure 6 for a clearer view). Both immigrants and natives face negative impacts on employment probabilities in the short and long run, with immigrants experiencing the greatest negative effects. An increase in the unemployment rate by 1% leads to a reduction in the probability of working any month for natives and immigrants by 0.7% and 2.3%, respectively, at entry into the labour market. While the adverse effect remains stable for natives over time, it diminishes for immigrants, lasting around 14 years for both groups.

Similar trends are observed for the employment probability of working at least 6 months, with immigrants experiencing a more significant negative impact persisting for over 12 years. The total number of months employed each year also shows a substantial negative effect for immigrants, particularly in the initial years.

A more detailed analysis, plotted in Figure 7, reveals differences among immigrant groups in the magnitude and persistence of negative impacts on employment probability in any month for immigrants (see tables with estimates in the Supplementary File for our employment outcomes). While all immigrant groups experience a significant initial negative impact, variations exist in its persistence over time [14].

In summary, both natives and immigrants experience negative impacts on employment probabilities because of worsening labour market conditions. Immigrants face more immediate negative effects, which diminish over time, while natives experience a gradually increasing negative impact. Additionally, differences are observed within immigrant groups regarding the duration of negative effects on employment probabilities.

6. Heterogeneity analysis

In this section, we address the impact of adverse shocks on earnings and employment outcomes by conducting a heterogeneity analysis. More specifically, we attempt to assess the extent to which we can find different results for natives and immigrants, distinguishing by educational level, gender and age of entry into the labour market.

6.1 Educational attainment

We distinguish between three educational groups:

1. compulsory education;

2. vocational or post-compulsory education; and

3. university education.

All regressions are estimated separately by native-born and abroad born.

Natives experience more significant and long-term negative impacts on annual earnings compared to immigrants, especially among those with lower education levels (see Table A.4.1. in the Supplementary File). Initially, native youths see a 1% drop in annual earnings in response to a 1% increase in the unemployment rate, intensifying to a 3.1% decline in 14 years. For less qualified natives, this could accumulate to a 30% drop in earnings over the same period. Similarly, highly educated natives face a comparable but lesser impact, with earnings decreasing by 1% upon entry into the labour market and dropping by 2.1% after 14 years. Post-compulsory educated natives exhibit similar trends, albeit with no significant negative effect in the first three years of experience.

Turning to immigrants, the persistence of negative impacts on earnings is mainly driven by low-educated workers. Initially, annual earnings for low-educated immigrants decrease by 1.5% for every 1% increase in the unemployment rate, persisting over 14 years to a 2.1% drop. Highly educated immigrants experience a greater short-term reduction in earnings, declining by 2.5% upon entry into the labour market, with this impact lasting for seven years and resulting in a 1.8% decline. No significant effect is observed on the earnings of immigrants with post-compulsory education two years after entering the labour market.

The impact of adverse shocks on employment probability by educational levels shows that natives face a persistent employment scar, regardless of educational attainment (refer to Table A.4.2. in the Supplementary File). A 1% increase in the unemployment rate reduces the employment probability of less educated native cohorts by around 0.6% initially, slightly increasing over time. Similarly, immigrants, particularly lower educated ones, experience negative and long-term effects on employment probability when the regional unemployment rate increases. For instance, the employment probability of lower educated immigrants decreases by 2.4% initially, persisting over 14 years to around 0.8%.

Overall, these findings are consistent across all our measures of employment probability and earnings, although the magnitude of the negative impact may vary slightly [15].

6.2 Gender

We carry out our empirical strategy, distinguishing between two groups: males and females. All regressions are estimated separately by natives and immigrants.

Natives experience long-lasting negative effects on earnings during periods of rising unemployment, as shown in Table A.5.1. in the Supplementary File. Initially, male natives see a 1.3% decrease in annual earnings in response to a 1% increase in the unemployment rate, intensifying to a 3.4% decline over 14 years. Female natives initially experience a 0.8% earnings loss, catching up to male natives’ impact after 6 or 7 years. Immigrants also suffer from declining earnings because of worsened labour markets, but the effects are less enduring. Male immigrants see a 1.9% reduction in earnings upon entering the labour market, lasting for 11 years, with similar trends for female immigrants. Notably, the negative impact is slightly smaller for immigrant women and diminishes about two years earlier than for immigrant men.

These findings regarding earnings by gender are consistent when we use the average monthly earnings and October earnings, but the negative impact is slightly lower in these cases. Interestingly, when using October earnings, the negative effect persists longer for male immigrants compared to female immigrants [16].

Natives also endure long-lasting negative effects on employment probability during periods of rising unemployment, particularly upon entering the labour market (see Table A.5.2. in the Supplementary File). Male natives experience a 0.8% decrease in employment probability in response to a 1% increase in the unemployment rate, intensifying to a 1.7% drop over 14 years. Female natives initially see a 0.5% loss, increasing to 1.4% after 14 years. Similarly, immigrants face reduced employment probability because of deteriorating labour markets, especially during their early years of experience. Male immigrants witness a 2.2% decrease in employment probability upon entering the labour market, persisting for 13 years, with comparable trends for female immigrants. However, the negative impact is slightly smaller for immigrant women compared to men.

These findings are consistent across all our measures of employment probability, although immigrants generally experience a higher magnitude of negative impact compared to natives, particularly concerning the total number of months employed each year [17].

6.3 Age of entry into the labour market

In this case, we carry out our empirical strategy, distinguishing between two groups:

1. those who enter the labour market at 16 and 22 years old; and

2. those who enter the labour market at 23 and 30 years old.

All regressions are estimated separately by natives and immigrants.

Across different age groups upon entry into the labour market, natives face significant and enduring negative impacts on earnings during periods of rising unemployment (refer to Table A.6.1. in the Supplementary File). Although older, young natives experience a smaller negative impact initially, the effect persists over time. For instance, a 1% increase in the unemployment rate leads to a 1% decrease in annual earnings for natives entering the labour market at ages 16–22, compared to a 0.6% decrease for those entering at ages 23–30. These adverse effects intensify over the years, with earnings dropping by 3.4% and 1.3% over 14 years, respectively. Similarly, immigrants also face negative and enduring impacts on annual earnings, particularly among younger age groups (Column 3). Immigrants entering the labour market between ages 16–22 experience a 2% reduction in annual earnings upon entry, while those entering at ages 23–30 see a 1.4% drop. However, the adverse effects for the latter group dissipate within six to seven years.

Consistent results are observed when examining average monthly earnings and October earnings, with slightly lower negative impacts in these cases [18].

Natives also suffer significant and long-lasting negative impacts on their employment probability during periods of rising unemployment, regardless of their age at labour market entry (refer to Table A.6.2. in the Supplementary File). Although older, young natives experience a smaller negative impact initially, it remains relatively constant over time. A 1% increase in the unemployment rate leads to a 0.7% decrease in employment probability for natives entering at ages 16–22, compared to a 0.4% decrease for those entering at ages 23–30. These adverse effects persist, reaching drops of 0.8% and 0.6% over 14 years, respectively. In contrast, immigrants face more severe negative impacts on employment probability, particularly among younger age groups. Immigrants entering between ages 16 and 22 experience a 2.3% decrease in employment probability upon entry, whereas those entering at ages 23–30 see a 2.2% drop. However, the adverse effects for the latter group persist for nearly the same duration.

Similar results are found when analysing other employment variables, with immigrants experiencing a higher magnitude of negative impact compared to natives, especially concerning the total number of months employed each year [19]. Additionally, the impact varies significantly based on the age of entry for both natives and immigrants.

7. Discussion of the results

Our research reveals two primary findings. Firstly, we observe a notable decrease in the number of new entrants to the labour market during periods of economic downturn, such as the great recession and the COVID-19 crisis, particularly impacting low-educated workers. Secondly, within this context of entry barriers, we find adverse effects on employment and earnings for those entering the labour market, with differential impacts between native and immigrant populations.

In terms of employment, our analysis indicates that low-educated immigrants experience more significant negative effects during crises compared to similarly educated natives. This observation aligns with existing literature, which suggests that immigrants with lower educational attainment face higher job loss rates because of their heightened sensitivity to economic fluctuations. Factors contributing to this vulnerability include employment in industries prone to closure or downsizing and selection for layoffs during downturns because of factors such as job tenure and job type (Bratsberg et al., 2018). Additionally, challenges related to language barriers and educational credential devaluation further constrain immigrants to low-skilled, precarious jobs, exacerbating their vulnerability to economic shocks (Bratsberg et al., 2018; Becker, 1975; Dustmann et al., 2016).

The buffer theory posits that immigrants exiting the labour market during adverse shocks may create job opportunities for native workers, potentially leading to job displacement for immigrants (Castles, 2011; Bernardi and Garrido, 2008). Consequently, the negative impact on employment is often more pronounced for immigrants compared to natives. Furthermore, less-educated workers may face displacement by higher educated individuals moving down the skill ladder during crises, contributing to their heightened vulnerability (Devereux, 2004; Papademetriou and Terrazas, 2009). Notably, our findings suggest that higher educated immigrants experience smaller adverse effects in the labour market, which may be attributed to their possession of more transferrable skills and qualifications.

Regarding earnings, our analysis highlights long-term negative effects, particularly for immigrants born in regions such as Africa, Latin America and rEU. This trend may be attributed to job displacement, with displaced, less-educated workers experiencing enduring reductions in earnings (Couch and Placzek, 2010). Notably, the rigid collective bargaining system in Spain may contribute to delays in earnings adjustment during economic downturns, prolonging the adverse effects on immigrant earnings.

For higher educated cohorts, downward occupational mobility during economic downturns may result in an earnings penalty, as individuals may accept lower quality jobs offering fewer opportunities for promotion (Devereux, 2004). This trend is consistent with labour market segmentation theory (Doeringer and Piore, 1971), which posits that structural factors shape disparities between native and immigrant groups. Immigrants often find themselves concentrated in the secondary labour market, characterized by lower wages, limited job stability and fewer opportunities for promotion compared to the primary labour market where native workers typically reside.

In high-unemployment contexts, limited opportunities in the primary sector compel tertiary-educated young workers to accept lower quality jobs in the secondary sector, leading to diminished earnings and employment prospects. This situation results in a “sticky floor” effect, where many remain trapped in the secondary sector, facing ongoing instability and limited promotion opportunities. Consequently, this impacts long-term earnings and employment outcomes, particularly for highly educated individuals, including immigrants, highlighting the challenges they face in adverse labour market conditions.

In spite of all the evidence found in this article, our study has several limitations. Firstly, it does not account for individuals working in the informal sector, a significant proportion of immigrants, especially during economic expansions (Bosch and Farré, 2013). This is because of the data set we used, with which we can only analyse people who work within the labour market, that is, the formal sector. However, it remains difficult to quantify the number of illegal immigrants and the size of the informal economy (Chiswick, 1998). For Spain, Connor and Passel (2019) estimated between 100,000 and 200,000 irregular immigrants that year. In a recent working paper, Gálvez-Iniesta (2020) finds that at the end of 2019, there were around 390,000–470,000 irregular immigrants in Spain.

Secondly, selection bias may influence our estimates, as adverse shocks may affect labour force participation, potentially biasing our measure of their impact on earnings and employment. It is possible that those natives and immigrants who are able to access the labour market under tough labour conditions show greater self-selection, probably because they are more educated or have more skills. Likewise, less educated workers may be more likely to respond to the adverse shock by leaving or not entering the labour force. All of this may be biasing our results downward.

Additionally, the lack of data in the CSWH on non-employed workers limits our ability to correct for selection bias using standard techniques of selection bias correction (Heckman, 1979) and, therefore, to examine the influence of selection on our estimates. Nevertheless, this possible limitation may be negligible, given that the participation rates of natives and immigrants are relatively high in Spain (Murillo-Huertas and Simón, 2017; De la Rica et al., 2014).

8. Concluding remarks

In this study, we examine the lasting impacts of tough economic conditions on the labour market outcomes of young native–immigrant cohorts in Spain, focusing on the period 2007–2021, which includes the financial crisis of 2008 and the COVID-19 pandemic. Our research fills a gap in recent literature by examining how economic downturns affect the long-term prospects of vulnerable young entrants into the job market. We explore various factors, such as country of birth, educational level, gender and age of entry, to provide a comprehensive analysis of the issue.

Our findings indicate that economic crises initially restrict entry into the labour market, reducing the number of new participants. Subsequently, these crises have a significant negative impact on both employment and earnings for those entering the labour market, with variations observed among natives and immigrants, as well as within different demographic and educational subgroups. Less-educated immigrants are particularly vulnerable, facing job losses and displacement from native workers. On the other hand, highly educated immigrants often experience job downgrading and struggle to find suitable positions, leading to income stagnation or decline. This over-qualification problem exacerbates during economic downturns, posing challenges for educational policies.

It is worth noting that the long-term effects of the COVID-19 crisis are not yet fully understood because of insufficient data. Future research should focus on analysing the long-term implications of the COVID-19 pandemic on labour market outcomes.

Figures

Figure 1.

Evolution of annual earnings by entry year and experience

Figure 2.

Evolution of average monthly earnings by entry year and experience

Figure 3.

Evolution of October earnings by entry year and experience

Figure 4.

Impact of a 1% increase in the unemployment rate on earnings dependent variables

Figure 5.

Impact of a 1% increase in the unemployment rate on annual earnings by country of birth

Figure 6.

Impact of a 1% increase in the unemployment rate on employment dependent variables

Figure 7.

Impact of a 1% increase in the unemployment rate on probability employment in any month of immigrants by country of birth

Table 1.

Descriptive statistics for natives and immigrants

	All workers			Natives			Immigrants
	Obs.	Workers	% workers	Obs.	Workers	% workers	Obs.	Workers	% workers
Gender
Male	756,235	131,293	50.66	615,807	98,063	50.74	140,428	33,230	50.41
Female	762,862	127,887	49.34	622,514	95,203	49.26	140,348	32,684	49.59
Age of entry
16–22	1,048,335	171,561	66.19	913,401	141,440	73.18	134,934	30,121	45.70
23–30	470,762	87,619	33.81	324,920	51,826	26.82	145,842	35,793	54.30
Educational level
Compulsory education	993,616	155,511	60.00	806,205	113,856	58.91	187,411	41,655	63.20
Post-Compulsory education	380,475	75,025	28.95	318,572	59,212	30.64	61,903	15,813	23.99
University education	145,006	28,644	11.05	113,544	20,198	10.45	31,462	8,446	12.81
Total	1,519,097	259,180	100	1,238,321	193,266	74.57	280,776	65,914	25.43

Notes:

This table displays information relative to the number of observations and workers included in our sample as well as the percentage of workers. Also note that postcompulsory education includes vocational training

Source: Own elaboration using the CSWH

Table 2.

Descriptive statistics for immigrants born in EU15, rEU and Latin America (I)

	EU15			rEU			Latin America
	Obs.	Workers	% Workers	Obs.	Workers	% Workers	Obs.	Workers	% Workers
Gender
Male	8,258	2,226	48.06	27,596	6,276	46.88	64,313	15,901	47.74
Female	9,617	2,406	51.94	33,724	7,111	53.12	71,514	17,408	52.26
Age of entry
16–22	6,654	1,589	34.30	28,430	6,070	45.34	66,753	15,297	45.92
23–30	11,221	3,043	65.70	32,890	7,317	54.66	69,074	18,012	54.08
Educational level
Compulsory education	7,931	1,947	42.03	42,872	9,114	68.08	82,802	19,026	57.12
Post-compulsory education	4,959	1,365	29.47	12,605	2,860	21.36	36,210	9,611	28.85
University education	4,985	1,320	28.50	5,843	1,413	10.56	16,815	4,672	14.03
Total	17,875	4,632	7.03	61,320	13,387	20.31	135,827	33,309	50.53

Notes:

This table displays information relative to the number of observations and workers included in our sample as well as the percentage of workers. Also note that post-compulsory education includes vocational training

Source: Own elaboration using the CSWH

Table 3.

Descriptive statistics for immigrants born in Asia, USA and Canada and Africa (II)

	Asia			USA and Canada			Africa
	Obs.	Workers	% Workers	Obs.	Workers	% Workers	Obs.	Workers	% Workers
Gender
Male	13,234	2,455	56.93	1,783	454	50.50	25,244	5,918	63.13
Female	9,510	1,857	43.07	1,680	445	49.50	14,303	3,457	36.87
Age of entry
16–22	11,463	2,058	47.73	1,240	301	33.48	20,394	4,806	51.26
23–30	11,281	2,254	52.27	2,223	598	66.52	19,153	4,569	48.74
Educational level
Compulsory education	17,578	3,079	71.41	1,647	388	43.16	34,581	8,101	86.41
Post-Compulsory education	3,191	723	16.77	836	220	24.47	4,102	1,034	11.03
University education	1,975	510	11.83	980	291	32.37	864	240	2.56
Total	22,744	4,312	6.54	3,463	899	1.36	39,547	9,375	14.22

Notes:

This table displays information relative to the number of observations and workers included in our sample as well as the percentage of workers. Also note that postcompulsory education includes vocational training

Source: Own elaboration using the CSWH

Table 4.

Impact of a 1-percent increase in the unemployment rate on the earnings of natives and immigrants

	Natives			Immigrants
	(1)	(2)	(3)	(4)	(5)	(6)
Variables	Annual earnings	Average monthly earnings	October earnings	Annual earnings	Average monthly earnings	October earnings
UR × exp0	−0.013*** (0.001)	−0.010*** (0.001)	−0.008*** (0.001)	−0.018*** (0.003)	−0.013*** (0.002)	−0.013*** (0.002)
UR × exp1	−0.011*** (0.001)	−0.009*** (0.001)	−0.008*** (0.001)	−0.016*** (0.003)	−0.012*** (0.002)	−0.012*** (0.002)
UR × exp2	−0.013*** (0.001)	−0.010*** (0.001)	−0.009*** (0.001)	−0.016*** (0.002)	−0.012*** (0.002)	−0.012*** (0.002)
UR × exp3	−0.017*** (0.001)	−0.013*** (0.001)	−0.012*** (0.001)	−0.016*** (0.002)	−0.012*** (0.002)	−0.012*** (0.002)
UR × exp4	−0.020*** (0.001)	−0.015*** (0.001)	−0.014*** (0.001)	−0.017*** (0.002)	−0.013*** (0.002)	−0.012*** (0.002)
UR × exp5	−0.021*** (0.002)	−0.016*** (0.001)	−0.015*** (0.001)	−0.017*** (0.003)	−0.012*** (0.002)	−0.011*** (0.002)
UR × exp6	−0.022*** (0.002)	−0.017*** (0.001)	−0.016*** (0.002)	−0.016*** (0.003)	−0.012*** (0.002)	−0.010*** (0.002)
UR × exp7	−0.024*** (0.002)	−0.018*** (0.002)	−0.017*** (0.002)	−0.014*** (0.003)	−0.011*** (0.002)	−0.010*** (0.002)
UR × exp8	−0.026*** (0.002)	−0.020*** (0.002)	−0.018*** (0.002)	−0.012*** (0.003)	−0.008*** (0.002)	−0.008*** (0.002)
UR × exp9	−0.028*** (0.002)	−0.022*** (0.001)	−0.020*** (0.002)	−0.011*** (0.003)	−0.008*** (0.002)	−0.007*** (0.002)
UR × exp10	−0.029*** (0.002)	−0.023*** (0.002)	−0.021*** (0.002)	−0.008** (0.003)	−0.005* (0.003)	−0.006** (0.002)
UR × exp11	−0.031*** (0.002)	−0.025*** (0.002)	−0.022*** (0.002)	−0.009** (0.004)	−0.006** (0.003)	−0.003 (0.003)
UR × exp12	−0.032*** (0.003)	−0.025*** (0.002)	−0.023*** (0.002)	−0.008 (0.005)	−0.006 (0.004)	−0.007** (0.003)
UR × exp13	−0.032*** (0.003)	−0.025*** (0.002)	−0.023*** (0.003)	−0.009* (0.005)	−0.008* (0.005)	−0.009** (0.004)
UR × exp14	−0.033*** (0.004)	−0.026*** (0.003)	−0.024*** (0.003)	−0.009 (0.008)	−0.005 (0.006)	−0.010** (0.005)
Exp. FE	YES	YES	YES	YES	YES	YES
Region FE	YES	YES	YES	YES	YES	YES
Entry cohort FE	YES	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES	YES
Constant	8.041*** (0.031)	6.429*** (0.026)	6.582*** (0.025)	8.380*** (0.053)	6.629*** (0.039)	6.812*** (0.039)
Observations	2,040	2,040	2,040	2,040	2,040	2,040
R-squared	0.990	0.980	0.967	0.966	0.928	0.889

Notes:

Robust standard errors in parentheses; ***p < 0.01; **p < 0.05; *p < 0.1

Source: Own elaboration using the CSWH

Table 5.

Impact of a 1% increase in unemployment rate on probability employment of natives and immigrants

	Natives				Immigrants
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)
Variables	Any month	At least six months	Number of months employed	October month	Any month	At least six months	Number of months employed	October month
UR × exp0	−0.007*** (0.001)	−0.006*** (0.001)	−0.011*** (0.001)	−0.006*** (0.001)	−0.023*** (0.002)	−0.020*** (0.002)	−0.084*** (0.007)	−0.019*** (0.002)
UR × exp1	−0.004*** (0.001)	−0.006*** (0.001)	−0.010*** (0.001)	−0.006*** (0.001)	−0.017*** (0.001)	−0.020*** (0.002)	−0.083*** (0.007)	−0.019*** (0.002)
UR × exp2	−0.005*** (0.001)	−0.008*** (0.001)	−0.013*** (0.001)	−0.008*** (0.001)	−0.017*** (0.001)	−0.020*** (0.002)	−0.082*** (0.007)	−0.019*** (0.001)
UR × exp3	−0.005*** (0.000)	−0.008*** (0.001)	−0.015*** (0.001)	−0.009*** (0.001)	−0.017*** (0.001)	−0.022*** (0.002)	−0.083*** (0.007)	−0.020*** (0.001)
UR × exp4	−0.006*** (0.000)	−0.010*** (0.001)	−0.015*** (0.001)	−0.010*** (0.001)	−0.017*** (0.001)	−0.023*** (0.002)	−0.084*** (0.007)	−0.021*** (0.002)
UR × exp5	−0.006*** (0.000)	−0.010*** (0.001)	−0.015*** (0.001)	−0.010*** (0.001)	−0.018*** (0.001)	−0.023*** (0.002)	−0.082*** (0.008)	−0.021*** (0.002)
UR × exp6	−0.005*** (0.001)	−0.010*** (0.001)	−0.015*** (0.001)	−0.010*** (0.001)	−0.018*** (0.001)	−0.022*** (0.002)	−0.079*** (0.008)	−0.021*** (0.002)
UR × exp7	−0.006*** (0.001)	−0.011*** (0.001)	−0.015*** (0.001)	−0.011*** (0.001)	−0.017*** (0.001)	−0.021*** (0.002)	−0.072*** (0.008)	−0.019*** (0.002)
UR × exp8	−0.007*** (0.001)	−0.012*** (0.001)	−0.017*** (0.001)	−0.012*** (0.001)	−0.016*** (0.001)	−0.019*** (0.002)	−0.067*** (0.009)	−0.018*** (0.002)
UR × exp9	−0.008*** (0.001)	−0.014*** (0.001)	−0.018*** (0.001)	−0.014*** (0.001)	−0.015*** (0.002)	−0.018*** (0.002)	−0.065*** (0.009)	−0.017*** (0.002)
UR × exp10	−0.007*** (0.001)	−0.014*** (0.001)	−0.019*** (0.001)	−0.013*** (0.001)	−0.012*** (0.002)	−0.015*** (0.003)	−0.055*** (0.011)	−0.013*** (0.002)
UR × exp11	−0.008*** (0.001)	−0.014*** (0.001)	−0.019*** (0.001)	−0.014*** (0.001)	−0.013*** (0.003)	−0.016*** (0.003)	−0.057*** (0.013)	−0.015*** (0.003)
UR × exp12	−0.008*** (0.001)	−0.015*** (0.001)	−0.019*** (0.002)	−0.014*** (0.001)	−0.010*** (0.003)	−0.012*** (0.004)	−0.050*** (0.016)	−0.011*** (0.004)
UR × exp13	−0.008*** (0.001)	−0.015*** (0.001)	−0.020*** (0.001)	−0.014*** (0.001)	−0.009*** (0.003)	−0.009** (0.003)	−0.035** (0.014)	−0.008** (0.004)
UR × exp14	−0.008*** (0.001)	−0.015*** (0.002)	−0.020*** (0.002)	−0.015*** (0.001)	−0.007** (0.003)	−0.007* (0.004)	−0.026* (0.016)	−0.006 (0.004)
Exp. FE	YES	YES	YES	YES	YES	YES	YES	YES
Region FE	YES	YES	YES	YES	YES	YES	YES	YES
Entry cohort FE	YES	YES	YES	YES	YES	YES	YES	YES
Year FE	YES	YES	YES	YES	YES	YES	YES	YES
Observations	2,040	2,040	2,040	2,040	2,040	2,040	2,040	2,040

Notes:

Robust standard errors in parentheses; *** p < 0.01; **p < 0.05; *p < 0.1

Source: Own elaboration using the CSWH

Notes

1. Between 2008 and 2009, employment in the construction sector fell by 24.5% in Spain.

2. We use the CSWH because it presents advantages versus other data sets such as the LFS, especially in terms of detailed information of income and labour market trajectories. Also, it comes from administrative records, so it specifies when the worker enters the labour market for the first time, and the same worker can be followed over time.

3. Note that irregular migration is not included in the data.

4. Likewise, it is possible that including the COVID-19 crisis in our analysis distorts the short-term effect of the Great Recession. As robustness check, therefore, we delete the years 2020 and 2021 to examine possible differences in our results. Nevertheless, we find similar results, as can be seen in an extended working paper.

5. We only focus on the General Social Security Regime and we have selected full-time contracts.

6. Note these three variables are deflated using the 2016 CPI.

7. We also estimated all regressions using as dependent variable the probability of being working at least three months or at least nine months, finding similar results.

8. Note that entering under tough economic conditions, individuals may self-select, potentially biasing our estimates of the impact on earnings and employment due to selection bias. Increasing emigration during the crisis may underestimate great recession’s impact on foreign-born workers.

9. When collapsing the data at the cohort–region–year level, the resulting variable referred to the employment probability is transformed from a binary variable to a proportion variable. As our dependent outcome takes values in the interval [0,1], we need to restrict E(y|x) to be in [0,1]. To do so, we estimate a fractional response regression that captures non-linear relationships, using a probit model: E(y|x) = Φ(xβ). So, the results presented in this paper are those relative to the marginal effects.

10. For those workers moving between regions, we do not compute their labour outcomes in different regions, but we compute them based on his/her entry region. We have quantified the number of workers in our sample who have a different region of birth and region of entry into the labour market, and those with a different region of residence and region of entry into the labour market. Thus, only 6.8% of the observations relating to natives present a difference between the region of birth and the region of entry (84,733 observations). In terms of workers, the percentage of native workers affected is 6.6% (12,887 native workers). In the second case, both natives and immigrants may be affected. We observe that only 5.8% of the workers in our sample have a different region of residence and region of entry (15,081 workers). Furthermore, the percentage of natives affected is 4.5% (8,857 native workers) compared to 9.4% of immigrants (6,224 immigrant workers). Carrying out our empirical strategy by deleting those workers who present this difference between the region of birth, region of entry into the labour market and region of residence, we have found little change in our results.

11. This is the simplest and most widely used way of addressing serial correlation in studies using group-structured panel data (Angrist and Pischke, 2008).

12. Figures A.1.1, A.1.2 and A.1.3 in Supplementary File illustrate the average evolution of immigrant earnings measures by entry cohorts over time, based on their country of birth.

13. Tables A.2.1, A.2.2 and A.2.3 in Supplementary File show the coefficient estimates using the three earnings variables.

14. Tables A.3.1–A.3.4 in Supplementary File show the coefficient estimates using the four employment variables.

15. See a working paper published on ResearchGate where you can find the full text, all figures and the estimate tables relative to this paper.

16. See the previous working paper mentioned.

17. See the previous working paper mentioned.

18. See the previous working paper mentioned.

19. See the previous working paper mentioned.

Supplementary material

The supplementary material for this article can be found online.