Labour market assimilation of foreign workers in Italy - CiteSeerX

3 downloads 0 Views 571KB Size Report
The general pattern for foreign workers appears to be a fragmented career either being ... example, defining rules over entering and staying for non-European Union nationals. ...... substantial groups are from Tunisia, Egypt, and Ghana.

Labour market assimilation of foreign workers in Italy Alessandra Venturini1 (Università di Torino, IZA, CHILD, CARIM) Claudia Villosio (LABORatorio Riccardo Revelli, Collegio Carlo Alberto)

Abstract This is the first paper that analyses the labour market assimilation of foreign (i.e. non-citizen) workers in Italy. It considers the daily wages and the days of employment of male workers in WHIP, a matched employer-employee panel dataset, from 1990-2003. The traditional human capital approach is augmented by a control for the probability of staying abroad, modelled by aggregate variables of the origin country. The human capital variables considered are age and experience, both in and out of employment. What emerges from the empirical analysis is discouraging. Foreigners who are able to get higher wages are the least likely to stay, but assimilation profiles do not change when return migration is taken into account. Foreigners employed in the private sector earn the same wages as natives upon entrance into employment, but the two wage profiles diverge with on-the-job experience. Neither do foreigners assimilate from an employment perspective: a differential in employment between foreign and native workers is found even upon entrance, which increases over time. In the construction sector the wage and employment differential is even larger, while manufacturing and services follow the aggregate trend. Africans immigrants have the fewest career prospects while Eastern European and Asian workers are less far behind. The general pattern for foreign workers appears to be a fragmented career either being confined to seasonal or temporary jobs or alternating between legal and illegal employment. JEL:J61 Key words: Wage assimilation, Employment assimilation

1

Corresponding Author: alessandra .venturini @unito.it. We would like to thank the participants at the OXREP seminar, Steiner Strom, Simone Bertoli, Andrea Cornia, Daniela Del Boca, Robert Rowthorn and an anonymous referee for valuable suggestions.

1

I INTRODUCTION Immigration is not new to Italy. During the 1970s, it ceased to be a country of net emigration and became a country of net immigration. A number of other South European countries, including Spain, Portugal and Greece, had similar experiences soon after. For all of these countries, the transition into a country of net immigration has been difficult. Until the 1970s, there were ministries in charge of emigration issues but no ministries in charge of immigration. Public attention, the political debate and scientific research focused upon the effects of immigrants on wages and employment in the native labour markets and upon matters of law – for example, defining rules over entering and staying for non-European Union nationals. Although empirical research showed that migrants in general do not compete with natives in the labour market (e.g. Dolado et al 1996, Bentolilla et al 2007, Amuedo-Dorante and de la Rica 2005, Gonzales and Ortega 2007, Gavosto et al 1999, Venturini 1999, Venturini and Villosio 2006, Carrington and Lima 1996)), migration was a highly sensitive issue in electoral debates, with each new government revising existing immigration law. It took time to adjust to this new and pervasive phenomenon and to cope with the internal demand for foreign labour and the sudden increase in supply of immigrants. Thus, in addition to the legal entrance of foreign workers, the Southern European countries implemented successive amnesties to legalise foreigners already working illegally (without a residence permit) and/or irregularly (without a work permit) in the country. Immigration increased very rapidly and in less than twenty years the total immigrant population rose from a few thousand to three and half million in Italy, one and half million in Greece and Portugal, and between three and six million in Spain, representing percentages of the total population ranging from 5% in Italy and Portugal to 7% in Spain and Greece (Munz 2008). Concern about the increasing pressure of immigrants on the welfare state directed attention toward foreigners’ working careers and assimilation paths, the underlying idea being that the more they assimilate, the less they need rely on welfare. Both social and economic assimilation research is relatively new in Italy (e.g. Ambrosini 2001, Reyneri 2003). This is the first paper to use a large dataset to answer the question: are migrants assimilating in wage and employment outcomes, relative to natives? In other words, even if migrants upon arrival are unable to use all their human capital in the labour market, is it true that the longer their presence in the host labour market, the more their wage and employment profiles become like those of native workers? The main body of the human capital assimilation theory was built upon the American experience (Chiswich 1978, Borjas 1985, Oaxaca 1973) and was imported from the gender literature into the migration literature, with the underlying assumption that immigrants remain in the destination country at least until they retire. In this framework, foreigners under- (over-) assimilate if they earn less (more) than natives with the same characteristics after a number of years in the host country. The reason they do not assimilate can be linked to poor or unsuitable skills upon arrival (Borjas 1985); arrival during an economic downturn (LaLonde and Topel, 1992, Rosholm, Scott and Husted 2000); or discrimination. Researchers also focused on the role played by the migrant’s community in favouring or hampering migrants’ economic integration (Borjas 1992, Hatton and Leigh 2007); on the importance of language skills which determine the abilities of workers of different nationalities to obtain an appropriate job and wage(Chiswich 1991, Dustmann and Van Soet 2002, Dustmann and Frabbri 2003, Clark and Drinkwater 2002), and on the role of integration policies in favouring economic autonomy (Stalker 1994, Salt, Singleton, and Hogarth 1994)2 Close attention has been 2

This field of research compares different policies as income support interventions, languages and training courses but also groups of migrants which differ according to their nationality or type of entrance, i.e. if they are political refugees, illegal migrants or workers.

2

paid to the selectivity problem (Borjas 1987, Borjas and Bratsberg 1996): that if migrants are the best of their native population, they are likely to out-perform the natives of their destination country. More recently the length of stay abroad has been investigated and much evidence has been produced on the non-permanent duration of migration, particular with European migration. The large-scale emigration from Southern European countries - Italy, Spain, Portugal and Greece started with repeated temporary stays in Northern European countries, which became more permanent after the 1973 recession, leading to the adoption in all host countries of a more restrictive immigration policy. But this pattern was also widespread for Asian migrants, – for example, Chinese workers (Chew & Liu, 2004) or Malaysian workers (Wickramasekara P. 2002)– as well as for migrants from New Zealand (Lidgard and Gilson, 2002). The most recent empirical research shows that a large share of migrants are not permanent (Proceedings on Circular Migration, CARIM, 2008) – in Germany, for instance, 60% of migrants were not permanent stayers (Constant and Zimmermann, 2007) – and return migration is becoming a very promising area for future research. In this framework, the work of Dustmann (1997, 2000, 2003) on the duration of stay abroad and the decision to return by migrants is particularly timely, as it revises conventional assumptions by introducing into the assimilation analysis a control for the probability of staying in the destination country. This paper, then, analyses the wage and employment assimilation patterns of foreigners in Italy, controlling for the probability of staying abroad. The analysis is based only on legal foreign workers who are free to return home when they decide to. Irregular migrants (migrants without work permits) are not included because no data on them is available, but also because their status conditions their length of stay until the arrival of an amnesty, which frequently forces them to stay longer than intended. We use a panel of administrative matched firm-worker individual data covering years from 1990 to 2003, and the available results for the male group alone show that foreigners do not assimilate, either in wage or in employment. The control for a selective return shows that the best foreign workers do not remain, but performance is very similar among both temporary and permanent . The construction sector demonstrates the least assimilation. Among the main ethnic groups, Africans have the fewest career prospects. The picture which emerges is not unexpected: the Italian labour market offers foreign workers only unskilled jobs, with frequent spells working in the underground economy, which damages their present and future lifetime incomes. The paper is organised as follows. Section two describes the evolution and characteristics of immigration in Italy. Section three sets out the model. Section four describes the dataset and the variables used, and section five presents the results. The conclusion compares these results with similar immigration patterns experienced by the other Southern European countries.

II THE CASE OF ITALY Italy became a popular destination for economic migrants at the end of the 1970s after the first oil shock, when Northern European countries adopted more restrictive immigration policies which made the entry of foreigners into their labour markets more difficult. In 1980 the foreigners resident in the country represented only 0.8% of the total population. The inflows to Italy and in general to the Southern European countries became more important later on: in 1990 foreign residents comprised 1.3% of the population. In the 2001 Census this figure had

3

risen to 2.3%; in 2003 it had risen again to 2.7%. The most recent figures, for January 2008, show that there 3.5 million foreign-born residents in Italy, comprised 5.8% of the total population. Foreigners frequently enter Italy illegally, and their subsequent legal status is often the result of the many amnesties that the Italian Government has granted. The first was in 1987-88 (118 foreigners legalised), the second in 1990 (217.000 foreigners legalised), the third in 1996 (250.000 foreigners legalised), the fourth in 1998 (218.000 foreigners legalised), and the last in 2002, which legalised the positions of three times more people (650.000) than previous amnesties Although the government establishes a quota of work permits each year, the number has always been insufficient to satisfy the potential supply of immigrants, but also insufficient to satisfy the national demand for immigrants. The repetition of amnesties – a policy, however, shared with the other Southern European countries – has created the expectation of further ones, and augmented the difficulties in controlling immigration3. Thus, illegal entry has become the main gateway into the country and the amnesties have been granted ex-post in order to regularise the irregular job positions of many foreigners. In fact, in order to become legal, migrants must show a regular job offer, and to obtain one, in general they have worked illegally for at least a couple of years. Initially, the flows came mainly from both neighbouring areas, such as North Africa, and from further away, including Asia (mainly the Philippines) and Latin America. With the fall of the Berlin Wall the inflows from the Eastern European countries began. Initially, the migrants came from neighbouring Albania, but later they also began to arrive from Romania and Moldova, attracted by the similarity of the language, and also from faraway Ukraine (See Figure 1). The information on foreign citizens provided by residency permits is the most accurate available, even if it records only the stock of legal immigrants. Figure 1 shows the foreign population divided by areas of origin, which increases over time, and more rapidly with each amnesty. The 1991-2 break indicates the revision made by ISTAT on the administrative data collected by the Ministry of the Interior, which since then has been cleaned of expired permits. The estimates on illegal immigrants are frequently derived from the number of workers who apply for legalisation, but they have been proven to be largely wrong. The spread of the estimates is always between 10% and 40% of legal workers (Strozza and Venturini, 2002).

3

See, for instance Venturini, 2004, chap.5.

4

Figure 1 Stock of resident permits for foreigners by main areas of origin 2000000 revisions in the residence permits series

1800000 1600000 1400000 1200000 1000000

EST EUROPE

800000 600000

AFRICA

400000 ASIA

200000

2003

2002

2001

2000

1999

1997

1996

1995

1994

1993

1992

1991

1990

1989

1988

1987

1986

1985

1984

1983

1982

1981

1980

1979

1978

1977

1976

1975

1974

1973

1972

1971

1970

1998

LATIN AMERICA

0

Source: ISTAT At the 2001 Census, 61.8% of foreign residents were resident in the North, whilst only 25% were resident in the Centre and 13.2% in the South and Islands, and the shares remained similar thereafter (confirmed by residency permits distribution in 2007)4. The male and female groups are on average quite balanced but exhibit large differences by country of origin, with males dominating among Africans (80% of males 2007) and females among the Asians (55% Filipinos, while among the Chinese males amount to 68%) and Latin Americans (59% Ecuadorian). Among Eastern European migrants, the variation is even larger, with large female inflows from Ukraine (78% female), balanced inflows from Romania (45% female), and a much higher proportion of males among Albanians (74%). According to the most recent ISTAT Labour Force Survey (2006) - which unfortunately under-reports the size of the phenomenon, in particular with regards to the most recent arrivals – there are 1,475,000 foreign workers in Italy, which represents 6% of the total labour supply. They are more concentrated in the North and in the Centre, where they represent respectively 7.3% and 7.1% of total employment, whilst they are less important in the Southern regions (2.4% of total legal employment). The immigrants arriving in Italy are not highly skilled, nor do they have high education levels, because the Italian labour market has high demand only for unskilled positions. According to the LFS source, their level of education is lower than the native one: 51.1% of immigrants have only compulsory education against 39.7% among natives, 40% have an upper-secondary diploma (45.2% for the natives) and 8.8% a university degree (15.1% for the natives). They also generally work in unskilled positions even if, in a few cases, they have higher educational qualifications (31% have unskilled jobs and 45% blue collar jobs, whilst among the natives the percentages are 9% and 27% respectively). Men usually work in the construction sector, in agriculture and the manufacturing industry, whereas woman work in family services and services in general; quite a few also work in industrial activities. 19% of foreign employment is in family services, of which 87% (based on Social Security data) is female. Unfortunately, very little is known about family services. The Social Security archive reports only the total number and nationalities of foreigners, and the few sociological studies to conduct local interviews with migrant Italy does not have a tradition of large European inflows as do, for instance, Greece and Spain, and at the beginning of the 1970s the majority of foreigners were located in Rome, where communities from the former colonies were based (Ethiopia and Eritrea). 4

5

workers provide a deeper but partial analysis. All the empirical economic research disregards this growing segment of the labour market. III THE ASSIMILATION MODEL IN THE PRESENCE OF RETURN MIGRATION The model used is the same for both wage and employment assimilation. It extends the classic model of wage assimilation (Chiswick, 1978) by explicitly including measures of human capital acquired on the job and out of the job (as for instance in Husted et al 2001) and controls for selection in return migration. The dependent variable is a measure of the individual labour market outcome [y] which depends on individual fixed effects [αi], individual time variant social and human capital variables [xit], worker’s job characteristics [zit] and time invariant individual characteristics [hi]. Moreover, it is influenced by fluctuations in the economic cycle [mrst] which affect both the region [r] and the sector [s] where the workers are employed and can be influenced also by the size of the migrant’s community [c] in the destination area [kcrt]. The literature on assimilation stresses the importance of the community to which the migrant belongs (Cutler and Glaeser, 1997). The effect of the community can be positive by favouring the job search process, but it can be also negative by reducing the social integration-interaction of immigrants, namely knowledge of the host country’s language (see e.g. Chiswich 1991, Dustmann, van Soet 2002, Shields, Wheatley Price, 2002), and in general of the informal rules on living which prevail in the destination countries. The process of assimilation, however, also depends on the characteristics of immigrants who remain in the destination country. As Borjas (1987), Borjas and Bratsberg (1996) and Dustmann (2001) stress in their articles on the return decisions of migrants, foreigners who remain may be either the best or the worst of the group5. The migrant decides to return if the migration project fails or in the opposite case, if the migration project is very successful and allows the migrant to go back home and start a business activity there. If those who remain are the best, the empirical estimates will be biased upwards (over-assimilation), while if the ones who remain are the worst, the estimates will be biased downwards (under-assimilation); in both cases they are inconsistent. If there is a systematic link between the decision to stay and labour market outcomes, a fixed effect estimate eliminates the bias. If it is not systematic, even fixed effect estimates give unreliable parameter estimates. The decision to stay in the destination country (or as Dustmann calls it, the “decision to remigrate”), Pr(Sit), is modelled in the framework of the emigration decision. In this framework, some authors (i.e. Dustmann (2001)) focused upon the role of income in the destination country on the individual decision to stay, and proxied the income in the country of origin with the migrant education level. However, recent research on return migration and return migration policies; e.g. Cassarino (2007) for the Maghreb areas and Mansoor and Quillin (2006) for many European and Central Asian countries, stresses instead the crucial role played by economic prospects in the countries of origin in attracting migrants back home. Thus, we model the migration choice as a function of the wage differentials between the two areas but for a given wage differential, the level of income in the country of origin plays a crucial independent role, as in Faini and Venturini, (1993, 2008). Given that individuals receive higher utility by consuming at home than abroad, as the income of the origin country increases, for any given wage differential, the level of wellbeing for any potential migrant grows. As income increases it reaches a level at which the potential migrant prefers to consume the normal good 5

Since his 1987 article Borjas stressed the selectivity of the migration decision as a function of the human capital return of migration. In his 1996 article with Bratsberg he also considers the selectivity of the return decision in a cross-sectional approach and always referring to a Roy return of human capital model.

6

“staying at home” and not migrate. Extending this result, higher income at home favours return, while low income at home discourages it. We model the probability of staying in the host country Pr(Sit), as in Dustmann (2001) and Fertig and Schurer (2007), with individual variables which are common to the assimilation equation and the return decision Xit and additional variables which affect only the re-migration decision and which serve as identification restrictions Zct, plus an individual component randomly distributed among the individuals. Pr(Sit=1) = Φ (βXit + µ Zct + υi), where the random effect υi has a normal distribution N (0, συ2)

(I)

The Heckman error correction (λit Inverted Mill Ratio) obtained from (I) is introduced in the assimilation equation which results (II):

yit = α i + xit β + zit δ + hi γ + k crt ξ + mrstη + λi t ω + τ t + ϕ r + φ s + ε it

(II)

Where [τt ] are time fixed effects, [ r] are region fixed effects, [ s]are sector fixed effects and [εit] an idiosyncratic error component. IV DESCRIPTION OF THE DATA AND VARIABLES USED The previously mentioned datasets either provide very limited information (residency permits and registration at the local municipalities or Census) or are available only for the most recent years and do not include information on wages (as the Italian LFS). They are, therefore, not suited for analysis of the wage and employment assimilation. We use, thus, the Work Histories Italian Panel (WHIP), a dataset randomly selected from all Italian Social Security Administration (INPS) archives, i.e. from the population – Italian and foreign - of those who have worked in Italy as employees or in self-employment, or have received income support or pensions by INPS even if for only part of their working careers. This is a panel with no attrition because it is derived from administrative information which the firm is obliged to provide. In this sample, entire working careers are observed. The period covered by WHIP is from 1985 to 20036. Workers in the public sector and selected professions (e.g. lawyers or notaries) are excluded. This limitation is not very relevant for foreigners, who are rarely employed in the public sector, where hiring is made more complex by the presence of national public competitive examinations which take place occasionally; nor do they belong to the professions, which usually employ natives. Only the dependent employment section of WHIP, which is a Linked Employer Employee Database, is used. Thus, in addition to information about individual and job characteristics, data on the characteristics of the firm in which workers are employed is also available. Since we focus on dependent employment in the private sector (79% of total Italian dependent employment), then besides public employment and self-employment, workers in the agricultural sector and housekeepers are also excluded. This last limitation is very important for an analysis of foreign labour market integration because a large share of immigrants work in these two sectors: in 2007, 3.8% worked in agriculture and 20% in domestic services, with a rapid increase from the 12% of 2002.

6

More information can be found at www.laboratoriorevelli.it/whip.

7

In particular, given the female monopoly of family services and its rapid growth, the share of females in the total number of residence or work permits is much larger (about 40-45%) than the figure reported in the WHIP dataset, where male employment dominates (84%). Agriculture and domestic work have, however, a very high share of illegal employment, and their forced exclusion from our analysis nevertheless increases homogeneity among the sectors. Foreign workers have been selected by place of birth (WHIP data does not contain information on nationality). Only workers born outside Europe and the main industrialised countries have been chosen, in order to avoid counting Italians born abroad as immigrants. Moreover, following Natale, Casacchia and Strozza (1999), we have also excluded workers born in Argentina, Brazil and Venezuela, because these countries have in the past received a large number of Italian immigrants, many of whom have returned to Italy in more recent years. The period analysed has been restricted to 1990-2003, when the largest inflows of foreigners began (see figure 1). The number of foreign workers present in the dataset in the 1980s was too small, so that the results could have been affected by measurement errors. Moreover, women have been excluded because, as already mentioned, our dataset does not include employment in the public sector, where a large share of native female workers are employed, nor family services, where a large share of female immigrants are employed. Finally, in order to compare foreigners with the most homogeneous group of Italian workers, we concentrate on full-time (male) workers in the 1845 age group. Assimilation in the labour market is evaluated using as labour market outcome [y] the log daily wage (wage assimilation) and the number of days worked per year (employment assimilation) of individual [i], in year [t]. This last variable has been computed in terms of full-time equivalent in the case of jobs starting during the course of the year, while no corrections are made for the wage variable since we focus only on full-time workers. The combination of these two measures gives the total annual earnings of individuals, which can be considered the aggregate measure of economic integration in the Italian labour market. Thus, the assimilation of foreign worker is evaluated both in terms of wages and quantity of work (for more information see Contini et al., 2000). The social and human capital variables [xit] included are age, number of months in employment, and not in employment. These last two variables have been built in order to capture the increase in human and/or social capital. The former captures human capital accumulated on the job – experience from current and previous jobs – and it is measured by the number of months in regular employment7. The latter measures the months spent “out of the job” which could be, according to the preferences of the two groups of workers, devoted to education, employment in sectors not covered by the dataset, unemployment, return back home for a while, or irregular employment. This second variable could thus play either a positive role in the labour market, by capturing the acquisition of human capital in the underground economy or the acquisition of more general social capital out of the labour market, or a negative role by reducing human capital accumulation. We expect this variable to play a different role in the employment performance of natives and foreigners. For natives, the periods of non-employment are expected to affect wages and number of days worked negatively, while for foreigners this negative impact should be lower since a longer presence in the host country, even if not in employment, could positively affect the migrant’s social capital, which has a positive return in the labour market.

Since the dataset starts in 1985, this variable is left-truncated for workers who joined employment before 1985. This happens less often for foreigners who mainly entered Italian employment after 1987. 7

8

In the case of employment assimilation (number of days worked per year), both experience on the job and out of the job refer to the previous years. Unfortunately, the dataset does not include any education variables. Whilst for natives the combination of age and skill level could proxy for education level, these two variables are not as satisfactory for foreigners. However, the number of years spent at school would not be a good proxy for the level of education or the level of productivity of a foreign worker because the quality of education differs considerably among countries, and in addition the education that is ready to use in the host country depends on proficiency in the language used in the host country. Controls for job characteristics of the worker [zit] include skill level (occupation) and type of contract, firm size, firm location and sector; while the time invariant individual characteristics [hi] included are country of origin and year of arrival. Fluctuations in the economic cycle [mrst] are controlled by the unemployment rate and the (log) value added. The unemployment rate is unfortunately available only at the regional level (from ISTAT, Labour Force Survey), while the value added is available at sector and region jointly (from ISTAT Regional Economic Accounts), and although it is not at firm level it strictly follows the fluctuation of local production and, thus, worker productivity. Two variables have been built to proxy the community of the migrant in the destination area [kcrt ]: the first is the share of the migrant worker community (nationality) on the regional total employment to proxy the size of the community itself, and the second is the share of the macro areas of origin (Africa, Asia, Eastern Europe, Latin America) on total branch and region employment. Both variables are problematic because they refer only to workers and not to the entire population of people of the same nationality. In addition they are contemporary the worker’s employment status. Hatton and Leigh (2007) stress the long-term effect of the community variable, and even use 10-year lags. Unfortunately, the migration phenomenon in Italy is quite recent, and we have too few observations to use long lags such the one suggested by Hatton and Leigh. For more detailed information on the variables see Appendix Table A1 The probability of staying Pr(Sit) is a function of individual variables (age at entry and years of stay (Xit)) and the annual GNP growth rate and level in sending countries, which proxy the average probability of finding a job and the effect of earning level in the country of origin on the probability to return (Zct), as well as dummies for sector of employment and main areas of origin8. Thus, contrary to the literature on return migration, which uses the family ties variables that is unavailable in our dataset for identification, we use as instruments the GNP variables because they affect the decision to stay in the host country but not the foreigners’ labour market outcomes in the host country and are valid exclusion restrictions. The individual component is strengthened by the use of a Random Effect model which includes an individual effect in the re-migration decision, randomly distributed among the individuals. A test on the correlation between errors in the wage equation and in the probability of staying rejects the null hypothesis that they are uncorrelated, confirming the validity of our approach (chi(2)= 628.56; Prob>chi(2)= 0.00). The first sections of Table 1 show the variables of interest for the assimilation analysis. On average, foreign workers earn 21% less than natives and work 20% less than natives. These differences are in large part due to the different characteristics between the two groups. Foreign workers are 8

An additional specification with the origin GNP interacted with dummies for the year of entrance has been implemented, which provides similar results, thus has not been reported.

9

younger than natives and have a shorter period of presence in the labour market, both in terms of employment and non-employment, and are employed in low skill positions (93% are blue collars). Foreign workers are concentrated in the small firms of the North, in blue-collar occupations and in the construction industry. The prevalence of foreigners from Africa declined during the period (in 2003 they were only 39%), and was contrasted by workers from Eastern European countries, which is the group that has grown most rapidly in recent years (in 2003 it reached 37%). Moreover, the data shows that average job tenure with the same employer for foreigners is much smaller than for natives; the estimated probability of a (legal) job tenure up to 5 years is only 27% against 55.7% for natives, while the foreigner estimated probability of staying (legally) in Italy up to 5 years is 54%, stressing both a lower (legal) participation in employment and, on average, shortterm migration projects. Table 1 – Descriptive statistics for foreigners and natives average 1990-2003 Variable

Daily wage Days worked Age Tenure (in months) with the same employer Total Months of employment Total Months out of employment Blue collar White collar Apprentices Atypical Firm size 0 20 Firm size 20 200 Firm size 200 1000 Firm size over 1000 North west North east Centre South Manufacturing Construction Services Africa Latin america Asia East europe

Foreigners Mean (SD)

47.39 240 33

Natives Mean (SD)

(16.7) (89) (8.7)

57.34 287 37

20.73

(25.4)

52.48

(50.3)

43.0 8.0 0.93 0.04 0.03 0.14 0.58 0.30 0.08 0.05 0.39 0.37 0.18 0.06 0.52 0.20 0.28 0.52 0.03 0.16 0.29

(38.7) (16.7) (0.26) (0.20) (0.17) (0.35) (0.49) (0.46) (0.27) (0.22) (0.49) (0.48) (0.38) (0.24) (0.50) (0.40) (0.45) (0.50) (0.17) (0.37) (0.45)

93.8 13.1 0.66 0.29 0.04 0.09 0.42 0.29 0.13 0.16 0.35 0.24 0.18 0.23 0.52 0.13 0.35

(54.6) (25.0) (0.47) (0.45) (0.20) (0.29) (0.49) (0.45) (0.34) (0.37) (0.48) (0.43) (0.38) (0.42) (0.50) (0.34) (0.48)

Share of foreign employment by nationality 0.63% (0.6%) and region Estimated probability of a job lasting up to: 1 year 70.6% 27.0% 5 years 9.9% 10 years Estimated Probability of staying in Italy up 78.9%

10

82.7% 55.7% 35.3%

(66.9) (50) (10.5)

T statistics for differences in means 62.06 91.84 22.25 268.7 147.16 37.26 -107.75 110.92 1.88 -16.72 -49.17 -3.96 26.7 50.63 -20.76 -58.99 1.7 80.57 -1.07 -38.57 32.96

to: 1 year 5 years 10 years Total 1990-2003 sample

53.7% 35.6% 33622

420205

V ESTIMATES AND RESULTS (i) Wage assimilation The Italian labour market is quite closely regulated and national agreements cover all workers: trade-union and non trade-union members. Thus, we do not expect large differences in the coefficients of the wage equation among the two groups. However, the 1993 Income Policy Agreement introduced a more decentralised bargaining system, which allows firms to adjust their wage structure according to their economic performance and to local labour market conditions. This raises the average share of top-up components on total wages to about 22% (Devincenti et al. 2008) and, thus, increases wage variability. The results (reported in Table C1 in the appendix) have been obtained by an OLS fixed effects estimates to control for unobserved heterogeneity among individuals The two wage equations for natives and foreigners are statistically different from one another: a test for common coefficient restrictions on a pooled regression rejects the null hypothesis that all the coefficients for foreigners are zero [F(11,404210) = 8.91 ]. The human capital variables differ for the two groups of workers. The return on age is higher among nationals than among foreigners, probably because it also reflects the returns on education, for which we are unable to appropriately control. Instead, the return to experience on the job is almost the same in the two groups, but natives have a more convex profile than foreigners (the square term is higher for natives than for foreigners). However, periods spent out of employment have a different return: for natives, this decreases their earning capacity, while for foreigners it does not harm their wage career (the variable is not significant). For both groups, periods out of employment mainly represent unemployment or employment in the black economy, but for foreigners they may also represent periods spent in their country of origin. Whatever the case, spending time out of employment does not have the negative effect on foreigners’ wages that it does for natives, either because they acquire additional human capital or because the jobs in which they are employed are very unskilled so that they are “punished” less by unemployment compared to natives. This result is not new: for instance Husted et al (2001) finds the same for labour migrants in Denmark. Turning to the selectivity control for the probability of staying, the negative sign of the IMR coefficient indicates a negative covariance between the error terms in the return decision and wage function. The un-observables that positively influence the worker’s earning also negatively affect the decision to stay in the host country (results of the probability of stay in Appendix in table B1 and variables description in Table A2). This implies that, other things being equal and, thus, for a given human capital level, the workers who earn more have a greater probability of leaving their present employment, either because they become self-employed as they have saved enough and want to return home, or because, given the limited wage opportunities available, they want to move to other countries, whilst the less “able” workers tend to remain in employment in Italy. Nevertheless, the coefficients of the wage equation 11

are not statistically different when return migration is taken into account. This result is, however, common to other research, (similar results are found by Fertig and Shurer (2007), for immigrants who arrived in Germany between 1969 and 1973, and by Constant and Massey (2003), for guestworkers in Germany in 1984). This could be because the subgroup of return migrants behaves like the stayers in terms of wage outcome. Figure 2 summarises the results of the assimilation model with respect to the effect of experience in the labour market on the wage profiles of foreigners and natives for the reference person (blue collar, employed in a small manufacturing firm in the North) upon entrance into the labour market. It shows that migrants start at the same level as natives, but since returns on age which proxies education are higher for natives, the two profiles diverge with time spent in the labour market. This result, however, is mitigated by the non-penalising effect for immigrants of periods of nonemployment which tend reduce natives’ wages. Whilst the average unconditional log wage differential throughout the period is 19%, there is no wage differential on entrance when differences in characteristics are taken into account, but after five years of experience the conditional log wage differential reaches 12% and after ten years, 19%. Figure 2 Effect of increasing labour market experience on the log wage profiles for foreigners and natives at entrance in the labour market

Log daily wage

3.50

3.00

2.50

2.00 0

1

2

3

4

5

6

7

8

9

10

11

12

13

Years Foreigners

Foreigners+ corr. return migr.

Natives

The macro variables play a different role in the two groups’ wage functions: whilst the value added variable has the expected positive sign for both groups, regional unemployment has a negative impact only on the natives’ wages. Even if the unemployment rate according to ISTAT (2007) is a little higher among foreigners (8.6%) than among natives (6.7%), migrants represent only 6% of the total labour supply and so unemployment is mainly composed of natives, who are not likely to compete for the same jobs. The community variable is not significant in any form (linear, log, share) or in different specifications and, contrary to expectations, does not yet seem to play any role in the foreign wage

12

equation, probably because the phenomenon is too recent, as the variable is not lagged as in Hatton and Leigh (2007) or because it is too small and composed only of employed workers. (ii) Employment assimilation Foreign workers are more mobile and less likely to be in stable jobs than native workers, as Table 1 and previous research (Venturini and Villosio 2000) shows. This section tests whether there is a differential between the two groups of workers in employment attachment and the effect of human capital variables on the process of employment assimilation. The dependent variable used in the employment assimilation equation is the number of days worked in a given year. This variable has also been used by other authors in the empirical analysis of employment assimilation (see Chiswick and Hurst 2000, Pichè 2002). Unfortunately, since our data has information only on employed individuals, it does not allow for analysis on unemployment probabilities which are more commonly found in the literature. The dependent variable assumes the value between one and 312. Fig.3 Effect of increasing labour market experience on the number of days worked in a year for foreigners and natives at entrance in the labour market 350

N.of days worked

300 250 200 150 100 0

1

2

3

4

5

6

7

8

9

10

11

12

13

Years Foreigners

Foreigners + corr. return mig.

Natives

The results (in the Appendix Table C2) show that the selectivity variable is not significant and there are no significant differences between the specification with and without it, as in the wage case (see Figure 3). The returns on age are slightly higher among natives than among foreigners. Experience on the job is lower for natives but declines at a lower (less convex) pace and previous periods spent out of legal employment have a negative impact on the number of days worked only for domestic labour. Thus, there is catch-up by foreigners due to experience which is, however, over-taken by the age effect (which, as previously mentioned, may also capture the missing education variable). The cumulative effect is that foreigners start at lower levels of employment and never catch up with natives, who reach a full-year employment level in few years, while foreigners do not. Days of employment by foreigners are less sensitive to the economic cycle than those of natives; the value added is positive only in the natives’ case, while the unemployment variable is never significant. The community variable, as earlier in the wage assimilation case, is never significant. 13

On average, the differential between natives and immigrants in the number of days worked in a year is about 20%. The estimates show that the differential on entrance for similar workers is around 10%. As time spent in the labour market increases this differential also increases; after 5 years of experience it reaches 11.2% and after 10 years, 12.8%. The picture that emerges is one of fragmented careers for foreign workers who do not have the job stability of natives and who are confined to more seasonal or short-term jobs or to alternating between legal and illegal employment. This evidence is in line both with the recent “temporary migration” literature which stresses the limited duration of migration made by repeated stays abroad, and with the legal-illegal work pattern of foreign workers in the Southern European countries (Reyneri 2007). This pattern may partly reflect the fact that large-scale illegal entry into the country has made only irregular jobs available to foreigners, which are more flexible than legal ones and sometime more rewarding. Foreign workers who become legal with a regular job might return to their previous illegal job, in part because they do not have a clear intention to settle. This is the first empirical research on the assimilation of immigrants into the Italian labour market. Hence, we cannot compare our results with those of other economic studies and we cannot derive very strong implications because they are driven from a 13 years’ panel analysis. However, one implication of the presence of these differentials between workers with similar characteristics in the labour market from entry onwards is that foreign workers, more so than natives, suffer from the employers’ practice of avoiding social security payments as a measure to reduce their labour costs. If this is the case, foreign workers are not only discriminated against at entry and in their current and future jobs, but also in their life cycle prospects in retirement, in either the host or originating country. As Ambrosini (2001) suggests, immigrants are needed for heavy, precarious, dangerous, low paid and socially penalized jobs. (iii) Wage and employment assimilation by sector and ethnic groups The overall picture is the result of the behaviour of immigrants from different countries of origin, who arrived in Italy at different periods of time (see Figure 1), and who are employed in different sectors. As in most host countries, in Italy immigrants from the same origin tend to concentrate in the same labour market sectors, partly because of heavy reliance on social networks as a means of gaining entrance into a sector, and partly because co-ethnic immigrants possess similar skills, experiences, or other attributes that favour their sector specialisation. In addition, employers may also prefer employing workers who share a common language and culture to increase productivity and social consensus. Unfortunately, the dataset is not large enough to distinguish by single national groups of workers, or by detailed sectors; however, by just taking the areas of origin and the three mains sectors, a different picture emerges. Table 4. Distribution of foreign employment by sectors and by ethnic groups. Average values in the 1990-2003 period Ethnic groups Africans Latin Americans Asians Eastern Europeans

Sectors (row %) Manufacturing Construction Services Total 55.4 34.8 48.1 35.5

17.2 21.5 8.1 39.7

27.4 43.7 43.8 24.8

14

17619 932 5551 9520

Sectors (col %) Manufacturing Construction Services 60.5 2.0 16.6 20.9

40.7 2.7 6.0 50.7

48.1 4.1 24.3 23.6

Total

16137

7466

10019

Table 4 shows that even at the aggregate level, sector specialisation by ethnic groups can be detected: the majority of immigrants from Africa are employed in the manufacturing sector; Latin Americans and immigrants from Asia are mainly employed in the services sector and a large share of Eastern European immigrants work in construction. Among the African migrants, Moroccans and workers from Senegal are the largest groups, but other substantial groups are from Tunisia, Egypt, and Ghana. The Eastern European group comprises Albanians, Romanians, and ex-Yugoslavs. For the Asian group, which comprises workers from India, Bangladesh and China, recall that we consider only the subgroup of workers employed in private firms and not the larger Asian community, which is mainly comprised of females employed in family services. Similarly, the Latin American group in our dataset is not representative of the large Latin American community present in Italy, which is mainly concentrated in family services and in addition is very small (3% of total migration workers) and thus is not considered further. On a sector-by-sector basis, foreign workers employed in manufacturing are mainly concentrated in the metal products sector and in the textile industries, while those in the Services sector are mostly employed in hotels and restaurants and in the commerce sector. To investigate possible different immigration projects and thus, different assimilation patterns, we interact the human capital variables with both area of origin and macro sector of employment variables . Table 5. Average daily wage and days worked by ethnic groups and by sectors of employment. Average values in the 1990-2003 period Daily_wage Foreigners Africans Asians Eastern Europeans Manufacturing Construction Services Natives Manufacturing Construction Services

Days_worked

Mean

Sd

mean

sd

47.01 47.80 47.62

(16.35) (19.02) (15.22)

240 243 236

(89.3) (88.5) (89.9)

49.37 45.45 45.49

(16.25) (12.58) (19.56)

258 212 231

(78.7) (93.5) (95.8)

58.05 46.26 59.85

(25.71) (17.35) (30.92)

292 249 282

(44.7) (66.4) (60.0)

Differences between the variables of interest across ethnic groups are not large. Greater variability is found in wages and worked days across sectors, with construction having the lowest number of average days worked in a year. Discriminating by sector, it is clear that the construction sector is an outlier, especially for immigrants, while manufacturing and, to a lesser extent, services are similar to aggregate patterns. The sector most exposed to seasonality and variability is construction and as such, has the highest turnover in employment. Moreover, it is the sector with the highest incidence of illegal work (in the aggregate economy the estimates are 11.3% (see ISTAT 2008)). The estimates show that wage and employment growth in this sector is lower than in the others for both natives and immigrants, but is higher for immigrants relative to natives. 15

The difference between natives and immigrants in wages and days worked is very similar across sectors at the beginning of the working careers, but after few years in the construction sector the differential is much larger than in the other two sectors, especially for wages: after 5 years the log wage differential in construction reaches 18% (with respect to 11% and 10% in manufacturing and services), and the differential in the days worked is 13% (8% in manufacturing and services). The result on wages is mitigated, however, by the positive impact of experience gained out of the job, most likely through experience in the underground economy, which contributes towards an individual’s human capital and hence increases wage. Immigrants’ wage and employment dynamics are very similar in manufacturing and services at the beginning; however in the medium run (after 7-8 years) differentials increase by more in services, which is a more dynamic sector, while manufacturing appears to be more stable and have a compressed wage distribution. This result is also in line with the negative sign of the selectivity control in the return decision, found only in manufacturing, which supports the hypothesis that the best workers leave. Across ethnic groups, the selectivity coefficient is negative, with the expected negative sign in the wage equation for all groups, supporting the interpretation that the workers who are able to get higher wages are the least likely to stay. For Eastern Europeans, this because they are likely to return to their home country or to migrate to higher wage areas; among Asians and the Africans, because they go home or to start their own business. The human capital variables play the usual role, with age being most important for Asians and Eastern Europeans. Again, this probably proxies an education component that is missing in the data (but is always lower than for natives). Experience on the job is not significant for Eastern Europeans, who are the more recent migrants (26 months in employment against the average 43); Furthermore, for this group, which most closely resembles the native workers, periods out of the job have a negative effect. Regarding employment assimilation, the selectivity variable is not significant in any group, as with the aggregate case. For Asian immigrants, the length of previous experience does not impact on the current number of days worked, which is in line with what happens in the service sector, where they are most employed. Moreover, this group, which is the only one showing a negative impact on days worked of non-employment periods, seems to be trapped in illegal employment more so than the other groups. Figure 4. Foreign-native differentials in wages and days worked by sectors at increasing experience in the labour market Differential in days worked (%) 0.20

0.35

0.18 Differential

Differential

Log wage differential 0.45

0.25 0.15 0.05

0.15 0.13 0.10 0.08 0.05

-0.05 0

1

2

3

4

5

Manufacturing

6 7 Years

8

9

Construction

10

11

12

13

0

1

2

3

4

5

6

7

8

9

10

11

Years

Services

Manufacturing

Construction

Figure 5. Foreign-native differentials in wages and days worked by ethnic groups at increasing experience in the labour market 16

Services

12

13

Differential in days worked (%)

0.25

0.15

0.20

0.13 Differential

Differential

Log wage differential

0.15 0.10 0.05

0.10 0.08 0.05 0.03 0.00

0.00 0

1

2

3

4

5

6

7

8

9

10

11

12

0

13

1

2

3

4

5

Africans

Asians

6

7

8

9

10

11

12

13

Years

Years

Africans

East Europeans

Asians

East Europeans

In the medium-long run, Africans appear to be more heavily penalised in both wages and employment with respect to the other groups: during the first 5 years of their career the log wage differential is almost the same and increases equally among the three groups, but then it slows down for Asians and Eastern Europeans immigrants, whereas it continues increasing for Africans. A similar pattern also emerges in the employment differential, which is almost constant or declining for Eastern Europeans and Asians, but increases for Africans. Table 6 Foreign-Native differentials in wage and days worked by ethnic groups and sectors at increasing experience in the labour market Log Daily Wage unconditional

Days Worked

conditional unconditional after 5 after 10 at entry years years 1.0% 12.0% 19.0% 19.6%

conditional after 5 after 10 at entry years years 10.0% 11.2% 12.8%

Average

19.1%

Africans Asians East Europeans

19.9% 14.3% 18.2%

1.6% 4.2% 2.2%

12.9% 13.5% 12.8%

19.7% 16.3% 16.9%

19.6% 18.1% 21.6%

10.7% 10.2% 9.8%

11.5% 10.9% 12.0%

13.1% 9.9% 12.4%

Manufacturing Construction Services

16.2% 1.8% 27.4%

1.7% 1.7% 1.0%

11.6% 18.1% 10.5%

17.9% 32.4% 20.4%

13.1% 17.7% 22.0%

10.0% 10.0% 10.0%

10.9% 14.1% 12.9%

11.9% 16.1% 13.5%

VI CONCLUDING COMMENTS Italy began to experience large-scale inflows of economic migrants much later than many other countries, Now, after twenty years of immigration, and with a total legally residing foreign-born population of about three and a half million immigration has become a key issue. This is the first economic analysis of the phenomenon which follows the human capital approach; namely, it explores whether the lifetime wage profile of immigrant workers or their employment outcomes matches those of native workers. Unfortunately, the only dataset available (WHIP) does not cover the entire working population, but only workers in private companies. Hence, it excludes the public administration, agriculture and domestic work. This implies that the conclusions drawn from the empirical analysis can only be 17

extended to the entire population with caution. The economic assimilation of foreigners has been divided into two components: wage assimilation and employment assimilation. Given the increasing evidence of the limited duration of stays among migrants and the increase in temporary migration, the results have been controlled for the probability of staying in the destination country, which according to recent empirical evidence, is mainly a function of the GNP growth rate and level of the sending country. What emerges from the empirical analysis is a discouraging picture. Firstly, migrants tend not to catch up, either in wages or in employment. The control for the probability of staying in the destination country shows that the unobserved characteristics positively influencing the assimilation variable have a negative effect on the decision to stay in the destination country meaning that the best workers leave. However, the specifications with and without the control are very similar. Secondly, foreigners employed in the private sector earn the same wages as natives on entrance in the labour market, but the two wage profiles diverge with increasing on-the-job experience. Natives present a higher return on age which may proxy the missing education variable. Thirdly, employment assimilation profiles are just as bad; the initial differential in days worked per year between native and migrant workers persists over time. However, the finding that migrants may face significant disadvantages is not new. On the one hand, the Italian labour market demands unskilled labour for labour-intensive production processess which maintain competitive an obsolete technology mix. On the other hand, sociological studies (for example, Reyneri 2007) have already suggested a marginalisation of foreign workers, who are employed in jobs which do not offer any career prospects. The analysis by sectors and ethnic groups reinforces this picture. The construction sector emerges as an outlier for its high incidence of informal employment and for having the lowest wage and employment prospects, while manufacturing and, to a lesser extent, services present patterns similar to the aggregate picture. Moreover, in construction the differentials with natives employed in the same sector, both in terms of wages and of days worked, are much higher than in the other two sectors and the differences increase with time spent in the labour market. Across ethnic groups, African workers appear to be more heavily penalised with respect to the other immigrants in the medium-long run, while Eastern European workers appear to be more similar to natives in the negative effects of non-employment on wages, but being the most recent immigrants (their experience in the Italian labour market is 26 months with respect to an average of 43 months), they are not yet able to make their on-the-job experience pay fully. The proxy for the community variable is not significant in any of the estimates. The way the variable is constructed - the share of the workers of the same origin in the region - probably captures a competitive effect of the presence of the same workers in a segmented labour market which, given the small sample, does not have any impact. Foreign wages and employment also seem non-sensitive to the unemployment rate prevailing in the region. These results are not surprising if compared to the evidence from other Southern European countries. For example, Amuedo-Dorantes and de la Rica (2007) analyse the employment assimilation of recent migrants in Spain and they find, as we do, that immigrants do not assimilate in employment. As well Lacuesta, Izquierdo and Vegas (2008) find that after an initial reduction the differential between Spanish and immigrant wages remains stable. Carneiro, Fortuna and Varejão (2007), use a similar dataset for Portugal, and also find a lower wage assimilation of foreign immigrants which declines over time, given the higher coefficient of the tenure variable in the case of foreigners. They also use a variable for ethnic concentration in the workplace which is either negative or not significant, as in our case. Nicolitsas (2007), by using Social Security data for 18

Greece, finds that foreigners have lower wages and that their wage differential remains, although smaller, when controls for sector, occupation and firm are introduced. These results confirm a pattern of labour market segmentation of foreign workers, who are paid less than comparable natives but who work in low skilled jobs and sectors. This corroborates other results obtained by empirical studies in Spain (Dolado et al 1996) and Italy (Gavosto, Venturini and Villosio 1999, Venturini and Villosio 2006), which find that immigrants do not displace natives but, rather, complement them.

19

References Allasino E., Reyneri E., Venturini A., Zincone G., (2004), ‘Discrimination of foreign workers in the Italian labour market’, International Migration papers, n.67, ILO, Geneva Ambrosini M. (2001) ‘The Role of Immigrants in the Italian Labour Market’, International Migration, 39(23), 61-83 Amuedo-Dorantes C., de la Rica S., (2005), ‘Immigrants' Responsiveness to Labor Market Conditions and Its Implications on Regional Disparities: Evidence from Spain’, IZA D.P.1557. Amuedo-Dorantes C., de la Rica S., (2007), ‘Labor Market Assimilation of Recent Immigrants in Spain’, British Journal of Industrial Relations. Baganha M., 1998, Immigrant involvement in the informal economy: the Portuguese case, Journal of Ethnic and Migration Studies, .24,(2), pp.367-385. Bentolila S., Dolado J.J., Jimeno J.F., (2007), ‘Does Immigration Affect the Phillips Curve? Some Evidence for Spain’, IZA D.P. 3249. Blackaby D., Drinkwater S., Leslie D.G.,Murphy P., 1997, A Picture of male and Female Unemployment Among British’s Ethnic Minorities, Scottish Journal of Political Economy, 44(29, pp.182-197. Blackaby D., Drinkwater S., Murphy P. and O’Leary N., 2001, English Language Fluency and the Ethnic Wage Gap in Britain”, Economic Issues, 6, 21-32. Borjas G., (1985), ‘Assimilation, Changes in Cohort Quality, and Earnings of Immigrants’, Journal of Labor Economics, 4, 463-89. Borjas G., (1992), ‘Ethnic Capital and Intergenerational mobility’, Quarterly Journal of Economics, 107, 123-150 Borjas G., Bratsberg D., (1996), ‘Who leaves? The Outmigration of the foreign-born’, Review of Economics and Statistics, 165-176 Borjas, G., (1987), ‘Self selection and the earnings of immigrants’, American Economic Review, 77, 531-553 CARIM, (2008), Proceedings on Circular Migration, RSCAS, Florence Carneiro A., Fortuna N.,Varejão J., (2007), ‘The Economic Assimilation of Immigrants in the Portuguese Labour Market’, mimeo. Carrington W.J., Lima P.J.F., (1996), ‘The Impact of the 1970s Repatriates from Africa on the Portuguese Labor Market’, Industrial and Labor Relation Review, 49(3), 330-347. Cassarino J.P., (2007), Return Migrants to the Maghreb: Reintegration and development challenges, MIREM, RSCAS/EUI, Florence. Chew K., Liu J., (2004), ‘Hidden in Plain Sight: Global Labor Force Exchenge in the Chinese American Population, 1880-1940’, Population and Development Review, 30(1), 57-78. Chiswick B. Hurst M., (2000),’The Employment, Unemployment and Unemployment Compensation Benefits of Immigrants’, in: Bassi, L. J. and Woodbury, S. A. (eds.), Long-Term Unemployment and Reemployment Policies (Research in Employment Policy, V. 2), Stamford, Conn., 87-115. Chiswick B. R., (1978), ‘The effect of Americanization on the earnings of foreign-born men’, Journal of Political Economy, 86(5), 897-921. Chiswick B., (1991), ‘Speaking, Reading, and Earnings among Low-skilled Immigrants’, Journal of Labor Economics, 9(2), pp.149-170 Chiswick B., Miller P.W., 2002, Immigrant Earnings: Language Skills, Linguistic Concentrations and the Business Cycle, Journal of Population Economics, 15 (1), pp.31-57. Clark K. Drinkwater S., (2002), ‘Enclaves, Neighborhood Effects and Employment Outcomes: Ethnic Minorities in England and Wales’, Journal of Population Economics, 13, 5-30. Constant A., Zimmermann K., (2007), ‘Circular Migration: Counts of Exits and Years Away from the Host Country’, IZA D.P. 2999. Constant, A., Massey, D. S. (2003) ‘Self-selection, earnings, and out-migration: A longitudinal study of immigrants to Germany’, Journal of Population Economics, 16 (4), 631-653 Contini B., Filippi M., Malpede C. (2000), "Safari tra la giungla dei salari. Nel Mezzogiorno si lavora meno?", Lavoro e Relazioni Industriali, 2 Cutler D.M., Glaeser E.L., (1997), ‘Are Ghettos Good or Bad?’, Quarterly Journal of Economics, 112, 827872. Dayton-Johnson J., Katseli L.T., Maniatis G., Munz R., Papademetriou D., 2007, Gaining from Migration. Towards a new mobility system, OCDE publication, Paris

20

Devicienti F. Maida A. and Pacelli L. (2008), ‘The resurrection of the Italian Wage Curve’, Economic Letters, 98(3), 335-341 Dolado J., Jimeno J.F., Duce R., (1996),’The Effects of Migration on the Relative Demand of Skilled versus Unskilled Labour: Evidence from Spain’, CEPR Discussion Paper 1476. Dustmann C. and Van Soest A., (2001), ‘Language fluency and earnings equation with misclassified language indicators’, Review of Economics and Statistics, 83, 663-674 Dustmann C., (1997), ‘Return Migration, Savings and Uncertainty’, Journal of Development Economics, 52, 295-316. Dustmann C., (2000), ‘Temporary Migration and Economic Assimilation’, Swedish Economic Policy Review, 7, 213-244. Dustmann C., (2003), ‘Return Migration, Wage Differentials and the Optimal Migration Duration’, European Economic Review, 47, 353-367. Dustmann C., (1994), Speaking fluency, writing fluency and earnings of migrants, Journal of Population Economics, 7, 135-156 Dustmann C., Fabbri F., (2003), ‘Language Proficiency and Labour Market Performance of Immigrants in the UK’, The Economic Journal, 113, 695-717. Dustmann C., Kirchkamp O., (2002), ‘The Optimal Migration Duration and Activity Choice after Remigration’, Journal of Development Economics 67(2), 351-372 Faini R. , Venturini A. ,(1993), ‘Trade, Aid and Migrations. Some Basic Policy Issues’, European Economic Review, 37, 435-442. Faini R., Venturini A., (2008), ‘Growth and migration’, CHILD, Working paper n.11. Fertig M., Schurer S., (2007), ‘Labour Market Outcomes of Immigrants in Germany: the Importance of Heterogeneity and Attrition Bias’, IZA DP 2915. Gavosto A., Venturini A., Villosio C., (1999), ‘Do Immigrants Compete with Natives?’, Labour, 13(3), 603621. González L., Ortega F., (2007), ‘How Do Very Open Economies Absorb Large Immigration Flows? Recent Evidence from Spainish Regions’, IZA D.P. 3311. Hatton T.J., Leigh A., (2007), ‘Immigrants Assimilate as Communities, Not Just as Individuals’, IZA DP 2538. Husted L, Skyt Nielsen H., Rosholm M, Smith N., (2001), ‘Employment and wage assimilation of male firstgeneration immigrants in Denmark’, International Journal of Manpower , 22(1/2), 39-71. ISTAT, (2007), Annual Report 2006, Roma, ISTAT Publishing office ISTAT (2008) ‘La misura dell’occupazione non regolare nelle stime di contabilità nazionale’ Statistiche in breve, Roma, ISTAT Publishing office Izquierdo Escribano A., Munoz-Perez F., (1989), L'Espana pays d'immigration, Population, n.2, pp.257-289. Lacuesta A., Izquierdo M., Vegas R., 2008, Assimilation of Immigrants in Spain : A Logitudinal Analysis, paper presented at the EALE Annual Conference, 18-20 September, Amsterdam. LaLonde R., Topel R., (1992), ‘The Assimilation of Immigrants in the U.S. Labor Market’, in Borjas G., Freeman R., (eds) Immigration and the Workforce, NBER, University of Chicago Press, Chicago, 6792. Lidgard J., Gilson C., (2002), ‘Return Migration of New Zealanders: Shuttle and Circular Migrants’, New Zeland Population Review, 28(1), 99-128. Mansoor A., Quillin B., (2007), Migration and Remittances, Eastern Europe and the Former Soviet Union, The World Bank. Martin P. L., (2003), Managing Labor Migration: Temporary Worker Programs for the 21st Century, ILO, IILS, Geneva. Munz R., (2008), ‘Demographic Change in EU27 and in Neighboring Mediterranean Regions’, paper presented at the Ninth Mediterranean Research Meeting, Florence Montecatini. Natale L., Casacchia O., Strozza S., (1999) ‘Migrazioni interne, migrazioni internazionali: Il nuovo ruolo del Mezzogiorno’, in Bonifazi C. (ed.) Mezzogiorno e migrazioni interne. Monograph No.10. Rome, IRP-CNR., 237-272 Nicolitsas D., (2007), Comparing “new” and “old” immigration countries, mimeo. Nielson H. S., Rosholm M., Smith N., Husted L., (2001), Qualifications, Discrimination, or Assimilation? An Extended Framework for Analyzing Immigrant Wage Gaps, IZA Discussion Paper n. 365 Oaxaca R., (1973), ‘Male-Female Wage Differentials in Urban Labour Markets’, International Economic Review 14, 693-709.

21

Piché V.,Renaud J. and Gingras L., (2002), ‘Economic Integration of New Immigrants in the Montreal Labor Market:A Longitudinal Approach’, Populations, 57, 57-82. Rannveig Agunias D., (2006), From a zero-Sum to a Win-Win scenario? Literature Review on Circular Migration, MPI. Reyneri E., (2003), Immigration and the Underground Economy in New Receiving South European Countries: Manifold negative Effects, Manifold Deep-rooted Causes, International Review of Sociology, 3 (1),pp117-27. Reyneri E.,(2007), ‘Migrants’ Involvement in irregular Employment in the Mediterranean Countries of the European Union’, Dept. of Sociology and Social Research, mimeo. Rosholm M., Scott K., Husted L., (2000), ‘The Times They Are A-Changin’, Organisational Change and Immigrant Employment Opportunities in Scandinavia’, W.P.n.7, Centre for Labour Market and Social Research. Salt J., Singleton A., Hogarth J., (1994), Europe’s international migration, London:HMSO. Schiff M., (2007), Optimal Immigration Policy: Permanent, Guest-Worker, or Mode IV?, IZA D.P. 2871. Shields M.A., Wheatley Price S., (2002), ‘The English language fluency and occupational success of ethnic minority immigrant men living in English metropolitan areas’, Journal of Population Economics, 15, 137-146. Stalker P., (1994), The work of Strangers: A survey of international labour migration, Geneva, ILO. Strozza S., Venturini A., (2002), ‘Italy is no longer a country of Emigration: Foreigners in Italy, How Many, Where They Come From and What They Do’, in Rotte R., Stein P. (eds) Migration Policy and the Economy: International Experiences, Studies and Comments 1 series, Hanns Seidel Stiftung, Munich. Venturini A. Villosio C., (2000), ‘Are Immigrants Assimilating in Italy: the Role of Big Town’, CHILD WP: 11/2000. Venturini A. Villosio C., (2006) ‘Labour Market Effects of Immigration into Italy: An Empirical Analysis’, International Labour Review, 145(1-2), 91-118 Venturini A., (1999) ‘Do Immigrants working illegally reduce the natives ’legal’ employment? Evidence from Italy’, Journal of Population Economics, 12, 135-154 Venturini A., (2004), Post-War Migration in Southern Europe. An Economic Approach, Cambridge University Press. Venturini A., (2007 ) ‘Circular Migration as an Employment Strategy for Mediterranean Countries’, CARIM Circular Migration series, n39. Vidal Coso E., (2007), ‘The labour complementarity between International Migration and National Labour Force in Spain, from a Gender Perspective’, mimeo CED Barcelona. Wickramasekara P., (2002) Asian labour Migration: Issues and Challenges in an Era of Globalization, International Migration papers n.57, International Migration programme, Geneva, ILO.

22

Appendix Table A1 – Human capital, macro and community variables Variables

Description

Source

Level of aggregation

Wage assimilation Months of employment Months out of employment Log VA Reg. unemployment rate Share of reg. foreign employment

Sum of months spent in regular employment up to year t since 1985 for natives and since entrance in WHIP Individual the Italian labour market for foreigners Sum of months spent out of the regular WHIP employment up to year t since fist employment Individual spell observed ISTAT Branch and Logarithm of value added in t national Region accounts ISTAT Regional unemployment rate in t Labour force Region survey Share of foreign employment of the individual’s Country of WHIP same country of origin on total regional origin and employment in year t Region

Employment assimilation Months of employment

Months out of employment Log VA Reg. unemployment rate Share of reg. foreign employment

Sum of months spent in regular employment up to year (t-1) since 1985 for natives and since WHIP Individual entrance in the Italian labour market for foreigners Sum of months spent out of the regular WHIP employment up to year (t-1) since fist Individual employment spell observed ISTAT Branch and Logarithm of value added in t national Region accounts ISTAT Regional unemployment rate in t Labour force Region survey Share of foreign employment of the individual’s Country of same country of origin on total regional WHIP origin and employment in year t Region

Table A2 – Return migration equation variables Variables Real GDP Growth rate of real GDP Years of stay Age at entry

Countries included:

Description

Source

Real Gross Domestic Product per Capita

Penn Word Tables Penn Word Tables

Level of aggregation

Country

growth rate of Real GDP per capita (Constant Country Prices: Chain series) Number of years of presence in Italy since WHIP Individual entrance Age of foreigner when entering in legal WHIP Individual employment Albania, Bangladesh, Bosnia and Herzegovina, Bulgaria, Chile, china, Colombia, Cote d`Ivoire, Croatia, Dominican Republic, Egypt, Ethiopia, Hungary, India, Lebanon, Libya, Macedonia, Morocco, Pakistan, Peru, Philippines, Poland, Romania, Senegal, Somalia, Sri Lanka, Tunisia, Turkey, Ukraine, Uruguay

23

Table B1 Probability of staying in the destination country (1) Coef.

P(stay) _cons real GDP growth rate of real GDP year of stay year of stay^2 age at entry med_afr lat_ame asia est

9.197 -0.00013 -0.005 -0.839 0.032 -0.044 0.034 0.378 -0.034 0.402

N. obs Wald chi2 rho

z 38.87 -6.22 -1.74 -40.42 25.31 9.1 0.32 1.68 -0.33 3.58

30327 5864.36 0.847

Includes also time dummies and controls for sector of employment in the last recorded job. An alternative specification with the country of origin GDP interacted with dummies for the year of entrance has been implemented, which provides similar results, thus has not been reported.

Table C1 Wage assimilation Fixed effect estimates of log nominal weekly wage 18-45 full time male 1990-2003 (robust s.e.) FOREIGNERS

Log daily wage Constant Age Age ^2 Months of employment Months of employment ^2 Months out of employm. Log VA Reg. unemployment rate Share of reg. foreign employm. Share of reg. foreign employm. ^2 Inverted Mills Ratio N obs F corr(u_i, Xb) = Prob > F = R-sq: within = between = overall =

(i) Coef. 2.48 0.040 -0.0002 0.0018 -0.000002 0.0005 0.075 0.003 -1.058 0.299

t 11.3 5.2 -4.0 3.0 -3.2 0.8 4.0 1.0 -0.7 0.6

30327 150.07 -0.5711 0.0000 0.3795 0.0562 0.1259

NATIVES

(ii) Coef.

t

2.58 0.031 -0.0001 0.0022 -0.000002 0.0010 0.073 0.002 0.037 -0.026 -0.029

17.3 3.5 -1.8 3.1 -2.4 1.4 3.4 0.8 0.0 -0.1 -3.8

30327 121.630 0.5439 0.000 0.368 0.0562 0.1236

24

(iii) Coef.

t

2.57 0.055 -0.0003 0.0022 -0.00001 -0.0006 0.059 -0.005

108.8 19.7 -18.2 9.8 -51.0 -2.6 17.2 -18.3

420205 6471.03 -0.3693 0.0000 0.5946 0.1951 0.3038

Includes also regional, sector and yearly dummies and controls for occupation and firm size. Age is current age minus 16 (the minimum legal age requirement) (ii) With correction for selection due to return migration

Table C2 Employment assimilation Fixed effect estimates of number of days worked in the year 18-45 full time male 1990-2003 (robust s.e.)9 FOREIGNERS N. of days worked in the year _cons Age Age ^2 months of employment (*) months of employment ^2 months out of employm. (**) Log VA Reg. unemployment rate Share of reg. foreign employm. Share of reg. foreign employm. ^2 Inverted Mills Ratio

(i) Coef.

t

230.37 4.067 -0.039 0.143 -0.001 0.023 11.16 0.753 -1.05 1.30

26.2 7.3 -2.8 4.5 -6.5 0.4 1.0 1.1 -0.3 1.0

(ii) Coef.

t

228.39 4.121 -0.031 0.138 -0.001 0.030 11.778 0.770 -1.422 1.372 -2.366

8.8 7.4 -2.0 4.4 -6.4 0.5 1.1 1.1 -0.4 1.1 -1.5

N obs 30327 30327 F= 18.64 18.60 corr(u_i, Xb) = -0.2897 -0.2517 Prob > F = 0.0000 0.0000 R-sq: within = 0.0776 0.0777 between = 0.0353 0.0404 overall = 0.0581 0.0632 Includes also region, sector and year dummies and controls for occupation and firm size (*) Refers to cumulative months of employment up to the previous year (**) Refers to cumulative months of unemployment up to the previous year (ii) With correction for selection due to return migration For a robustness check we also conducted Tobit estimates, with similar results

NATIVES (iii) Coef. 257.88 4.563 -0.057 0.0077 -0.0004 -10.197 18.760 -0.012

420205 307.9 -0.2222 0.0000 0.0556 0.0923 0.0896

Table D1. Wage and employment assimilation by sector of employment a) Natives Wage assimilation Age Age ^2 Months of employment Months of employment ^2

manufacturing coeff t 0.0582 21.5 -0.0003 -17.7 0.0023 10.2 -0.000007 -41.3

construction coeff t 0.0674 23.3 -0.0005 -14.8 0.0009 3.9 -0.000006 -17.2

25

services coeff t 0.0544 19.6 -0.0001 -4.7 0.0025 11.0 -0.000008 -34.1

t 255.7 56.1 -28.7 2.1 -15.3 -35.3 10.2 -0.3

Months out of employm. Employment assimilation Age Age ^2 Months of employment Months of employment ^2 Months out of employm.

-0.0008

-3.5

-0.0009

-3.8

-0.0005

-2.3

4.153 -0.069 0.0092 -0.00010 -0.077

8.3 -22.9 2.6 -2.9 1.8

2.776 -0.022 0.138 -0.00031 -0.170

5.3 -4.0 3.0 -4.1 3.8

4.087 8.0 -0.056 -14.9 0.0099 2.1 -0.00018 -4.0 -0.012 -0.3

construction coeff t 3.5 0.0337 -1.7 -0.0002 0.0011 1.6

services coeff t 2.6 0.0255 0.0000 -0.1 3.6 0.0027 -0.00001 -3.7 0.0010 1.3 -0.0101 -0.6

b) Foreigners Wage assimilation Age Age ^2 Months of employment Months of employment ^2 Months out of employm. IMR Employment assimilation Age Months of employment Months of employment ^2 Months out of employm. IMR

manufacturing coeff t 3.2 0.0293 -1.7 -0.0001 3.7 0.0027 -2.6 -0.000003 0.0011 1.5 -2.8 -0.0300 1.155 0.116 -0.001 -0.035 -0.055

2.5 3.1 -4.3 -0.5 -0.03

0.0015 -0.0176

2.0 -1.3

1.535 0.077

2.2 1.5

2.035 0.005

3.5 0.1

-0.066 -6.906

-0.5 -0.7

-0.094 -1.926

-1.0 -0.8

Table D2. Wage and employment assimilation by ethnic groups Wage assimilation Age Months of employment Months of employment ^2 Months out of employm. IMR Employment assimilation Age Months of employment Months of employment ^2 Months out of employm. IMR

Africans coeff t 3.9 0.0347 2.2 0.0016 -2.8 0.0000 0.0004 0.6 -3.4 -0.0301 2.451 0.069 -0.001 0.029 0.183

1.9 1.8 -2.1 0.4 0.1

26

coeff 0.0505 0.0005

Asians t 14.4 1.9

Est Europeans coeff t 6.5 0.0471 0.0006 1.0

-0.0007 -0.0557

-1.3 -3.8

-0.0013 -0.0357

-1.9 -2.8

3.179 -0.159

3.4 -1.6

2.563 0.134

2.8 2.1

-0.381 -3.349

-1.8 -0.7

-0.023 -2.425

-0.1 -0.5

Suggest Documents