What Happens to the Careers of European Workers When Immigrants “Take Their Jobs”?

Source: authors’ calculation based on ECHP data.

Notes: Columns 1 and 2 report statistics by individual-years, summing up to the total sample size. Columns 3 and 4 report frequencies and shares of individual who have ever been of each tier. The total frequency is higher than the number of individuals suggesting that some individuals have been employed in different tiers over the period considered.

↵Note: Each column reports the estimate from a different regression where the dependent variable is defined as equal to 0 if the individual is at the same tier level as when he first entered the sample, 1 if he is at a higher tier, and −1 if he is at a lower tier. The coefficient reported in Columns 1–3 is the coefficient on the share of foreign born at time t. In Column 4, earlier periods are also controlled for. Column 5 reports also the coefficient on the share of immigrants in the next higher tier level at time t. The first column uses the whole sample of 11 countries, the remaining columns use data from the subsample where instrumental variables are available. All regressions are performed at the individual level and include controls for education, marital status, tenure, and industry, as well as year fixed effects. In brackets, we report the standard error using two-way clustering at the individual and at the year-country-occupation levels. *,**,*** indicate significance at the 10, 5, and 1 percent level, respectively.

↵Notes: Each column reports the estimate from a different regression where the dependent variable is defined as equal to 0 if the individual is at the same or at a lower tier level than when he first entered the sample and 1 if he is at a higher tier. The coefficient reported in Columns 1–3 is the coefficient on the share of foreign born at time t. In Column 4, earlier periods are also controlled for. Column 5 reports also the coefficient on the share of immigrants in the next higher tier level at time t. The first column uses the whole sample of 11 countries, the remaining columns use data from the subsample where instrumental variables are available. All regressions are performed at the individual level and include controls for education, marital status, tenure, and industry, as well as year fixed effects. In brackets, we report the standard error using two-way clustering at the individual and at the year-country-occupation levels, *,**,*** indicate significance at the 10, 5, and 1 percent level, respectively.

↵Notes: Each column reports the estimate from a different regression where the dependent variable is defined as equal to 0 if the individual is at the same or at a higher tier level than when he first entered the sample and 1 if he is at a lower tier. The coefficient reported in Columns 1–3 is the coefficient on the share of foreign born at time t. In Column 4, earlier periods are also controlled for. Column 5 reports also the coefficient on the share of immigrants in the next higher tier level at time t. The first column uses the whole sample of 11 countries, the remaining columns use data from the subsample where instrumental variables are available. All regressions are performed at the individual level and include controls for education, marital status, tenure, and industry, as well as year fixed effects. In brackets, we report the standard error using two-way clustering at the individual and at the year-country-occupation levels, *,**,*** indicate significance at the 10, 5, and 1 percent level, respectively.

↵Notes: Each column reports the estimate from a different regression where the dependent variable is defined as the individual tier level at time t. The coefficient reported in Columns 1–3 is the coefficient on the share of foreign born at time t. In Column 4, earlier periods are also controlled for. The first column uses the whole sample of 11 countries, the remaining columns use data from the subsample where instrumental variables are available. All regressions are performed at the individual level and include controls for education, marital status, tenure, and industry, as well as year and individual fixed effects. In brackets, we report the standard error using two-way clustering at the individual and at the year-country-occupation levels. *,**,*** indicate significance at the 10, 5, and 1 percent level, respectively.

↵Notes: Each column reports the estimate from a different regression where the dependent variable is defined as 1 if the individual is unemployed and 0 if regularly working. The coefficient reported in Columns 1–3 is the coefficient on the share of foreign born at time t. In Column 4, earlier periods are also controlled for. The first column uses data of the whole sample, the remaining columns use only data of the subsample where instrumental variables are available. All regressions are performed at the individual level and include controls for education, marital status, and for employed people (Panel B) tenure and industry. Individual and year fixed effects are also included. In brackets, we report the standard error clustered at the individual and at the year-country-occupational levels. *,**,*** indicate significance at the 10, 5, and 1 percent level, respectively.

↵Notes: Each column reports the estimate from a different regression where the dependent variable is defined as the log of net wage and salary earnings. The coefficient reported in Columns 1–3 is the coefficient on the share of foreign born at time t. In Column 4, earlier periods are also controlled for. The first column uses data of the whole sample, the remaining columns use only data of the subsample where instrumental variables are available. All regressions are performed at the individual level and include controls for education, marital status, and for employed people (Panel B) tenure and industry. Individual and year fixed effects are also included. In brackets, we report the standard error clustered at the individual and at the year-country-occupational levels. *,**,*** indicate significance at the 10, 5, and 1 percent level, respectively.

↵Notes: Each column reports the estimate from a different regression. In Panel A, the dependent variable is defined as 1 if an employed person receives only wage and salary and no self-employment income and 0 otherwise. In Panel B, the dependent variable is defined as the log of net self-employment income. The coefficient reported in Columns 1–3 is the coefficient on the share of foreign born at time t. In Column 4, earlier periods are also controlled for. The first column uses data of the whole sample, the remaining columns use only data of the subsample where instrumental variables are available. All regressions are performed at the individual level and include controls for education, marital status, tenure, and industry. Individual and year fixed effects are also included. In brackets, we report the standard error clustered at the individual and at the year-country-occupational levels. *,**,*** indicate significance at the 10, 5, and 1 percent level, respectively.

↵Notes: Each column reports the estimate from a different regression where the dependent variable is defined as 1 if the individual changed occupation toward one with an ISCO code that is different with respect to the one he/she had when he entered the sample, and 0 otherwise. The coefficient reported in Columns 1–3 is the coefficient on the share of foreign born at time t. In Column 4, earlier periods are also controlled for. The first column uses data of the whole sample, the remaining columns use only data of the subsample where instrumental variables are available. All regressions are performed at the individual level and include controls for education, marital status, tenure, and industry, as well as individual and year fixed effects. In brackets, we report the standard error clustered at the individual and at the year-country-occupational levels. *,**,*** indicate significance at the 10, 5, and 1 percent level, respectively.

↵Notes: Each column reports the 2SLS estimate from different regressions of an outcome variable (from top to bottom, respectively, Occupation mobility in tiers, Unemployment status and Log wage earnings) over the contemporaneous share of immigrants. Columns differs by the subsample selected, which is by tier (either the first two or the last two), by gender, and by age at entry in the sample. All regressions are performed at the individual level and include controls for education, marital status, tenure, and industry, as well as year fixed effects. All regressions but those on the Occupation mobility outcome variable include individual fixed effects. All columns report the estimate for the specification with all interaction dummies (Country*Occupational Level, Country*Year, Occupational level*Year). In brackets, we report the standard error clustered at the individual and at the year-country-occupation level, *,**,*** indicate significance at the 10, 5, and 1 percent level, respectively.

Source: authors’ calculation based on ECHP data and O*NET data.

Notes: Column 1 provides the percentage of native workers with tertiary education. Monetary values in ECU until 1998, in Euro from 1999 onward. The scores in Column 4 are the average scores in complex, mental, and communication skills. A score equal to 78 in complex skills for Tier 4 implies that 78 percent of all workers use complex skills less intensively than workers in Tier 4. The scores in Column 5 are the average scores in manual and routine skills. Statistics weighted using individual weights.

Source: authors’ calculation based on ECHP data.

Notes: Monetary values in ECU until 1998, in Euros from 1999 onward.

Source: authors’ calculation based on ECHP data.

Source: authors’ calculation based on ELFS data.

↵Notes: The table reports the first stage regressions of Specification 4 of Tables 3–5. In brackets, we report the standard errors using two-way clustering at the individual and at the year-country-occupation levels. *,**,*** indicate significance at the 10, 5, and 1 percent level, respectively.