Abstract
We analyse the role of inter-regional labor mobility in absorbing labor demand shocks in the Euro Area (EA). We find that mobility of foreign-born workers is strongly cyclical, while this is not the case for natives. Foreigners’ higher population-to-employment elasticity reduces the variation of overall employment rates over the business cycle: because of them, the impact of a one standard deviation change in employment on employment rates is six percent lower at the country level and seven percent lower at the regional level. Additionally, we compare Euro Area mobility to that of another currency union, the United States. We find that the population-to-employment elasticity estimated for foreign-born individuals is similar in the EA and the United States, while EA natives are significantly less mobile across countries than US natives are across states in response to labor demand shocks. This last result confirms that in the Euro Area there is room for improving country-specific shock absorption through higher labor mobility. It also suggests that immigration helps labor market adjustments.
This is a preview of subscription content, log in via an institution to check access.
Notes
Peri and Sparber (2009) show this mechanism at work in the USA. D’Amuri and Peri (2014) show this mechanism for the EU before and during the Great Recession; Foged and Peri (2016) show this mechanism at work in Denmark; Basso et al. (2017) show that, due to their specialization, migrants can “push” natives in the intermediate part of the wage distribution reducing polarization of the labor market for natives.
Foreign-born individuals are individuals that were born outside of the country analysed, both when the analysis is carried out at the country and at the regional level.
Our main sample includes both early euro adopters (Austria, Belgium, Finland, France, Germany, Greece, Ireland, Italy, Luxembourg, the Netherlands, Portugal, Spain) and late adopters (Cyprus, Slovakia, Estonia, Latvia, Lithuania and Slovenia). We do not include only Malta as the data are available from 2009 onwards only. For more details on sample selection, see Table 13 in Appendix.
Only for Germany, we use citizenship instead of place of birth, since this last information is not available in the EU-LFS. We do not have information on the exact place of birth of the individuals, but we only know whether they were born in the country, outside the country but in the EU, or outside the EU.
We look at the aggregate change of these populations in response to shocks and not to the behaviour of individuals in each category. Hence, we cannot follow individuals across time and space.
Refugees are excluded from our sample because asylum applicants and displaced persons who have been granted temporary protection reside in group quarters until their asylum application has been accepted. Moreover, according to Eurostat, the coverage of recent migrants is limited in the EU-LFS data. For these reasons, even the most recent data do not cover yet the recent refugee migration waves.
As a robustness check, we also run all the main regressions of the paper including a control for log employment in the other group or using overall employment as a measure of the shock and the results are essentially unchanged.
As already mentioned in the previous note, as a robustness check, we also run all the main regressions of the paper including a control for log employment in the other group or using overall employment as a measure of the shock and the results are basically unchanged.
We chose as base year 2006 to optimize the trade-off between power and exogeneity of the instrument. In the context of Bartik instruments, a recent paper by Jaeger et al. (2018) suggests to use the share from a baseline year as far as possible for the period analysed. However, we would lose power in going much back in time with respect to 2006.
Cadena and Kovak (2016) for the USA find an elasticity that is about twice as large.
A more detailed comparison on mobility in the two currency areas is carried out in Sect. 6.
We further investigated whether the migration response to negative labor demand shocks differs from that to positive shocks and whether the elasticity to large employment shocks is different to that to smaller ones. In neither of the cases, we find evidence for asymmetric or nonlinear elasticities. (A full discussion of these results can be found in “Appendix 2”.)
The construction sector shows a strong cyclical behaviour and employs many immigrants (see the particularly notable case of Spain, illustrated in Gonzalez and Ortega 2013). In a robustness check, we analyse whether it is the driver of our results. Table 15 in Appendix 2 shows that this is not the case: The population to employment elasticity for foreign-born is not significantly different when excluding construction workers from the sample. In performing that exercise, we excluded from the sample all the people employed in the construction sector.
In the EU-LFS data, individuals’ age is available in 5-year intervals. Hence, we assume that cohort sizes by year of birth are constant within such age brackets and subtract (add) in each year t from (to) the actual population value the number of individuals in the 10–14 (60–64) age class in year t − 1 that are expected to enter (exit) working age in year t. These additional estimates of Eq. 1 that are not reported in the main text for brevity are very similar to the main ones reported in Tables 4 and 5 both when using actual employment variations and when instrumenting it with the Bartik IV described in Sect. 3. In a further robustness analysis (not reported, but available upon request), we also run all the main regressions limiting the sample to individuals aged 15–54 to test whether changes in retirement rules may affect the estimates. All the results are robust to the age definition of the sample.
On average in the 2006–16 interval, the share of highly educated individuals was slightly higher among natives (25.5 percent) than among foreign-born (21.7 percent). In this last category, EU foreigners tend to have a share of highly educated that is equal to the one found for natives (25.3 percent), while such share is lower among the extra-EU (20.0 percent; Figure 7 of Appendix). If anything, the fact that the share of highly educated individuals is higher among natives than among the group of migrants and that the highly educated tend to be more mobile implies that the mobility gap between natives and foreigners would be larger when controlling for education.
Ottaviano and Peri (2012) estimate an elasticity of substitution of 20 between immigrants and natives of similar skills, Manacorda et al. (2012) an elasticity of 10. Even the exact estimate of the elasticity of substitution can be important to calculate specific wage effects, the literature agrees that immigrants and natives are gross (but not perfect) substitutes.
See Appendix Table 19.
In Appendix Tables 20 and 21, we observe larger and more persistent effects of a labor demand shock on employment/population ratios in Euro Area regions with fewer migrants. In these areas, not only the effects of a sector-driven shock on employment and employment-to-population rates are larger, but they persist up to two years (although not precisely estimated).
Another source of bias is the one identified by Teulings and Zubanov (2014) that may lead to an underestimation of the persistence and it is likely to affect our estimates. For these reasons, we prefer to consider these results as mainly suggestive.
The previous literature has shown some conflicting results on the response of migration during upturns and downturns. Saks and Wozniak (2011) find that US internal migration is strongly procyclical as the benefits of moving rise during upturns; in a recent paper Dao et al. (2018) find instead counter-cyclical migration responses in the US using population and inter-state migration data. Monras (2018) finds that most of the response to local shocks occurs through changes in in-migration flows.
The sample is composed of all people aged 15–64, not living in group quarters, and not currently enrolled in school, residing in US states in the period 2007–2016.
There are two substantial differences between US commuting zones and European NUTS 2 regions: (i) the latter are based on jurisdictional units and not based on commuting patterns; (ii) the average population of a US CZ is about 200,000 (15–64 years old not enrolled in school) – there are 741 CZs using the 1990 definition – while the average population of a NUTS 2 region is 3,300,000 – there are 65 regions in our sample.
In order to avoid dealing with breaks in the occupation and education classifications in the EU-LFS data, we measure the qualification level of each occupation as of 2011. We perform the same analysis on US Census American Community Survey data using the same base year.
We cannot fully replicate the local projection analysis of Sect. 5.2 on US data as in the US an out-of-state-born American is still a citizen, differently from the Euro Area. If we were to divide US states and commuting zones based on the share of out-of-state-born population, we would have to consider also American citizens and, thus, we would not be able to compare the results with those of Sect. 5.2.
When looking at mobility across regions, the inclusion of country-by-year fixed effects in the main estimates would increase the population to employment elasticity for natives (Table 27 in Appendix), but not for the foreign-born. This result indirectly confirms that natives tend to react to within country shocks by relocating, but are less mobile across countries.
References
Arpaia, A., A. Kiss, B. Palvolgyi, and A. Turrini. 2016. Labour Mobility and Labour Market Adjustment in the EU. IZA Journal of Migration 5: 1–21.
Barone, G., A. D’Ignazio, G. De Blasio, and P. Naticchioni. 2016. Mr. Rossi, Mr. Hu and Politics. The Role of Immigration in Shaping Natives’ Voting Behavior. Journal of Public Economics 136: 1–13.
Bartik, T. J., 1991. Who Benefits from State and Local Economic Development Policies? W.E. Upjohn Institute for Employment Research.
Basso G., G. Peri and A. Rahman, 2017. Computerization and Immigration: Theory and Evidence from the United States. NBER Working Paper 23935, National Bureau of Economic Research, Inc.
Blanchard, O.J., and L.F. Katz. 1992. Regional Evolutions. Brookings Papers on Economic Activity 1: 1–75.
Beyer, R.C.M., and F. Smets. 2015. Labour Market Adjustments and Migration in Europe and the United States: How different? Economic Policy 30(84): 643–682.
Borjas, G. J. 2001. Does Immigration Grease the Wheels of the Labour Market? Brookings Papers on Economic Activity, Economic Studies Program, The Brookings Institution, 32(1): 69–134.
Cadena, B., and B. Kovak. 2016. Immigrants Equilibrate Local Labour Markets: Evidence from the Great Recession. American Economic Journal: Applied Economics 8(1): 257–290.
Dao, M., Furceri, D. and P. Loungani. 2018. Regional Labour Market Adjustment in the United States: Trend and Cycle. The Review of Economics and Statistics, forthcoming.
D’Amuri, F., and G. Peri. 2014. Immigration, Jobs, and Employment Protection: Evidence from EUROPE Before and During the Great Recession. The Journal of the European Economic Association 12(2): 432–464.
Decressin, J., and A. Fatás. 1995. Regional Labour Market Dynamics in Europe. European Economic Review 39(9): 1627–1655.
Edo, A., Giesing, Y., Öztunc, J. and P. Poutvaara. 2017. Immigration and Electoral Support for the Far Left and Far Right. Ifo WP 244.
Fahri, E. and I. Werning. 2014. Labour Mobility Within Currency Unions, NBER Working Paper No. 20105.
Foged, M., and G. Peri. 2016. Immigrants’ Effect on Native Workers: New Analysis on Longitudinal Data. American Economic Journal: Applied Economics, American Economic Association 8(2): 1–34.
Foote, A. 2016. The Effects of Negative House Price Changes on Migration: Evidence Across U.S. Housing Downturns. Regional Science and Urban Economics 60: 292–299.
Gonzalez, L., and F. Ortega. 2013. Immigration and Housing Booms: Evidence from Spain. Journal of Regional Science 53(1): 37–59.
Halla, M., A.F. Wagner, and J. Zweimueller. 2017. Immigration and Voting for Far Right. Journal of the European Economic Association 15(6): 1341–1385.
Hoffmann, M., Maslov, E., Sorensen, B. E. and I. Stewen. 2018. Are Banking and Capital Markets Union Complements? Evidence from Channels of Risk Sharing in the Eurozone, paper presented at the “Euro at 20” IMF-Central Bank of Ireland conference in Dublin, 25–26 June 2018.
House, C., Proebsting, C. and L. L. Tesar. 2018. Quantifying the Benefits of Labour Mobility in a Currency Union, NBER Working Paper No. 25347.
Huttunen, K., J. Møen, and K.G. Salvanes. 2018. Job loss and Regional Mobility. The Journal of Labour Economics 36(2): 479–509.
Jaeger, D. A., Ruist, J. and J. Stuhler. 2018. Shift-Share Instruments and the Impact of Immigration. NBER Working Paper No. 24285.
Jordà, O. 2005. Estimation and Inference of Impulse Responses by Local Projections. American Economic Review 95(1): 161–182.
Manacorda, M., A. Manning, and J. Wadsworth. 2012. The Impact of Immigration on the Structure of Wages: Theory and Evidence from Britain. Journal of the European Economic Association 10(1): 120–151.
Mincer, J. 1978. Family Migration Decisions. Journal of Political Economy 86(5): 749–773.
Modestino, A.S., and J. Dennett. 2013. Are American Homeowners Locked into Their Houses? The Impact of Housing Market Conditions on State-to-State Migration. Regional Science and Urban Economics 43(2): 322–337.
Monras, J. 2018. Economic Shocks and Internal Migration CEPR Discussion Paper 12977.
Mundell, R.A. 1961. A theory of Optimum Currency Areas. American Economic Review 51(4): 657–665.
Obstfeld, M., and G. Peri. 1998. Regional Non-adjustment and Fiscal Policy. Economic Policy 13 (26): 205–259.
Ottaviano, G.I.P., and G. Peri. 2012. Rethinking the Effect of Immigration on Wages. Journal of the European Economic Association 10(1): 152–197.
Peri, G., and C. Sparber. 2009. Task Specialization, Immigration, and Wages. American Economic Journal: Applied Economics, American Economic Association 1(3): 135–169.
Peri, G. 2018. Immigrants and Poverty: How Do They Cope with it, How Do They Affect Natives? In the Human and Economic Implications of Twenty-First Century Immigration Policy, Susan Pozo, ed. W.E. Upjohn Institute for Employment Research.
Ruggles, S., Genadek, K., Goeken, R., Grover, J. and M. Sobek. 2017. Integrated Public Use Microdata Series: Version 7.0 [dataset]. Minneapolis: University of Minnesota.
Saks, R.E., and A. Wozniak. 2011. Labour Reallocation over the BUSINESS Cycle: New Evidence from Internal Migration. Journal of Labour Economics 29(4): 697–739.
Stock, J. and M. Yogo. 2005. Testing for Weak Instruments in Linear IV Regression. In: Andrews DWK Identification and Inference for Econometric Models. New York: Cambridge University Press; 2005. pp. 80–108.
Teulings, C.N., and N. Zubanov. 2014. Is economic Recovery a myth? Robust Estimation of Impulse Responses. Journal of Applied Econometrics 29(3): 497–514.
Acknowledgements
Prepared for the Central Bank of Ireland (CBI) and the International Monetary Fund (IMF) Conference “Euro at 20,” and for the IMF Economic Review. We thank two anonymous referees, Alan Barrett, Olivier Blanchard, Emine Boz, Matteo Bugamelli, Federico Cingano, Giovanni dell’Ariccia, Pierre-Olivier Gourinchas, Mathias Hoffmann, Philip Lane, Juri Marcucci, Gian Maria Milesi-Ferretti, Doug Miller, Marianna Riggi, Paolo Sestito, Luigi Federico Signorini, and the IMF Economic Review guest editors Philippe Martin and Sebnem Kalemli-Ozcan for their comments on previous drafts. The views expressed in this paper do not necessarily reflect those of the Bank of Italy.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendices
1.1 Appendix 1: Description of the Data
See Table 13.
Appendix 2: Additional Results and Robustness Checks
See Tables 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26 and 27.
In Table 14 we split the first-stage estimates of Table 3, panel B, between the subgroups of EU versus extra-EU-born among immigrants, hence identifying three groups, finding similar results.
Table 15 reports a robustness check in which we exclude from the sample all individuals who declare working in the construction sector. The discussion of these results is reported in the main text, Sect. 4, footnote 14.
In the results presented in the main text, we found that—compared to natives—foreigners’ show larger population-to-employment elasticity both at the country and at the region levels. We now test whether such estimated average elasticity can have different values during upturns and downturns. Tables 16 and 17 report additional evidence that rules out asymmetric population-to-employment elasticities in response to positive/negative shocks: in columns 1 and 2 of both tables we replicate the main analysis (regression Eq. (1)) also interacting the shock with a dummy equal to 1 if the employment change was negative (including the indicator for a negative employment change). At both country (Table 16) and regional (Table 17) levels, the results show that the value of the elasticity does not change significantly during downturns, with point estimates for the additional interaction that are very close to zero and never statistically significant. 2SLS results, not reported here for brevity, confirm these findings. We can thus conclude that the values of the population-to-employment elasticities do not change significantly during labor market upturns and downturns. Immigrants respond much more actively than natives do to positive and to negative area employment shocks and their mobility is not hindered by recessions.
Similarly, in columns 3 and 4 of Table 16 (Table 17) we also interact the main independent variable of Eq. 1 with two dummies that are equal to one, respectively, if the employment change for the country (region) falls within the first or the fourth quartile of the distribution of the employment changes (including again the indicators of the main effects). This is a check to see whether the elasticity to large employment shocks is different to that to smaller ones. Also in this case, point estimates for the interactions are very close to zero (and it is statistically significant only for foreign-born in case of small employment shocks). All in all, these results provide no evidence of significant differences in the population-to-employment elasticity during large upturns and downturns, or for small employment variations.
In Tables 18, 19, 20, 21, 22, 23, 24 and 25, we report additional robustness checks on the local projections results that we discuss in the main text.
In Table 26 we report the employment elasticities breaking down the sample in young and old, similarly to what we did for low and high educated in Table 7 in the main text.
In Table 27 we show estimates of the population elasticity to employment at the regional level, in a format comparable to Table 5, when including country-by-year fixed effects (Fig. 7).
Share of college graduates on population. Note: Our calculations on EU-LFS data
