Journal of Ethnic and Migration Studies Vol. 29, No. 3: 519–534
May 2003
Socio-economic integration of ethnic Greeks from the
former USSR: obstacles to entry into the Greek labour
market
George Halkos and Dimitrios Salamouris
Abstract The purpose of this study is to collect data on the demographic and
socio-cultural characteristics of ethnic Greeks from the former USSR and then assess the
extent to which these characteristics affect their integration into the Greek labour market.
To achieve this target we employ logistic regression analysis to test a model to predict
the case of being employed or not by considering various factors as explanatory variables.
A field survey carried out in the broad area of West Athens (total of 6,994 respondents)
is used to estimate the parameters of this model. From past experience we identify 13
major variables contributing to employment or not. The results generate an overall
model with six of the 13 variables statistically significant. The final part of the paper
discusses the policy implications of the findings, particularly the necessity for better
professional and language training of the ethnic Greek immigrants from the ex-USSR.
KEYWORDS: ETHNIC GREEKS; MIGRATION; SOCIO-ECONOMIC INTEGRATION; LABOUR MARKET; LOGISTIC REGRESSION
Introduction
During the twentieth century approximately 400,000 ethnic Greeks from the
(former) Soviet Union settled in Greece. The Census of 1928 recorded 230,000
while, during the decade 1989–99, 146,102 ethnic Greeks from the former USSR
and holding foreign passports were ‘repatriated’.1 Among this latter group,
103,573 had applied for, and received, Greek nationality by February 2001
(Fakiolas 2001). Today, these people face adaptation difficulties in Greek society:
a plethora of problems related to employment, knowledge of the Greek language
and understanding the Greek way of life. Geographically, 15 per cent were
resettled in Thessalonica, 20 per cent in Thrace and Macedonia, and 60 per cent
in Athens (mainly in Western Athens).
For a long time, Western Athens has been a degraded area in terms of its
social, economic, environmental and urban-planning levels. This resulted in
unstructured development and environmental deterioration. In the 1920s
refugees from the Black Sea and Asia Minor created the initial settlements in the
area. They settled in the peripheral suburbs of West Athens and formed the
initial residential cores. During the 1940s Western Athens consolidated its role as
a place for receiving low-income people, newly arrived as a result of the internal
migration from the rest of the country. All these people contributed to the
development of the area since industries as well as large numbers of small
enterprises were based in West Attica (the broader area of Athens). As a result
ISSN 1369-183X print/ISSN 1469-9451 online/03/030519-16  2003 Taylor & Francis Ltd
DOI: 10.1080/1369183032000109185
Carfax Publishing
520
G. Halkos and D. Salamouris
of poor town planning, the low-quality environment and cheap rents, but also
of the increased opportunities for employment, West Athens has become the
main residential area of the metropolis for immigrants and ethnic Greeks from
the former USSR over the last 15 years.
The arrival of such a massive migration stream in such a short time period has
exposed the lack of appropriate infrastructure and policy for the migrants’
smooth social and employment insertion in the local society. Local government
efforts have proved inadequate to resolve these problems, mainly due to limited
funds available. Ethnic Greeks or repatriates from the ex-Soviet Union constitute
a population group with particular characteristics; these are strongly related to
the dissimilar social, political and economic system that they come from. After
their arrival in Greece they faced acute problems of housing, employment,
communication and cultural adaptation.2
This paper examines the demographic and certain social and cultural characteristics of this population group, as well as the extent to which these characteristics affect their integration into the Greek labour market. For this purpose, we
conducted a large-scale field survey. A number of variables were selected and
then used to develop an ordinal logistic regression model. The main contributions of our paper are the following:
• We record for the first time all ethnic Greeks from the former USSR living in
the broad area of West Athens in the form of a field census.
• Very few researchers have dealt with the issue of the repatriation of Greeks
from the ex-USSR. Moreover almost all the existing studies restrict their
analysis to the presentation of basic descriptive statistics or to small-scale
qualitative surveys. In this paper, a logistic regression analysis is applied to
the task of calculating the odds ratio of ethnic Greeks being employed.
• The reliability of our results is all the greater since the sample size examined
is the biggest compared to other studies attempted. In fact the sample size
recorded approaches the population size in this area.
In the next section a review of current similar research and its main findings
is provided. Then, the methodology of approaching and recording this population group is described, followed by the main statistical results of this effort.
We then present the model employed and the empirical results obtained
therefrom. Finally, the problems that lead to difficulties for ethnic Greeks in
finding employment in the local labour market are specified in order to suggest
policy options.
Review of the literature
International migration has undoubtedly attained a growing importance on both
global and domestic policy agendas in the last two decades, a trend confirmed
by many authors. For instance Rainer Münz (1996), reviewing the scale and
geography of migration to and within Europe for the second half of the
twentieth century, including the changing causes, patterns and trends of migration in key periods, claims that migration has a greater impact on population
size and structure compared to the balance of births and deaths.
Hans van Amersfoort (1996) concentrates on the limits of governmental
control over migration, noting that the variables addressed by governmental
Integration of ethnic Greeks from the former USSR
521
measures form only a small part of the relevant variables. To confront effectively
these variables we have to overcome the obstacles of the conflicting interests of
various economic sectors and the conflicting roles of the various administrative
institutions that they have established for themselves. Van Amersfoort claims
that the ‘root-cause’ approach does not appear to be the case at the moment for
better prospects for the management of migration.
Moving closer to the theme of the present paper, Michael Blos et al. (1997)
investigate the impact of migration policies on the labour market performance of
migrants in Sweden and Switzerland. They compare the income and employment of groups of foreign residents relative to natives, looking at socio-demographic characteristics and educational policies. They assert that the Swiss policy
has been economically oriented, considering migrants as temporary guests
without any explicit integration policy. The Swedish policy has given more
attention to political and humanitarian questions, rather than to economic
aspects, aiming at the long-run integration of foreigners once admitted to the
country. Harvey Krahn et al. (2000), using a sample of 525 adult refugees who
were in professional or managerial positions prior to their arrival in Canada,
explore issues of access to high occupations in the Canadian labour market. They
found that these refugees, compared to Canadian-born individuals, experience
much higher rates of unemployment, part-time employment and temporary
employment, despite their high educational attainment.
There are just a few Greek research studies concerning ethnic Greeks.3 These
studies record information on a number of issues like the reasons of emigration,
marital status, level of education, skills, knowledge of the Greek language and
the correspondence of the ethnic Greeks’ professional qualifications with the
demand on the Greek labour market.
Some of the main obstacles for the adaptation of ethnic Greeks into the Greek
labour market are their low level of knowledge of the Greek language and
insufficient capital for the start-up and development of their own enterprises, the
long period required for the recognition of their degrees and other studies as
equivalent to awards from Greek universities and technical schools, and the
generally high level of unemployment in Greece (Fakiolas 2001). In a study of
repatriates and immigrants carried out in the early 1990s, three-quarters of
respondents considered that insufficient knowledge of the Greek language had
a negative influence on the preferences of Greek employers, some of whom were
found to employ ethnic Greeks at lower wages compared to those offered in the
market and relative to their qualifications (Glytsos 1995). Rossetos Fakiolas
(2001) claims that even the high educational level of ethnic-Greek repatriates
cannot help them against unemployment. When the settlement of the ethnic
Greeks started in the 1980s the growth rate of Greek GDP was quite slow – an
average 1.7 per cent yearly for the period 1979–97. This resulted in an increase
in overall unemployment, which climbed to 11.9 per cent (more than 530,000
people) at the end of 1999, compared to the 9.4 per cent average rate for the EU.
There is also a massive influx of foreign undocumented economic immigrants,
whose labour market impact and working conditions have been commented on
by, inter alia, Droukas (1998), Lazaridis (1996) and Lianos et al. (1996). No official
recorded information is available on the number of undocumented immigrants
in Greece. Lianos et al. (1996) claim that the total number of immigrants in 1992
was already 271,000, among them 181,000 undocumented. By 1994, according to
the same authors, the number of undocumented immigrants had virtually
522
G. Halkos and D. Salamouris
doubled to 350,000, largely because of the mass influx from Albania in the early
1990s. In 1998 around 371,000 immigrants submitted applications for legalisation; about 200,000 did not apply, either because they were afraid that submitting personal information to the Greek authorities would lead to their direct
expulsion or due to the lack of the necessary certificates (Fakiolas 2000; Naxakis
and Chletsos 2001). Following the two legalisations in 1998 and 2001, it is
estimated that about 450,000–500,000 immigrants are now legal; and by and
large they compete with the ethnic-Greek settlers in the same sectors of the
labour market.
Methodological approach
For this study we conducted a large-scale field survey, carried out during
1998–99. The initial problem we faced was the location of the target population.
In order to overcome this, we developed close collaboration with the ten
municipalities of West Athens, as well as with the local repatriates’ unions. In
particular, we consulted the municipal rolls of the municipalities where ethnic
Greeks live and the archives of the local ethnic Greeks’ unions. However, the
research and evaluation of the data collected from these two sources raised
several problems. We realised that the actual number of ethnic Greeks included
many who did not appear on the municipal rolls, since during their first period
of stay after their arrival in the area, they were guests of their friends or relatives
and not permanent residents. In other cases where we obtained apparently good
data regarding place of residence through the municipal rolls, we found that we
could not actually utilise the addresses because of the repatriates’ high mobility,
searching for a better place of residence and a better means of livelihood. On the
other hand the files of the unions were incomplete since only a limited number
of people (approximately 4 per cent) were found to be participating in these local
unions.
Nevertheless the logistical contribution of the municipalities and unions was
substantial. They assisted our attempt to approach this population group in a
more direct way in order to assure the maximum possible accuracy of the data,
and thus the reliability of the study. They further contributed by suggesting
appropriate persons as interviewers, most of whom were also ethnic Greeks
belonging to our target group. Knowing the language, they played an important
role in achieving better communication with the target population. In addition
they knew the places where most of the ethnic Greeks live, thus helping to locate
them easily. The administration of the survey was finally implemented door-todoor, with all the difficulties this method implies.
We distributed individual questionnaires accompanied by personal interviews. The questionnaire was designed to provide personal data on the economically active population, demographic characteristics, employment and
education characteristics and other data. Collectively these data enabled us to
determine the degree of adaptation to Greek society. The first questions were
general and referred to the personal characteristics of the individual respondent:
sex, nationality, religion, educational level, marital status and other private
information. Most of the other questions were formatted using the Likert scale.
A pilot questionnaire was first tried out on a small sample of respondents in an
effort to ensure that the instructions and questions were clear and comprehensible. This kind of pre-test provided us with a limited amount of data, which
Integration of ethnic Greeks from the former USSR
523
also gave an idea of what we could expect from the upcoming full-scale study.
It was also a chance to check out the redundancy of some of the initial questions
included in the draft questionnaire and which were eventually excluded.
Twenty interviewers were employed in the ten municipalities of Western
Athens. The main problem they encountered was the disbelief and suspiciousness on the part of the ethnic Greek respondents. Some of them even denied they
were Greek repatriates, while others feared that they would be persecuted.
However, most of them were eventually convinced of the true nature and the
positive results of this research for them.
This is the first time a field survey has been conducted on this particular
population group in Western Athens. In total, 6,994 persons were surveyed,
3,436 men and 3,558 women. This population constitutes the reference set on
which the remainder of our study is based.
Statistical results of the survey
The information gathered from this study reflects the particular characteristics of
the ethnic Greeks, as well as the development of their social and economic
condition by the time they first arrived in Greece. From our field research we
conclude that the main reason for immigration is the ‘minority syndrome’; the
choice of Greece as their destination country instead of any other country is due
to the fact that they always considered Greece as their spiritual homeland. As far
as their socio-economic condition is concerned we discovered that in their
country of origin they had created a satisfactory standard of living: the majority
had a secure job and housing, while unemployment was at a low level. Thus it
can be argued that the reason for immigration had no strong economic motive
as the financial status of the ethnic Greeks in the former USSR was broadly
satisfactory.
As can be seen from Figure 1, the repatriates’ immigration flow has gradually
reduced since its peak in 1990. Overall, 71.4 per cent of the sample moved to
Greece in the period 1990–93, only 12.3 per cent during the period 1994–99, and
15.7 per cent before 1989 (0.5 per cent did not answer). This temporal pattern is
explicable in the light of the stabilisation of the political situation in the countries
of origin as well as the discovery that, in Greece, their expectations were no
longer satisfied. As far as their financial situation in Greece is concerned, 43 per
cent of ethnic Greeks responded that it is worse than before ‘repatriation’, 34 per
cent said it is the same, 22 per cent better and 1 per cent refused to answer.
Virtually all of the ethnic Greeks from the former USSR came from Kazakhstan (44.3 per cent), Russia (26.5 per cent) and Georgia (25.5 per cent); much
smaller numbers from other republics such as Uzbekistan (1.9 per cent) and
Kyrgyzstan (0.5 per cent). We found that the Greeks coming from Kazakhstan
and Georgia (almost two-thirds of the sample) had preserved their Greek
character (Greek nationality, ethos and family structure) to a greater extent than
those from Russia. After the collapse of the USSR, unstable political conditions,
continuous economic crisis, the creation of war zones and ethnic tensions
resulted in a high number of immigrants from these two republics. The vast
majority of the ethnic Greeks nominated West Athens as their place of permanent residency. As far as the reasons for choosing West Athens are concerned,
80 per cent of the people said because their friends and family already lived
there. Nearly three-quarters (73.1 per cent) of the survey respondents lived in
524
G. Halkos and D. Salamouris
Figure 1. Repatriation flow of ethnic Greeks from the former Soviet Union
one municipality, Aspropirgos, where rents were particularly low and where
community contacts were highly concentrated.4
From the analysis of the demographic characteristics we found that 64 per cent
of the ethnic Greeks moved with their families, and 36 per cent moved alone.
Detailed statistics concerning the age, sex and marital status of the targeted
population are set out in Table 1. This shows that two-thirds of the population
were aged 20–49, which is the most productive age. Less than one-fifth were
aged 60 ⫹ , mainly women. Regarding marital status etc., 61.9 per cent of the
population surveyed were married, 27.2 per cent were single, 7.4 per cent were
widowed and 3.4 per cent divorced.
Table 2 cross-tabulates the level of unemployment by age and sex. A high
percentage of unemployment is observed, especially in the young and middleage groups and in females. As can be seen, 2,492 out of 6,994 (35.6 per cent) were
unemployed, of whom two-thirds are female and one-third male. For the most
productive ages of 20–29, 30–39 and 40–49, the total unemployment rate is 38.2
per cent, 36.9 per cent, and 35.6 per cent respectively. The unemployment rates
for the remaining age groups can also be considered as high compared to the
level of unemployment of the native population.5 Looking at the other side of
the coin, Table 3 presents the answers to the survey question ‘What is your
occupation in Greece today?’. It can be seen that only 1.6 per cent of them were
employers or self-employed. The largest group – 41.1 per cent – were employees.
Those who were not employed for various reasons (housewives, etc.) corresponded to 12.1 per cent, while 9.6 per cent were pensioners and the residual
35.6 per cent unemployed, as noted above.
Table 4 sets out the level of education of the ethnic Greeks. Overall 6.5 per cent
are university graduates and 7.4 per cent are HND-equivalent or polytechnic
graduates, which means that a total of 13.9 per cent are higher education
graduates. Next, 40.1 per cent of the population are secondary and secondary
technical school graduates, 26.5 per cent finished primary school, while a
considerable percentage of the population (15.1 per cent) are illiterates. It is
worth mentioning here that, compared to the native population of Greece and
especially relative to the younger generations, this educational profile is poor,
which may in the long run affect the social and economic integration of the
ethnic Greeks.
166
687
202
70
11
3
3
1,142
14–19
20–29
30–39
40–49
50–59
60–69
70–89
Total
1,904
M
Single
Age
group
174
437
76
38
13
13
11
762
F
F
7
254
574
656
303
240
131
2,165
4,332
1
97
483
753
331
279
223
2,167
M
Married
0
0
2
1
4
21
45
73
M
518
0
2
20
34
43
94
252
445
F
Widowed
0
0
18
15
13
6
2
54
M
240
0
10
54
72
22
18
10
186
F
Divorced
6,994
348
1,487
1,429
1,639
740
674
677
Total
M⫹F
27.2
4.9
16.1
4.0
1.5
0.3
0.2
0.2
Single
%
Table 1. Marital status by sex and age
61.9
0.1
5.0
15.1
20.2
9.1
7.4
5.1
Married
%
7.4
0.0
0.0
0.3
0.5
0.7
1.6
4.3
Widowed
%
3.4
0.0
0.1
1.0
1.2
0.5
0.3
0.2
Divorced
%
100
5.0
21.3
20.4
23.4
10.6
9.6
9.7
Total
M⫹F %
Integration of ethnic Greeks from the former USSR
525
526
G. Halkos and D. Salamouris
Table 2. Unemployment by age and sex (%)
Age group
14–19
20–29
30–39
40–49
50–59
60–69
70 ⫹
Total
Male %
Female %
Total %
13.2
30.0
19.3
20.4
25.3
38.5
18.7
24.0
12.2
47.2
54.1
51.6
63.8
55.1
15.6
46.8
12.6
38.2
36.9
35.6
45.1
47.5
16.8
35.6
Table 3. Occupation in Greece today
Employers
Self-employed
Employees
Unemployed
Pensioners
Housewives etc.
Total
Male
Female
Total
Male
%
Female
%
Total
%
9
73
1,873
826
266
389
3,436
4
22
1,003
1,666
405
458
3,558
13
95
2,876
2,492
671
847
6,994
0.1
1.0
26.8
11.8
3.8
5.6
49.1
0.1
0.3
14.3
23.8
5.8
6.6
50.9
0.2
1.4
41.1
35.6
9.6
12.1
100.0
Table 5 presents the unemployment rates by level of education, for the main
categories listed in Table 4 (i.e. excluding the numerically less important students and ‘no answers’). We observe high unemployment rates for all levels of
education; and female unemployment rates (46.8 per cent overall) are significantly higher than male (24.0 per cent overall) at all levels of education. The high
unemployment rate of people with higher education (43 per cent for university
graduates and HND-equivalent graduates) underlines the acute absorption
problem of the scientific workforce in the Greek economy. According to the
results of this research, most of the higher education graduates are underemployed or they work in a different field than their specialisation.
Table 4. Educational characteristics
Level of education
Male
University
HND, Polytechnic
Students
Secondary School
Secondary
Technical School
Primary School
Illiterates
No answer
Total
214
194
32
974
474
1,017
444
87
3,436
Total
Male
%
Female
%
Total
%
243
325
53
871
485
457
519
85
1,845
959
6.2
5.7
0.9
28.4
13.8
6.8
9.1
1.5
24.5
13.6
6.5
7.4
1.2
26.4
13.7
839
614
128
3,558
1,856
1,058
215
6,994
29.6
12.9
2.5
100.0
23.6
17.3
3.6
100.0
26.5
15.1
3.1
100.0
Female
Integration of ethnic Greeks from the former USSR
527
Table 5. Unemployment by level of education (%)
Level of education
University
HND, Polytechnic
Secondary School
Secondary Technical School
Primary School
Illiterates
Total
Male %
Female %
Total %
26.2
25.3
21.2
21.9
26.7
28.2
24.0
58.9
53.4
48.2
53.6
49.2
36.8
46.8
43.5
43.0
33.9
37.9
36.9
33.2
35.6
The main explanation of the above situation is that most of the ethnic Greeks
have education and work specialities which do not correspond to the needs of
the Greek labour market. This divergence, which creates structural unemployment, is mainly attributed to the strategy of the ex-USSR central government for
the specialisation of economic activity in each republic. Hence the development
of particular skills of the ethnic Greeks depended on the place of residence. As
mentioned above, the majority of ethnic Greeks living in West Athens come
from the Republics of Kazakhstan and Georgia. In Kazakhstan the Greeks were
mainly working in the mineral industry and a few of them in agriculture and
services. In Georgia there are some rural areas which consist only of Greeks
working mainly in agriculture. However, the progressive urbanisation of the
Greeks of Georgia, as a result of industrial development, led increasing numbers
of these ruralites to be employed in industrial production and to a lesser extent
in the service sector. Overall, the central planning of particular sectors of
economic activity in these regions, in combination with the technical orientation
of education in these sectors resulted in the creation of a strictly specialised
labour force many members of which are unable to find appropriate employment in Athens.
In Table 6 our survey results on the level of proficiency of the Greek language
(written and verbal) are presented. Straightaway, we can see that the level of
knowledge of Greek is poor for older people. The higher the age group, the
higher the percentage of ethnic Greeks who speak the Greek language at a poor
or moderate level. This conclusion holds for both oral and written knowledge of
Greek. Moreover, the level of knowledge for written Greek is worse than the
level of knowledge of verbal Greek. In synthesis, 77 per cent of the surveyed
population knows poor or moderate written Greek while 50 per cent knows poor
or moderate verbal Greek.
At ages 14–19 the language problem is relatively minor since only 20 per cent
of them said they have ‘poor’ or ‘moderate’ knowledge of written Greek, while
80 per cent said ‘good’ or ‘excellent’. This is because the majority of these
children had their primary school education in Greece. The problems start
becoming acute at the age groups 20–29, 30–39 and 40–49 where 56, 82 and 89
per cent respectively assessed their level of knowledge of the written Greek
language as poor or moderate. By comparing these results with those of Table
2, which shows the unemployment rates for these particular age groups, we can
see that the relation between (high) unemployment and (low) level of knowledge of the Greek language is quite close. Since these age groups comprise most
of the economically active population, it is necessary to develop a policy which
528
G. Halkos and D. Salamouris
Table 6. Level of knowledge of written and verbal Greek, by age (%)
Written
Verbal
Age
group
Poor
Moderate
Good
Excellent
Poor
Moderate
Good
Excellent
14–19
20–29
30–39
40–49
50–59
60–69
⬎ 70
4.5
20.0
37.6
47.4
58.8
75.4
73.5
15.8
36.0
44.7
41.4
32.1
20.0
22.7
67.0
33.7
16.6
10.2
8.3
4.0
3.6
12.7
10.3
1.1
1.1
0.8
0.6
0.2
1.5
1.6
2.5
2.5
4.8
10.3
14.3
9.3
25.0
40.5
49.6
60.2
72.0
69.8
67.5
55.7
53.6
45.8
33.0
17.3
14.9
21.7
17.7
3.4
2.0
2.1
0.5
1.0
will help these people learn the Greek language. For the age groups of 50–59,
60–69 and ⬎ 70, the problem is dramatic since only a small percentage of them
said their level is ‘good or excellent’. These older ethnic Greeks may never attain
fluency in Greek, especially written Greek (and some are illiterate anyway).
The model
In the remainder of the paper we propose to develop an ordinal logistic
regression model. This method of statistical analysis was preferred over multiple
regression for a number of reasons. First, the dependent variable is ordinal and
not continuous. Second, the ordinal logistic regression model is a more appropriate monotone function for our data-set in contrast to the least squares criterion
of a multiple regression analysis. At the same time, logistic regression was
preferred to discriminant analysis as the latter relies on meeting the assumptions
of multivariate normality and equal variance-covariance matrices across groups.
These assumptions are not required with logistic regression.
The regression coefficients of the proposed logistic model quantify the relationship of the independent variables to the dependent variable involving the
parameter called the Odds Ratio (for more details on the properties and
applications of logistic regression see Collett 1991; Hair et al. 1998; Hosmer and
Lemeshow 1989; Kleinbaum 1994; Kleinbaum et al. 1999; Sharma 1996). As odds
we define the ratio of the probability that employment will take place divided
by the probability that it will not take place. That is
Odds (E|X1, X2 … , Xn) ⫽
Pr(E)
1 ⫺ Pr(E)
(1)
where X1, X2, … , Xk the k independent variables. Instead of minimising the
squared deviations as in multiple regression, logistic regression maximises the
likelihood that an event will take place.
ln
Pr
⫽ 0 ⫹ 1X1 ⫹ 2X2 ⫹ … ⫹ kXk
1 ⫺ Pr
(2)
or
Pr ⫽
1
k
1 ⫹ e ⫺ ( ⫹ i ⫽兺 1BiXij)
(3)
Integration of ethnic Greeks from the former USSR
529
where Pr is the probability of being employed given the independent variables
X1, X2, … , Xk. Equation (2) models the log of the odds as a linear function of the
independent variables and it is equivalent to a multiple regression equation with
log of the odds as the dependent variable.
The logit form of the model is a transformation of the probability Pr(Y ⫽ 1)
which is defined as the natural log odds of the event E(Y ⫽ 1). That is
logit [Pr(Y ⫽ 1)] ⫽ loge[odds (Y ⫽ 1)] ⫽ loge
冋1 ⫺Pr(YPr(Y⫽⫽1)1)册
(4)
Empirical results
As our key interest is in terms of the main effects we have ignored interactions.
We have considered a number of variables, namely, sex (male or female), age (in
years), religion, nationality, number of family members with special needs,
family status (single, married, divorced, widowed), number of children, graduate or not (0 and 1), difference in profession in Greece from the one in the source
country (0 or 1), written knowledge of Greek language (ordinal variable with
values 1,2,3,4), verbal knowledge of Greek language (ordinal variable with
values 1,2,3,4), education level (ordinal variable with values 1,2,3,4,5,6,7), and
years of education.
From these 13 explanatory variables, six were found to be statistically significant in influencing the level of employment. Prior to our study we expected to
see the ethnic Greeks absorbed easily by the large numbers of industries and
small-medium enterprises based in the area of West Athens. But due to the
incompatibility of their professional qualifications with the demands of the local
labour market, this did not prove to be the case. To further explore this reason,
we have constructed a variable that reflects the difference between the professions of ethnic Greeks before coming to Greece and after they settled down.
The logit form of the fitted model may be presented as
logit[Pr(Y ⫽ 1)] ⫽ 0 ⫹ 1 Sex ⫹ 2 Age ⫹ 3 DiffProf ⫹ 4 GreekWr ⫹ 5 GreekVr
⫹ 6 Educlev
where Y denotes the dependent variable as 1 for being employed and 0 for being
unemployed. The six explanatory variables are sex, age, difference in profession
in the ex-Soviet Union and Greece (DiffProf), the written and verbal knowledge
of the Greek language (GreekWr, GreekVr), and the education level (EducLev).
Based on the fitted model and the information provided, we can compute the
estimated odds ratio (OR). This is the probability that ethnic Greeks will be
employed in different professions relative to ethnic Greeks being employed in
similar professions controlling for sex, age, written and verbal knowledge of the
Greek language and education level. Hence
OR
0.165
⫽ 1.18
(DiffProf ⫽ 1 vs DiffProf ⫽ 0兩Sex, Age, GreekWr, GreekVer, EducLev) ⫽ e
This algebraic expression gives us the adjusted odds ratio equal to 1.18 which
means that the odds of employment are about 1.18 times higher for a repatriate
who accepts a different profession than for a repatriate who does not accept such
a position. But, as can be seen from Table 7, the Wald statistic is not significant,
which indicates that there is no statistical evidence in these data that the
difference in profession significantly increases the probability of employment.
530
G. Halkos and D. Salamouris
Table 7. Logistic regression results
Regressors
Coefficients
Standard
errors
Wald
Statistic
(Chi-sq)
0.849
⫺ 1.290
0.005
0.165
0.179
0.223
0.083
0.299
0.069
0.003
0.114
0.056
0.064
0.021
8.065
345.9
3.610
2.075
12.204
12.086
15.781
Constant
Sex
Age
DiffProf
GreekWr
GreekVr
EducLev
95% CI for
EXP(B)
Sig.
ei
Lower
Upper
0.005
0.000
0.057
0.150
0.001
0.001
0.000
2.337
0.275
1.005
1.179
1.196
1.250
1.087
0.240
1.000
0.942
1.072
1.102
1.043
0.315
1.011
1.476
1.335
1.418
1.132
Note: The explanatory variables are sex, age, difference in profession in the former Soviet Union and
Greece (DiffProf), written and verbal knowledge of the Greek language (GreekWr, GreekVr), and
education level (EducLev).
Working with the results of Table 7, we can realise the importance of the
ˆ
variable Sex. We may compute the difference ei ⫺ 1, which estimates the
percentage change (increase or decrease) in the odds
⫽
Pr(Y ⫽ 1)
Pr(Y ⫽ 0)
for every 1 unit change in Xi holding the other entire X’s fixed. The coefficient
ˆ
ˆ
of Sex is ˆ 1 ⫽ ⫺ 1.290, which implies that e1 ⫽ 0.275 and e1 ⫺ 1 ⫽ ⫺ 0.725. This
means that we estimate the odds of employment to decrease by 72.5 per cent,
holding all the rest fixed. Similarly, the coefficient of Age is ˆ 2 ⫽ 0.005, which
ˆ
ˆ
implies that e2 ⫽ 1.005 and e2 ⫺ 1 ⫽ 0.005. This means that for each additional
year of age we estimate the odds of being employed to increase by only 0.5 per
cent, holding fixed all the rest. Similarly, looking at the difference in profession,
the written and verbal knowledge of the Greek language and the education level
we expect the odds of employment to increase by 17.9, 19.6, 25.0 and 8.7 per cent
respectively, all the others remaining fixed in each case.
The standard errors of the estimates are given in the appropriate column of
Table 7 and the squared ratios of estimates to their corresponding standard
errors are presented in the column Wald Statistic (Chi-square). The significance
levels of the individual statistical tests (i.e. the P-values) are presented in the
column Sig (Significance) and correspond to Pr ⬎ Chi-square. Note that the
constant term and the variables Sex, Greek language written and verbal and
Educational level are significant at all statistical levels. The variable Age is
statistically significant at the 0.1 level while the variable Difference in Profession
is not statistically significant at the usual significance levels (0.05, 0.01 and 0.1).
The overall significance of the model is given by X2 ⫽ 23.447 with a significance level of P ⫽ 0.000 and 7 degrees of freedom. Based on this value we can
reject H0 (where H0: 0 ⫽ 1 ⫽ 2 ⫽ 3 ⫽ 4 ⫽ 5 ⫽ 6 ⫽ 0) and conclude that at least
one of the coefficients is different from zero (X20.05,7 ⫽ 14.067). That is, to assess
the model fit we compare the log likelihood statistic ( ⫺ 2 log L̂]) for the fitted
model with the explanatory variables with this value that corresponds to the
reduced model (the one only with intercept). The likelihood ratio (LR) statistic
for comparing the two models is given by the difference
Integration of ethnic Greeks from the former USSR
531
LR ⫽ ⫺ 2 log L̂R ⫺ ( ⫺ 2 log L̂F) ⫽ 256.012–232.565 ⫽ 23.45
where the subscripts R and F correspond to the Reduced and Full model
respectively. This value must be compared with X20.05,7 ⫽ 14.067, which implies
again a rejection of H0.
The Hosmer and Lemeshow value equals 6.327 (with significance equal to
0.347). The non-significant X2 value indicates a good model fit in the correspondence of the actual and predicted values of the dependent variable.
Conclusion and policy implications
During the last 15 years, the migratory stream of ethnic Greeks from the former
USSR to Greece has led to the formation of a particular population group in
Greek society which shows acute signs of social exclusion. The dramatic change
of the political situation in the former USSR created economic decline and social
disorganisation. Ethnic Greeks, especially those of the eastern republics like
Kazakhstan, as well as those from Georgia and to a lesser extent from Russia,
experienced political and economic upheavals. As minorities they felt vulnerable, unprotected, and threatened by the ongoing political, religious and social
changes, due also to their earlier historical experiences of persecution, exile, etc.).
They chose to move to Greece which they had always considered as their
historical country of origin. They thought that in Greece they would find an
appropriate ‘repatriation’ which would permit them to live properly, covering
their needs for employment, education, housing, insurance and social welfare.
Their profile statistics as described above show a healthy population from a
demographic point of view, mainly in productive age, with reasonably good
professional experience and levels of education (especially those coming from
Kazakhstan). All these data reveal that the Greek labour market should be able
to easily absorb the ethnic Greeks of productive age, with some low additional
costs for professional retraining where necessary. However, the reported unemployment rates for all population age groups in both sexes – with especially high
rates of female unemployment – are much higher than the official unemployment rates for Greece. For reasons of comparison, it is worth repeating that the
official unemployment rate for Greece is 11.9 per cent, or 532,700 people, in 1999,
when our field survey was implemented. The corresponding figure for Attica is
12.7 per cent for 1999. Moreover we observe that the majority of the ethnic
Greeks who are employed are working in different areas from their early
profession in the former USSR, or they are underemployed. Furthermore, it is
quite disappointing that unemployment is higher for those with high levels of
education and professional experience. One of the most important reasons for
the high unemployment level of the ethnic Greeks seems to be their inadequate
knowledge of the Greek language and their difficulty in adapting to the reality
of the free market in Greece. The adjustment difficulty was all the greater since
their arrival coincided with a period of significant realignment in the Greek
production system, and by extension in the need for particular specialisations in
the labour market. If we add to the aforementioned the suspicion of Greek
society towards ‘outsiders’ and the acute psychological stress the repatriates
suffer during their efforts to adapt, their sense of social isolation becomes
palpable. This is something which the Greek state should never have allowed to
532
G. Halkos and D. Salamouris
become a reality. The support and extension of existing structures and systems
for effective professional and social integration is considered as a necessity.
The elimination of phenomena of social exclusion requires an organised
political intervention, which would activate and coordinate the responsible
services and utilise the existing structures at the national as well as at the local
level. The content of this policy should include above all the development of
employment capabilities and the materialisation of training programmes, including those focused on language competence. Housing is another area that
requires further attention, at least in the West Athens context. Any intervention
for the solution of these problems presupposes the determination of the particular characteristics of the repatriates; hence our work in this regard should lead
to effective planning of policy action. Also valuable in this context is the
experience acquired by the ethnic Greek ‘Centres of Social and Labour Adaptation’, developed from continuous and direct contact with the problems of this
particular population group.
It is important for ethnic Greeks to be able to access the labour market and the
legal economy. The creation of new work opportunities presupposes research on
skills identification and on the correspondence of the professional qualifications
of the ethnic Greeks with the demand for labour from the local economy. In
particular we see the need for all training programmes to be accompanied by the
learning of the Greek language, which has so far proved to be a major
impediment in finding work. In this way we hope to break the circular flow of
the repatriates’ exclusion and of the resulting poor social and housing conditions
of the areas of West Athens where they have settled.
Acknowledgments
We would like to thank Professor Russell King and two anonymous referees for
their very helpful and constructive comments on an earlier draft of the paper.
We also thank all the local ethnic Greek societies and in particular the ‘Pan
Hellenic Ethnic Greek Scientists Society’ and the Society ‘FAROS’ for their
cooperation.
Notes
1
Although these ethnic Greeks, also known as Pontian Greeks (after their region of historical
origin in Asia Minor), are often referred to as ‘repatriated Greeks’, technically this term is
incorrect since the migrants are not returning to their native land (where they were born) but to
a kind of ‘ancestral home’. For an ample discussion on this point see Keramida (2002).
2 It should be mentioned that in the period of their settlement in Greece they were supported by
EIAPOE, the National Foundation for the Support of Ethnic Greeks Settling in Greece. This
Foundation was established in 1990 and was very active in the early 1990s in creating reception
and educational facilities, including building thousands of houses, mainly in Thrace, to house
them (Fakiolas 2001; Ribas-Mateos 2000).
3 For the record, these studies comprise Fakiolas (2001), Georgas and Papastylianou (1993),
Hatzivarnava (2001), Kasimati (1992), Marbakis et al. (2001), Mavrea (1998), Terzidis (1995),
Tsoukalas (1994), Vergeti (1998).
4 The residential distribution of the ethnic Greeks in the ten municipalities we surveyed was as
follows: Aspropirgos 73.1 per cent, Aigaleo 7.8, Agia Barbara 7.0, Peristeri 3.6, Agia Anargiroi 1.6,
Kamatero 1.6, Ilion 1.5, Zefiri 1.4, Petroupoli 1.3, Xaidari 0.3 (no answer, 0.8 per cent).
5 However, an important note of caution should be sounded here. The survey data we collected are
not ‘true’ unemployment rates since they are not measured against the total numbers of persons
in each age/sex cohort who are potentially available for work, but against other notional totals
Integration of ethnic Greeks from the former USSR
533
in each age/sex group. Hence they are perhaps more accurately called ‘inactivity’ rates rather
than unemployment. This explains why the female rates are so high in Table 2. Nevertheless, if
we consider only males in the most economically active age groups, the indications are still that
ethnic Greek repatriates are twice as likely to be unemployed as the native Greek population.
References
Blos, M., Fischer, P.A. and Straubhaar, T. (1997) ‘The impact of migration policy on the labour market
performance of migrants: a comparative case study’, New Community, 23(4): 511–35.
Collett, D. (1991) Modeling Binary Data. London: Chapman and Hall.
Droukas, E. (1998) ‘Albanians in the Greek informal economy’, Journal of Ethnic and Migration Studies,
24(3): 347–65.
Fakiolas, R. (2000) ‘Immigration and unregistered labour in the Greek labour market’, in King, R.,
Lazaridis, G. and Tsardanidis, C. (eds) Eldorado or Fortress? Migration in Southern Europe. London:
Macmillan, 57–78.
Fakiolas R. (2001) ‘The adaptation of the ethnic Greeks to Greek society’, in Amitsis, G. and
Lazaridou, G., Legal and Socio-Political Dimensions of Immigration in Greece. Athens: Papazisis
Publications, 129–55 (in Greek).
Georgas, D. and Papastylianou, A. (1993) Cultural Cultivation of Repatriates from Albania and the Former
USSR in Greece: Psychological Procedures of Adjustment. Athens: General Secretariat of Emigration (in
Greek).
Glytsos, N. (1995) ‘Problems and policies regarding the socio-economic integration of returnees and
foreign workers in Greece’, International Migration, 33(2): 155–72.
Hair, J.F., Anderson, R.E., Tatham, R.L. and Black, W.C. (1998) Multivariate Data Analysis. Englewood
Cliffs, NJ: Prentice Hall.
Hatzivarnava, H. (2001) Social Exclusion and Ethnic Greeks from the Former USSR. Athens: Papazisis
Publications (in Greek).
Hosmer, D.W. and Lemeshow, S. (1989) Applied Logistic Regression. New York: John Wiley and Sons.
Kasimati, K. (1992) Greek Repatriates from the Former USSR: Social and Economic Accession. Athens:
General Secretariat of Emigration, Center of Social Morphology and Social Policy (in Greek).
Keramida, F. (2002) “‘Repatriates” or “refugees” and other vexed questions: the resettlement of
Pontian Greeks from the former Soviet Union in Greece and the politics of policy discourse’, Studi
Emigrazione, 145: 231–59.
Kleinbaum, D.G. (1994) Logistic Regression: A Self-Learning Text. New York: Springer-Verlag.
Kleinbaum, D.G., Kupper, L.L., Muller, K.E. and Nizam, A. (1999) Applied Regression Analysis and
Other Multivariate Techniques. Pacific Grove, CA: Duxbury.
Krahn, H., Derwing, T., Mulder, M. and Wilkinson, L. (2000) ‘Educated and underemployed: refugee
integration into the Canadian labour market’, Journal of International Migration and Integration, 1(1):
42–57.
Lazaridis, G. (1996) ‘Immigration to Greece: a critical evaluation of Greek policy’, New Community,
22(2): 335–48.
Lianos, T., Sarris, A. and Katseli, L.T. (1996) ‘Illegal immigration and local labour markets: the case
of Northern Greece’, International Migration, 34(3): 449–84.
Marbakis, A., Parsanoglou, D. and Pavlou, M. (2001) Immigrants in Greece. Athens: Hellinika
Grammata (in Greek).
Mavrea, K. (1998) ‘Dimensions of social exclusion: repatriates from Albania and the former USSR in
Greece’, Greek Review of Social Research, 96–7: 185–218 (in Greek).
Münz, R. (1996) ‘A continent of migration: European mass migration after the twentieth century’,
New Community, 22(2): 201–26.
Naxakis, H. and Chletsos, M. (eds) (2001) Immigrants and Immigration in Greece. Athens: Patakis
Publications (in Greek).
Ribas-Mateos, N. (2000) ‘Old communities, excluded women and change in Western Thrace’, Papers:
Revista de Sociologia, 60: 119–50.
Sharma, S. (1996) Applied Multivariate Techniques. New York: John Wiley and Sons.
Terzidis, K. (1995) ‘Aspects of the social and economic integration problems of the Greek repatriates’,
Greek Review of Social Research, 88: 100–21 (in Greek).
Tsoukalas, K. (1994) ‘Repatriation and division of labour’, Greek Review of Social Research, 84–5: 28–33
(in Greek).
534
G. Halkos and D. Salamouris
van Amersfoort, H. (1996) ‘Migration: the limits of governmental control’, New Community, 22(2):
243–57.
Vergeti, M. (1998) Ethnic Greeks from the Former USSR 1985–1995. Thessalonica: Kiriakidis Bros. (in
Greek).
Author details
George Halkos is Assistant Professor of Econometrics and Statistics in the Department of Economics,
University of Thessaly, Volos, Greece.
Dimitrios Salamouris is a PhD student in the Department of Economics, University of Thessaly.
They can be contacted at:
E-mail: halkos@uth.gr