Recent returnees: labour market performance based on LFS data

Table 8.1 summarises the descriptive statistics of our LFS sample. In the first column, we present descriptive statistics for the entire sample. The second column - non-migrants - describes those Lithuanian nationals who either have lived in Lithuania their entire lives or have returned from abroad more than a year ago. The third column illustrates the main socio-demographic characteristics of recent returnees. In the fourth column, we show the differences between nonmigrants and recent returnees, including whether these differences are statistically significant.

Table S.l Descriptive statistics of the sample, LFS data for the period 2013-2018

AH sample 1

Non-migrants

2

Recent

returnees

3

Differential

  • (t-test)
  • 2-3

Age (median)

50-54 years- old

50-54 years- old

30-34 years- old

2.703***(0.273)

Male

0.446

0.446

0.525

-0.0787**(0.039)

Single

0.231

0.231

0.506

-0.276***(0.033)

Married

0.580

0.580

0.346

0.235***(0.039)

Divorced, legally separated, or widowed

0.189

0.189

0.148

-0.316***(0.051)

Low education

0.178

0.178

0.080

0.098***(0.030)

Medium education

0.547

0.546

0.679

-0.133***(0.039)

High education

0.276

0.276

0.241

0.035 (0.035)

Employed (Y/N)

0.562

0.562

0.475

0.087** (0.039)

Observations

77,776

77,614

162

Source. Own elaboration based on LFS data 2013-2018.

Notes. Standard errors in parentheses; *** denotes significance at 0.01; ** denotes significance at 0.05. To ensure anonymity, Lithuanian Statistics provided the age variable grouped in 5-year intervals, hence the differential indicates the difference between two age groups. Classification of education levels follows the approach of EU LFS: High refers to tertiary education, Medium - to upper secondary and post-secondaiy non-tertiary education, and Low to the rest (CIRCABC, 2019).

It is immediately apparent that recent returnees differ from non-migrants. Recent renirnees are younger, and more of them are male and single compared to the rest of the Lithuanian population. Fewer returnees have low education, and more of them have acquired medium education. These observations go in line with similar research performed previously (Hazans & Philips, 2011; Martin & Radu, 2012). Unlike in previous research, however, in our dataset we do not see significant differences between the shares of degree holders among the general population and the returnees.

Given that recent renirnees are different from non-migrants at least in some regards, we control for their socio-demographic characteristics when assessing their position on the labour market. The first column in Table 8.2 summarises the results of the regression, which estimates whether returning from abroad affects one's likelihood of working. We find that the recent renirnees are 86 per cent less likely to be working than the general population.

In order to check whether the impact is different for more- and less-educated renirnees, we then limit the sample to renirnees only and estimate the effect of having a higher degree on whether a person is working using a logistic regression (see column 2 in Table 8.2). The results are significant, meaning that work outcomes among returnees are differentiated by education. To be more precise, renirnees who are degree holders have double the odds of working compared to renirnees who do not hold a degree.

If recent renirnees are more likely to be not working compared to the general labour force, are they more likely to reach out to public employment office to find work? To answer this question, we run another logistic regression (see column 3 in Table 8.2) with a dependent variable set to 1 if a respondent contacted the public employment office within four weeks prior to taking the survey and 0 otherwise.10 We find that renirnees are 66 per cent less likely to contact a public employment office to find work compared to non-returnees. Regarding other institutional involvement, returnees also appear less likely to participate in regular education, even when we control for age and other socio-demographic characteristics (see column 4 in Table 8.2).

Given findings in the literature that returnees are also more prone to be self- employed (Karolak, 2016; Martin & Radu, 2012), we also assess whether returnees are more likely to be self-employed than non-migrants (see colunm 3 in Table 8.2). In line with other research, we find that returnees have more than twice the odds of being self-employed. Based on econometric analysis, some scholars (Demurger & Xu, 2011; McCormick & Wahba, 2001) argue that savings and skills acquired abroad foster entrepreneurship upon return. Nevertheless, drawing on interviews with Polish returnees. Karolak (2016) points out that many businesses started by returnees soon fail, the primary motivation for starting a business is the hesitancy to work ‘for a Polish boss’, and that few self-employed returnees employ others. This calls into question whether self-employment among Lithuanian returnees is a sign of entrepreneurship or difficulty integrating into the Lithuanian labour market.

Table 8.2 Work-related regression results

(1)

DV:

Employed

(Y/N)

(2)

DV:

Employed

(Y/N)

(3)

DV: Public ' employment office(Y/N)

(4)

DV:

Student

(Y/N)

(5)

DV: Self- employed (Y/N)

Returnee

  • 0.141***
  • (0.051)
  • 0.342***
  • (0.130)
  • 0.482***
  • (0.137)
  • 2.630***
  • (0.723)

Employment status a year ago (ref. employed)

Unemployed

  • 0.021***
  • (0.001)

Inactive

  • 0.005***
  • (0.000)

Age (ref.:

15-29-year- olds)

30-44-year-

olds

  • 0.577***
  • (0.042)
  • 1.866***
  • (0.227)
  • 0.035***
  • (0.002)
  • 2.054***
  • (0.155)

45-59-year-

olds

  • 0.438***
  • (0.032)
  • 2.367***
  • (0.309)
  • 0.008***
  • (0.001)
  • 2.254***
  • (0.173)

60+ year-olds

  • 0.103***
  • (0.008)
  • 1.806***
  • (0.351)
  • 0.000***
  • (0.000)
  • 2.523***
  • (0.214)

Females (ref. Males)

  • 0.855***
  • (0.027)
  • 1.522***
  • (0.123)
  • 1.489***
  • (0.057)
  • 0.623***
  • (0.020)

Marital Status (ref. Single)

Married

  • 1.277***
  • (0.077)
  • 1.117
  • (0.125)
  • 0.200***
  • (0.014)
  • 1.102
  • (0.065)

Divorced, legally separated, or widowed

  • 1.044
  • (0.072)
  • 1.018
  • (0.137)
  • 0.202***
  • (0.032)
  • 1.089
  • (0.076)

Education (ref. Low)

Medium

  • 3.531***
  • (0.168)
  • 1.195
  • (0.154)
  • 0.782***
  • (0.058)

High

  • 6.412***
  • (0.354)
  • 0.576***
  • (0.081)
  • 0.499***
  • (0.038)

Higher degree (Y/N)

  • 2.113**
  • (0.797)

Employed (Y/N)

  • 0.123***
  • (0.012)
  • 0.079***
  • (0.003)

Constant

  • 10.648***
  • (0.627)
  • 0.757
  • (0.138)
  • 2.433***
  • (0.324)
  • 4.588***
  • (0.144)
  • 0.111***
  • (0.010)

Observations

77,776

162

4,845

77,776

43,084

Pseudo R:

0.694

0.018

0.140

0.706

0.026

Source. Own elaboration.

Notes. *** denotes significance at 0.01; ** denotes significance at 0.05. Results are reported in odds ratios. DY stands for a dependent variable. Y/N informs that the variable could take only two values: 1 - for Yes, 0 - for No. Reference categories are specified next to each variable. Robust standard errors are noted in parentheses.

So far, we have learned that returnees - especially those with less education - are less likely to be working than non-migrants but more likely to be self-employed. However, we know little about the reasons behind these patterns. These findings could imply that recent returnees are at a disadvantage in the Lithuanian labour market compared to non-migrants, or on the contrary, returnees might simply still be looking for jobs given that they returned within the previous year.

To assess returnees' labour market position more in depth, we limit the sample to the working population only and estimate the impact that being a returnee has on one’s wage. If we find that returnees earn less for the same jobs than nonmigrants, we could argue that recent returnees are at a disadvantage. As shown in Table 8.3, we indeed find a somewhat negative effect, although significant only at 10 per cent. The model has some limitations. Although we control for occupation, occupation is coded in nine broad categories (from managers to elementary occupations), but returnees and non-migrants might occupy different positions within these broad categories. It is likely that at the initial stages of return, mobile

Table S.3 OLS regression results regrading wages

(1)

Dependent Variable: Net salaiy

Returnee

  • -1.130*
  • (0.635)

Occupation

  • -0.280***
  • (0.012)

Age (ref.: 15-29-year-olds)

30-44-year-olds

  • 1.065***
  • (0.096)

45-59-year-olds

  • 1.226***
  • (0.097)

60+ year-olds

  • 0.781***
  • (0.115)

Females (ref. Males)

  • -1.764***
  • (0.050)

Marital Status (ref. Single)

Married

  • -0.015
  • (0.084)

Divorced, legally separated, or widowed

  • -0.110
  • (0.100)

Education (ref. Low)

Medium

  • 1.059***
  • (0.116)

High

  • 2.857***
  • (0.129)

Constant

  • 7.584***
  • (0.152)

Observations

35,811

R:

0.120

Source: Own elaboration.

Notes'. *** denotes significance at 0.01; ** denotes significance at 0.05; * denotes significance at 0.1. Occupation is coded in nine categories based on ILO categorisation (ILO, 2012), where 1 indicates managers and 9 stands for elementary occupations. Reference categories are specified next to each variable. Robust standard errors are noted in parentheses.

workers may occupy relatively lower positions within these categories than nonmigrants. Nevertheless, we know little about the further career progress of the returnees. Other research shows that over the years the effect of former migration dissipates or may turn into advantage (Barcevicius, 2016; Grabowska, 2016).

Given that circular migration has been observed in Lithuania in the past (Barcevicius & Zvalionyte, 2012), we attempt to address it in our research by looking at respondents who say that their current place of work is abroad. We indeed find stark differences between recent returnees and non-migrants regarding their place of work: in our sample, 0.3 per cent of non-migrants work abroad compared to 39 per cent of recent returnees.11 These statistics could help explain some of the employment patterns explored in this chapter. Nevertheless, we are careful not to overstate the impact of circular migration. Returnees who work abroad might work both in high-skilled and low-skilled occupations; they can work on a temporary or full-time basis (for example, remotely), so circular migration could have both an upward and downward pressure on their wages as well as an effect on employment rates.

Overall, our limited sample size does not allow us to explore what implications a foreign workplace has on returnees' employment patterns. We believe it is an important area of research to explore by introducing complementary methods and data samples, as it could provide insights into the effects of emigration and return migration on important questions, including the effects of migration on labour productivity and utilisation in the country.

 
Source
< Prev   CONTENTS   Source   Next >