BETWEEN STATE SOCIALISM AND MARKETS: Effect of Education on Incomes in 27 Countries

A major consequence of the rise of the capitalist labour market throughout East European societies was the tremendous increase of the financial value of education (Vecernik 1999; Domanski 2001). One of the most convincing interpretations of this change pointed to the growing role of meritocratic rules of distribution of incomes. In its original formulation, given by Michael Young (1958), meritocracy was referred to rules of fair distribution of rewards. The widely accepted definition says that meritocracy exists when individuals are rewarded according to their merits. In the famous equation by Young, merits were operationally defined in terms of ‘IQ plus effort’. Certainly meritocracy serves as a hypothetical model, which – in contemporary societies – comes through in promotion of most intellectually gifted, ambitious and, first of all, most educated individuals, who possess rare and highly demanded skills (Marshall et al. 1997). The concept of meritocracy laid the ground for an influential paradigm of interpretation of social inequalities in Western societies (Bell 1973; Saunders 1996; Breen and Goldthorpe 2002).

There is a good reason to believe that this paradigm could embrace transformations in post-communist societies. Interpretations referring to the rise of the meritocracy dwell on regularities found in the past, before the fall of the communist system. The size of association between education and incomes was extensively examined, based mostly on survey data obtained in Poland, with generally unambiguous results. They showed that effect of education, as measured in years of schooling, was much lower in Polish society, at the time, in comparison to Western countries (Meyer et al. 1983; Słomczynski 1989; Domański 1993). According to prevailing interpretation, these findings confirmed the core claim of theoretical reflection on specific features of social stratification in planned economies – namely that it was inherently flawed by ‘decomposition’ of educational investments and financial rewards (Wesołowski 1978; Słomczynski and Wesołowski 2002).

This thesis on decomposition, which was regarded as the differentia specifica of the communist stratification, became the natural starting-point for major interpretations that were incorporated into the market transition debate, which emerged after the collapse of the communist regime. The main conceptual feedback came from theories of modernisation. Most research has seemed to confirm their anticipation that with the passing from traditional to modern types of social systems, the role of individual achievement would increase the role of an ascription decline. Recent analyses, such as DiPirete and Grusky (1990), Erikson and Goldthorpe (1992), and Hout (1988), provided detailed empirical elaboration of the themes of increasing openness of social barriers and declining ascription. Theoretical anticipation of this appears as early as Parsons (1949) and was clearly summarised in Treiman's (1970) account. According to the thesis, referring to the ‘logic of industrialisation’, the increasing efficiency of production in large organisations shifts decisions governing occupational outcomes to bureaucracies. Universalistic criteria such as education gain force, the ability of kin networks to perpetuate rank fades, and intergenerational mobility rises.

The central component of the decline of non-meritocratic criteria is the claim that schooling looms ever larger as the primary rank of distribution of rewards. The functional interpretation holds that this reflection of expanding universalism should increase with time (Kerr et al. 1960; Treiman 1970; Erikson and Goldthorpe 1992). ‘Opportunity (meritocratic) society’ has to emerge since the logic of modernisation dictates that candidates to occupational positions have to meet formal requirements regardless their origin, sex, race, as well as membership of political party and ideological affiliations. Education becomes the key to high rank. At the same time, more educated individuals must be properly rewarded. Paralleling demands for skilled and responsible incumbents of more functionally important roles, one needs distribution of financial rewards, which would motivate them to fill them and perform them in an efficient way. Resulting from this is the functional necessity to maintain relatively strong correlation between education and incomes.

Theories of convergence, in pursuing this track further, predicted that social stratification in communist states would approach modern capitalist societies. Bell (1973) and Blau and Duncan (1967) concluded that, across many levels of economic development, patterns of stratification were increasingly approximating the United States. Such notions might suggest that convergence should gain momentum in the post-communist world. It proved a fruitful avenue in interpretation of the increasing effect of education on incomes in post-communist countries.

It is true, that the views presented here are not universally accepted. Although the critical role of formal schooling in determining incomes is well established in the sociological and economic literature (Spilerman and Lunde 1991), in the last three decades neoclassical economics as well as sociological stratification theory have been widely criticized for neglecting the impact of institutions (Kalleberg and Sorensen 1979; Parkin 1979; Erikson and Goldthorpe 1992). Many sociologists became interested in understanding how educational credentials are used as a mode of closure in distributive struggle, and how they relate to the organization of the employment system and to individuals’ education returns. For this purpose, comparative labour market research with involved societies differing and changing institutional context has proved to be highly important (Kerckhoff 1995; Shavit and Muller 1998).

Notwithstanding this critique, there are certain reasons to believe that, in the 1990s, incomes were related to education. Higher education in particular was supposed to become more relevant because of technological progress imposing higher demands for technical and academic expertise on the one hand, and because of general increase in the supply of higher-education graduates on the other. Vecernik (1999) and Domański (2003) documented this tendency for the 1990s in the Czech Republic and Poland. Bian and Logan (1996) and Zhou (2002) reported a significant increase in the income return to human capital in urban China, although recent studies in the Chinese case are riddled with contradictory evidence – for example, Xie and Hannum (1996) found exactly the opposite. According to Gerber and Hout (1998), market reform had not made significant headway in Russia. They analysed data collected in the 1990s and – as the income analysis suggests – there was no evidence for increasing return to human capital from 1991 to 1995. Gerber and Hout stated that political capital in Russia (made by current and former cadres) outstripped the gains of the educational capital on the whole. Basically, the only exception to the pattern of increased returns was that in the former East Germany; the rate of return to a year of education falls in the early part of transition (Kruger and Pischke 1995). However, the type of transition that occurred in East Germany was unique. Because of the special relationships with the former West Germany, East Germany was pulled immediately into a powerful market system with competition for skilled workers in the West. It did not go through the slower transition which took place in the other East European countries whose labour markets were insulated.

Among existing empirical studies, analyses for Poland stand out for their documentation of an almost linear increase of education effect at consecutive points in time, starting from the 1980s (Domański 2003). While consistent with predictions of modernisation and convergence, the empirical evidence is far from unequivocal in the key question: to what extent East European societies approached Western patterns as far as monetary returns to education are concerned. The limitation of almost all studies lies in that they have used nationally bounded research designs. Certainly, they are inadequate for studying potential shifts of post-communist stratification to the West. The only way to unsettle this question is to determine the direction of change, which must rely on direct comparison of data from both Eastern and Western societies. In this study I take a step to address these problems. My analytical focus is on a study of the relationship between education and incomes in 27 countries, reinterpreted in the light of the theories referred to above. The use of strictly comparative design allows one to assess prospects of convergence in a more satisfactory way. However, a caveat is in order. I report only an analysis of income determination in 1996–1999. The state of the art in the field does not permit us to extend this strictly comparative design back to the period before the collapse of the communist regime. Henceforth the foregoing analysis will be far from conclusive in the determination of how far the distribution of incomes in Eastern Europe approaches the Western model in the long term. One can only establish how it developed in the 1990s, since comparative data for communist societies for the 1980s do not exist.

An important implication of this empirical blind spot is that we are left with a large area of indeterminacy with respect to the mechanism of systemic transition. However, we can reduce it and improve the power of our theory in showing an interconnectedness between stratification and emerging markets in the course of institutional changes in Eastern Europe. I turn to this now, drawing upon comparative data from 27 countries from 1996 and 1999 from the International Social Survey Program.

One can set out three competitive scenarios with regard to the ‘recomposition’ of education and incomes in transitional economies. First, this association remains meagre despite its increase in the 1990s, which was empirically documented for the Czech Republic and Poland. Implicit in this hypothesis is a cautious forecast of rise of a meritocracy in the distribution of incomes in coming years. Because the former centrally planned economies inherited a system of low-wage differentials between groups with different skills, one could expect the rate of return to education to increase as these countries moved towards a market economy. But the human capital and experience gained under communism may not be very useful in an emerging market economy and the inherited education system will not necessarily promote fungible skills (Terrel 2000). If correct, this argument suggests that rate of return would persist during the transition period. In this regard, one may assume, at best, that institutional transformations in Poland did not bring about distribution of rewards that are typical for developed market economies. If true, this prediction would be empirically confirmed by relatively smaller monetary returns for education in post-communist societies as compared to the Western countries.

The second more optimistic scenario assumes that we are already approaching market economies if considered in terms of benefits from educational credentials. In this case, one should not expect a significant gap between East and West in the size of this effect in the late 1990s. Building on this we might hypothesise that the emergence of capitalist labour markets really will entail a modern logic – emphasising human capital versus (for example) political connections. This framework has in fact been central to market transition theory from the onset (Szelenyi and Costello 1996). This interpretation is seriously questionable in that it fails to take into account comparative data from the 1980s. One cannot exclude the possibility that, even in the past, any gap between market and communist economies existed with respect to effect of education on individual incomes. In view of the latter interpretation, the following difficulty arises: to the extent that human capital conferred the same financial rewards in post-communist and market societies in the 1990s, this might reflect lack of changes instead of a move towards the West in distributive patterns.

One can add a third scenario which lies in between: the effect of education in East European societies moves forward, although it still remains behind that taking place in developed capitalist societies. Construed as such, the rising significance of the meritocratic distribution thesis does not require roughly the same financial returns to education in both systems. They remain distinct but not as much as two decades ago: in such a case, if in post-communist countries benefits from human capital increased, it may suggest that the spirit of capitalism has slowly pervaded the distribution of incomes.

Data for this analysis were obtained from the International Social Survey Program. To investigate cross-national variability in education returns, I collected data from a relatively large number of advanced capitalist and post-communist societies. The ISSP, which has been carried out since 1985, provides the best source of data with which to systematically test hypotheses about convergence in distribution of incomes. For each country, random sampling designs are applied with the size of samples varying from 900 to 2500 men and women aged 18 and over. For the purpose of our analysis, the samples are confined to economically active people.

All variables were coded in the same way to maintain common codes across countries. We take advantage of the existence of suitable items which with to measure individual incomes, education and relevant statistical controls. In 1999 the ISSP covered 33 countries, and in 1996 32 countries. However, in this study I employed 27 national data sets for 1999 and 16 for 1996, after leaving countries having a large number of missing data and those for which the reported income was logically suspect. These ISSP data contain information on key variables of interest at application to address the hypotheses developed above. I used the logarithm of personal income as a dependent variable. In the ISSP surveys, information on respondents’ total incomes are collected. Thus, personal income reflects both rewards from current job and other sources based one one's skill or position. In order to capture ‘effect of meritocracy’, I referred to formal education, which is a conventional measure of merits: I operationalized it in number of years of schooling (from 0 to 20). As regards independent control variables I adopted, whenever possible, generic categories whose meanings are widely shared to avoid artificial ambiguities in interpretation. I use gender as a dummy variable (male=0) to examine gender-based variations in the redistribution of economic rewards. It is also reasonable to assume that one's market position depends on access to opportunities and resources, which is significantly affected by positional power. For this reason I control for supervisory position expressed as a dummy variable (coded 1 for supervisors, otherwise 0). In a planned economy, state-owned firms have been major institutions of redistribution. To the extent that expansion of the market caused development of private firms, we had to take into consideration division into state and private sectors which are distinct as regards rules of distribution of incomes. Sector of economy is coded 1 for private firms, otherwise 0. I also took into account age – as measured in number of years – and weekly hours of work.¹

Previous analyses suggested that the most responsible factor for decomposition of education and incomes be embedded in peculiar features of planned economy. State communism inevitably produced a variety of mechanisms which defied logic of market and meritocratic rules of rewarding people. Several streams of explanations emphasised that allocation of financial resources directed by state tend to decrease effect of educational credentials on incomes as compared to the allocation driven by the market rules (Nee 1989; Domański 1990; Kornai 1992). The first step is to determine the size of this effect in the late 1990s when capitalism only started to emerge in East-European societies.

As a preliminary analysis, I use the conventional multivariate OLS regression model to examine income determinants in selected countries. This set of analyses aimed at detecting net monetary increments of income from years of schooling. The parameter estimates for the latter in metric and standardised forms is presented in Table 1. The estimated models also included a set of control variables (i.e. sex, age, supervisory position, sector, hours of work, and size of residential location). Table 1 summarises net effect of education for 27 countries in 1999. Due to a lot of missing information on incomes, I deleted Chile, The Netherlands and Russia from the original ISSP data set. I also excluded the former GDR and Northern Ireland, which were rendered irrelevant from the viewpoint of analytical focus of this study.

Let us be clear that post-communist societies are represented here by seven countries, e.g., Bulgaria, the Czech Republic, Hungary, Latvia, Poland, Slovakia and Slovenia. On first sight, one can hardly establish whether returns to education in these societies are greater, smaller, or may be the same, relative to developed capitalist economies. In looking at Table 1, we see that in 1999 years of schooling most strongly affected incomes in the Phillipines, with a net increase of 12% of incomes per one year. The magnitude of this effect also appears relatively large in England, Japan, Norway and Switzerland, where net return to education was about 10%. It showed the lowest, though, in Cyprus, Spain, the Czech Republic and Slovakia, where one year of schooling yielded a 3% increase. It stood for little more – about 4% – in Canada and Sweden. Then we have a set of countries with 5% net returns for meritocratic achievements, which include Austria, Ireland, Israel, New Zealand, Portugal, and (out of post-communist countries) Hungary, Latvia, and Slovenia. Poland is more or less in between, with higher returns than in the United States, and the same as in Australia and West Germany. In Poland one year of schooling increased income by 7%, which provided it first place among East-European societies.

We can detect rough changes in income regimes in the transitional period as reflected in a comparison of income determinants in 1996 and 1999. Table 2 presents parameter estimates of the same model for 1996. We restricted analyses for 16 countries (including six from Eastern Europe) for which strictly comparable data exist. The estimates for 1996 show that, as in 1999, there were no discernible patterns in the magnitude of returns to education across countries. Data show significant and varying effect across individual societies, but one cannot give a substantive interpretation of it in terms of degree of modernisation, unequal economic development, historical tradition, culture, as well as with respect to division between post-communist and capitalist societies. It shows that in 1996 educational attainment performed best in Slovenia, increasing incomes by 0.09, and had the lowest impact in New Zealand where it stood for 2% increment for one year. It appeared among the lowest also in the Czech Republic, Latvia and Sweden, where a one-year increase in schooling was related to a 3% increase of monetary returns. At the same time, Poland may be counted as a country belonging to the group representing a moderate level of meritocracy. The 0.06 coefficient for education was lower in 1996 than in 1999, although substantial (in magnitude) and warrant to say that human capital played an important role.

To determine the key question, concerning convergence of the post-communist stratification to the market patterns, we performed a simple test by comparison of average metric coefficients of years of schooling. For the post-communist bloc they were derived by calculating average effects for seven nations in 1999 (presented in Table 1) and for the available six in 1996 (from Table 2). For definition of the ‘market economies’, I used three alternative specifications ranging from the most comprehensive set of countries, through moderate, to the narrow. Theories of modernisation are far from conclusive in adjudicating between competing conceptions of which countries constitute the ‘true developed West’. Notwithstanding various controversies, they tend, mostly, to identify this concept with Western Europe, adding to this – originating in European civilisation – the United States, Canada, Australia and New Zeland. I took this idea as my starting point to distinguish three alternative measures referred to in the foregoing analyses to the ‘type of system’.

The first set of Western societies embraces all West European countries, except Israel and Cyprus, including the United States, Canada, Australia and New Zealand. It includes all principal economic regimes labelled by Esping-Andersen (1999) ‘social democratic’ (basically the Nordic countries), ‘conservative’ (Continental Europe) and ‘liberal’ (the Anglo-Saxon nations). All in all, this widest category consists of 16 societies for the 1999 data set (nine for 1996), leaving aside Japan and Phillipines. I contrasted this set of countries with seven East-European nations (for 1999) to develop a dichotomy variable ‘type of system 1’ (Western countries=1).²

The second definition of Western societies focuses more narrowly on the mostly developed economies, excluding out of 16 cases Portugal and Spain. Finally, in case of the most narrow definition of Western societies, we restrict them to the nine countries (for 1999) which seem closest to the market principles in matching individual incomes to merits, i.e. where the effect of education is relatively least mediated by the redistributive policy of the state. According to Esping-Andersen (1999), the liberal, Anglo-Saxon system emphasis on targeted means-testing for the poor, and private occupational plans for core workers will reinforce rather than mute market-based stratification. Quite another configuration emerges in the comprehensive Scandinavian welfare states. The accent on universal benefits and services creates a more homogeneous population in terms of distribution of social resources. Although no clear-cut patterns exist, there is some evidence that the types of social democratic regimes stand out for their greater degree of co-ordinated and centralized bargaining system, higher minimum wage, and stronger protection against dismissal. The opposite are liberal, and more deregulated economies that have experienced a sharp increase of earnings inequality since the late 1970s.

This third group includes Australia, Canada, England, France, Ireland, Italy, New Zealand, Switzerland and the United States. I did not take into account Scandinavian countries, which are highly marked by institutional arrangements of welfare policy, nor Austria and West Germany, which belong to the ‘corporatist’ (conservative) version of capitalist economy (Esping-Andersen 1993). In the case of the 1996 data set, I include in this group only six countries, without Canada, Italy and Switzerland, due to lot of missing cases. Certainly it is often extremely hard to draw a clear boundary between the various variants of capitalist economies and, at the same time, one can indicate substantial differences between the Untied States and Italy, or France. Despite methodological convenience these three distinctions fail to take into account the complexity and variability that characterise all large-scale societal environments.

Table 3 summarises this step by comparing the means for post-communist and capitalist countries for 1996 and 1999. Means for the former are presented in the first row. The key lies in contrasting them with means presented in the three bottom rows which show mean size of returns to educational credentials for Western countries. The latter are presented, consecutively: (i) for the most overwhelming group of sixteen countries (second row), (ii) for the fourteen countries, i.e. without Portugal and Spain (thirrd row), and (iii) for the nine countries which most closely resemble meritocratic distribution of incomes in that individuals maximise their returns on educational assets (fourth row).

As one can easily see, years of schooling increased income significantly less in the former communist camp, compared with the size of this effect in Western economies. The first row's coefficient shows that in 1999 increases in returns to educational levels in post-communist societies, on the whole, were 0.048, and in 1996 it was much the same, amounting to 0.050. Corresponding means for Western societies were higher, especially for 1999. The net return for 16 countries was about 0.062, for 14 it was 0.068, and for nine it was 0.066.

As Table 3 shows, education appeared less important as an asset in post-communist countries. This may suggest that they still lag behind the logic of market income regime. In order to test this interpretation in a rigorous way, data for analysed counties were pooled. I estimated the ‘main effects’ of years of schooling (alongside with other controls) and the ‘interaction effect’ between this covariate and the dummy variable ‘type of system’.

At this step we turn from simple cross-sectional research design to the analysis of data defined at different levels of observations, such as countries and individuals within these countries. The design of such studies reflects several specific methodological problems summarised in a term of ‘multilevel problems’ (see Bryk and Raudenbush 1992; Hox 1994). To investigate the independent (additive and interaction) effects of individuals and contextual variable referred to as ‘type of system’, a design must be used that leads to low or preferably zero correlations between individuals’ characteristics and contextual variable. The sources of both within- and cross-subject ‘clustering’ violate the independently and identically distributed assumption in the OLS estimation, and one needs to explicitly model the covariance structures that deal with these sources. Because respondents are hierarchically nested within the systems, a hierarchical model must be used for the analysis of individuals and system effect. I employed the well-known hierarchical linear regression model also known as the random coefficient model, which has been described in several articles and books (e.g., Mason et al. 1988; Bryk and Raudenbush 1992). The hierarchical linear model HLM 5 procedure (see Raudenbusch et al. 2000) can deal with these problems and is adopted in the estimation model in this study.

The parameter estimates for these three models, using the three alternative specifications of the ‘type of system’ for 1999, are summarised in Table 4. Columns 1, 2 and 3 show the main effects of sex, age, years of schooling, etc., and interaction effect of years of schooling and ‘type of system’. The positive and significant effect of the ‘type of system’ (for models 2 and 3) indicate that in the transitional period, residents of Western societies had (on average) higher incomes as compared to inhabitants of post-communist societies. It should be mentioned that there was no statistically discernible difference between post-communist and Western countries with the latter understood most widely, i.e. when 16 countries were included. The slope coefficient for the ‘type of system’ (−1.26) did not benefit people living in Western societies, other things being equal. It appears to be substantively higher when post-communist countries are contrasted with 14 Western societies – after excluding Portugal and Spain (model 2) – and with nine contras (model 3), narrowed to the ‘true market economies’. The net returns to living in the West, defined as consisting of 14, or nine countries, were, respectively, about 90% (b=0.9) and 187% (b=1.87), other things being equal. This implies that during first decade of transition of Eastern European countries to the capitalist system there remained sharp difference in individuals’ incomes.

Certainly, we are mostly interested in the question of whether our contextual variable really affected educational returns. According to the results from the ‘interaction effects’, we find that education returns were significantly associated with living in Eastern, or in Western societies. The positive and significant effect for interaction terms (for each model) shows that residents of Western societies enjoyed higher educational returns. A one-year of schooling increase in Western countries resulted in about 2% higher income, compared with Eastern European societies. Overall, decisive changes in the 1990s, with both expansion of the market and the weakening role of the state, did not lead to a convergence of the patterns of income determinants between the two systems. These findings seem to suggest that economic changes involved multifaceted processes that produced substantial changes in some areas but persistence in others.

As regards the main effects, the logic of multilevel design allows interpretation of regression slopes in terms of average cross-country effects. They inform us about global mechanisms of distribution of incomes as averaged for the 27 societies under the study. The parameter estimates for years of schooling indicate that the net increment for each year stood at 3–4% in 1999; gender gap amounted to 25–26%; supervisors had significantly higher incomes than rank and file (the reference category) by about 25–26%; and each additional hour of work gave net premium of 1%.

One may assume that, at the end of the 1990s, mechanisms of distribution changed as transitional processes accelerated. In order to detect these earlier patterns I estimated the same income determinants for 1996 using two models for the two alternative specifications of the ‘type of system’ (Table 5). In model 1, East-European countries are contrasted with nine, and in model 2, with six Western countries regarded as best fitting the rules of meritocratic distribution. These findings reinforce the interpretation based on the results in Table 3, which suggest that in 1996 differences between post-communist and market economies were less salient with respect to association between education and incomes. As the regression slope for interaction term indicates, interaction between effect of education and type of system proved statistically significant only for model 1. The magnitude of this effect for Western countries appeared to be larger. In 1996, returns to education were 2% higher. As regards all other effects, they were patterned in much the same way in 1996 and in 1999.

Before we discuss the estimated patterns of incomes, it is necessary remind to ourselves that the consensus view is that stability and continuity are the backbones of social stratification. In the course of history we were rarely faced with abrupt changes, such as revolutions. The fall of the communist system became a turning point in the transition of East-European countries to a market economy. The market transition theories (see Cao and Nee 2000) attributed the main impetus of changes in mechanisms of stratification in post-communist countries to the emergence of market institutions. Indeed, in testing modernisation and market theories’ predictions, the most unambiguous empirical results on social stratification reported changes in the distribution of incomes. Processes associated with market-based power appeared to lead to higher returns to human capital than under a planned economy – it was evidenced in systematic increase of association between education and individual incomes. Confirmatory evidence was reported for the Czech Republic, Hungary, Poland, and Slovakia, where higher educational credentials appeared consequential in its marketability under the new economic regimes.

Nonetheless, the crucial question remains: to what extent has the effect of education changed over time? Existing empirical studies were qualified in that they did not explicitly attempt to determine how far the new rules of income distribution in Eastern Europe resemble opportunity structures in developed market economies. Implicit in the theory of modernisation is the idea that returns to education are the highest in the most developed capitalist societies, which are identified with modern Western countries. They should strengthen over time because technological and economic efficiency require that ability and motivation, as developed within educational systems, have to become the dominant criterion of selection in labour markets. This requirement is met through systems of qualification and certification and through personnel policies of employing organisations. Thus ‘achievements’ progressively prevail over ‘paricularism’ and ‘ascription’.

Bearing this in mind, a crucial issue was to check whether post-communist societies entered such a defined path of modernisation. This empirical blind spot was evidenced in the fact that empirical studies on convergence of stratification systems relied on indirect measures. To determine this question, I focused on a direct comparison of effect of education on incomes in a number of countries representing both capitalist and post-communist societies.

Our analysis provides evidence that, at the end of the 1990s, economic benefits to education were still less in countries undergoing transition than in capitalist societies. Two further questions that in turn arise are the following. First, these results provide only guidelines for understanding the most interesting question: to what extent post-communist societies approached market-based patterns of distribution of incomes. Although we went beyond former studies, our findings display only a static picture, restricted to 1996 and 1999. Certainly, we fall short of comparative cross-systemic data, on education returns, going back to the 1980s, i.e. before the communist system collapsed. Having these caveats in mind, one has to concede that we will never find unequivocal empirical arguments, which proved that effect of education under the planned economy was really lower. Contrary to many theories, one cannot exclude the idea that this association might have been higher or the same as in the Western societies at the time.

The second question concerns the changing meaning of meritocracy. One cannot exclude the idea that in the mostly developed economies formal education may be less and less important in the determination of incomes as compared to more marketable meritocratic assets, such as: high motivation, perseverance, or ability to learn quickly – that seem to enter into employers’ decisions. The growing importance of the new types of credentials, specialized courses, etc., may explain why years of schooling turned out to matter more in the 1990s in Poland in comparison to the United States and some other countries that were placed higher on the scale of modernisation.

Henryk Domański is Professor of Sociology in the Institute of Philosophy and Sociology Polish Academy of Sciences in Warsaw. Since July 2000 he has been Director of the Institute, and Head of Department of Social Structure Research and Department of Studies on Methods and Techniques of Sociological Research. His main fields of interests are studies on social stratification and mobility and methodology of social research. He has conducted numerous surveys and was a chair of several projects. He has authored 20 books (mostly in Polish), primarily on labour market segmentation, inequality of sex, comparative social stratification and methodology of social research. Among his most important publications are: ‘Dimensions of Social Stratification. A Comparative Analysis’ (issued as Vol. 19 of International Sociology in 1989), On the Verge of Convergence. Social Stratification in Eastern Europe (2000) and Women on the Polish Labour Market (2001).