185 covered by collective agreements were on average 78% higher than the agreed levels in the agreements (Neumann 2002: 25). In regards to sectors, the union wage premium was highest in those sectors of the competitive economy, which had high rates of unionisation and belonged to the traditional sectors of the Hungarian economy having survived transition. Furthermore, the service sectors of the economy (with the exception of financial services and public administration) showed the highest pay differences. This is related to the fact that services are notoriously low paid and union firms are scarce. Hence, the impact of the union wage premium in the service sector is larger (Neumann 2002: 16). However, wage rates in non-union firms in the dynamic branches of the manufacturing sector were higher than in union firms. These were those sectors, which were dominated by foreign-owned firms. TNCs, displaying low unionisation levels, paid higher than average wages across the board and contributed most strongest to wage increases (Neumann 2002: 16, 22). In contrast, the union wage premium and its contribution to wage increases was the smallest in indigenous firms (Neumann 2002: 22). Furthermore, intra-firm wage inequality was highest in Hungarian-owned enterprises without collective agreements. Again, the level of educational attainment, age and occupational status were the most influential factors determining intra-firm remuneration inequality. Indigenous firms employing mainly unskilled labour also had the highest share of incomes corresponding to the minimum wage (Neumann 2002: 25). The higher incidences of the absence of collective agreements as well as higher rates of unskilled low pay in indigenous firms are linked to their smaller size and the absence of trade unions (Neumann 2002: 25). Their higher cost sensitivity together with increased competitive pressure has led to a situation, whereby small and medium-sized indigenous employers have increased the amount of registered minimum wage recipients in their firms in order to lower their social and welfare contributions as well as their tax payments (Koltay 2002: 56). Employees invariably receive non-reported side payments or contributions inkind to supplement their official minimum wage or are encouraged to sell their skills to the firm via quasi-self-employment (Szanyi 2004: 198-201). As a result, figures concerning the true extent of underpayment are scarce and unreliable (Koltay 2002: 55-56). Nevertheless, these illicit practices contribute to precarious forms of employment and labour insecurity characterising indigenous employment (Ferge/Juhász: 2004: 235). 4.3.3 Social Implications The technological bias of transition has distinct social implications. The wage premium for education feeds through into the dispersion of household incomes. The educational level of the household head is the main determinant in influencing the 186 household income position. Furthermore, the level of education also determines the risk of experiencing poverty. The same socio-demographic groups suffering from low pay are also prone to experience poverty. Household Income Inequality Concerning the distribution of net household incomes, Molnár (2004: 12, 14) shows that only the highest four income groups were able to increase their income shares. Accordingly, simple inequality as measured by the top/bottom ratio increased by 0.5 points. In total, the level of household income inequality grew by 11% between 1993 and 2001 (Molnár 2004: 15). Furthermore, the growth in inequality was most pronounced between 1997 and 2001, when the economy recovered from the transitional recession and following the implementation of the 1995 austerity policies. The shares of the poorest income groups remained largely the same. In terms of educational profile, the educational wage premium rose considerably. The proportion of university graduates in the top income decile increased by almost 10% from 1993 to 2001. Moreover, the general level of educational attainment of the population increased throughout the observed period, as the share of primary school leavers dropped by almost 13%. This in part explains the lower levels of primary school leavers in the lowest income groups in 2001, although they still constituted the majority of households in the lowest income decile. Low income households with vocational training experienced the largest increases between 1993 and 2001 (Molnár 2004: 12, 14). These findings are inline with I.G. Tóth’s (2004: 80-81) decompositional analysis of household income inequalities. The level of educational attainment of the household head had an increasingly determining role. In 1987, it was responsible for 8% of inequalities. This explanatory share was 25% in 1996 and 23% in 2000. The employment status was the second most important factor. Inactivity of the household head was responsible for 12% of inequality in 1987 and could explain 11% in 2000. In 2003, 20% of household income inequality was down to either inactivity or unemployment. Five percent of income inequality in 2000 was due to the number of children and seven percent was the result of the household head being a member of the Roma ethnic minority. Similarly, Gábos and Szivós (2004: 103-104, 113-114) show that in their sample of socio-demographic groups situated in the lowest income quintile in 2003, 71% were members of the Roma minority, 47% were families with at least three children, 50% were single parents with children, 63% were unemployed, 53% were dependents and 48% had only primary school education. Income Poverty and Deprivation The lowest income decile and quintile are also used as one indicator for the poverty rate in Hungary, which until the harmonization of EU poverty thresholds through the 187 Laeken Indicators had no official poverty threshold. However, this income level is generally regarded to be too low. The income of the lowest income decile is lower than the minimum pension, which is used as an eligibility requirement for social assistance (Havasi 2002: 59). As Ferge and Tausz (2002: 195) show for 2001, the minimum monthly pension stood at HUF 18,000 (€ 74), in contrast the subsistence minimum was calculated to be HUF 26,000 (€ 107). An estimated 30% of the members of the lowest income decile had an income below the calculated subsistence rate in 2000. In 1989, this figure was only 10% (Ferge/Tausz 2002: ibid).149 In terms of relative income poverty, the most commonly used indicator is the threshold of 50% of the median equivalised household income. Hence, households are classified as poor, if their income is equal or below this threshold. The use of income bands allows identifying those households that are at risk at experiencing poverty and those whose incomes are below the threshold (Förster/Pearson 2002: 13). Again, this method is prone to underestimations, as real incomes fell during the 1990s. The decline in real earnings resulted in a reduction of the median income levels by more than a third and effectively reduced the poverty threshold (Förster/Pearson 2002: 13). Bearing general income developments in mind, poverty rates measured as 50% of average per capita incomes are higher than those figures based on the official equivalised median incomes. For instance, median incomes were 14% lower than mean incomes in 1995 (Havasi 2002: 59). Nevertheless, the Hungarian poverty rates measured as 50% of and below the median income line have remained stable. Table 9 Income Poverty Indicators for Hungary (%) 1991/1992 1995/1996 1999/2000 2000/2001 Net Mean Earnings (HUF) b 59,978 53,769 54,744 55,785 50% Mean 12.8 18.3 14.6 14.4 Poverty Gap 50% Mean 33.2 29.8 25.3 27.3 50% Median 10.2 12.8 9.1 10.3 Poverty Gap 50% Median 31.3 29.9 26.3 26.8 60% Median 5.5 a 8.0 10.9 9.3 Poverty Gap 60% Median 22.1 a 23.1 22.6 21.7 Quintile Boundary 30.9 31.2 25.5 26.7 a figures for 1993, b in 2000 prices Source: Gábos/Szivós (2004: 95) and Havasi (2002: 59) 149 In comparison the income boundary (i.e. the highest income) of the lowest income quintile was € 122 per month in 2000 (Havasi 2002: 58). The figures in Euros are calculated from MNB yearly exchange rate averages. 188 The table displays these measurement differences. An estimated 10% of the population in 1990 and in 2000 had an income of 50% and below the highlighted median household income, with poverty peaking at 13% in 1995 during fiscal adjustment and following the transitional recession (Gábos/Szivós 2004: 95). According to the calculations shown in the table, almost 13% of the population received a monthly income equivalent to or less the mean of ca. € 296 in 1991. In 2000, 14% of the population earned lees than the average income of ca. € 105 per month.150 Concerning the severity of poverty, the figures relaying the poverty depth, i.e. the share of those classified as poor whose income is below the defined thresholds, a slight improvement has taken place. This is linked to the resumption of economic growth after 1996/1997, which raised all incomes (Gábos/Szivós 2004: 96). These developments are also displayed in the evolution of those incomes lying below the quintile boundary of the lowest 20% income group. Hence, 27% of those members of the lowest quintile received incomes below the boundary income limit of 4.6% of the mean household net income in 2000 (ca. € 122 per month). The 60% median threshold (€ 312 per month for 2001) relays the proportion of the population, who are at risk of experiencing poverty (Gábos/Szivós 2004: 98). Following these calculations, the share of households at risk of experiencing poverty was slightly smaller. However, the proportion of households earning less than the threshold of 60% of median net incomes declined only slightly by one percent, remaining at ca. 1/5th of all households throughout the observed period. Unsurprisingly, the socio-demographic groups experiencing poverty match those groups previously described as members of the low-income quintile and experiencing low pay. They constitute Hungary’s working poor. Again the level of educational attainment is the most important factor. In 2003 16% of the households experiencing poverty were headed by persons with only primary education. One in four of poor households had heads with uncompleted primary school education. Single mother households accounted for 12% of the poor. Households with the main earner either unemployed or inactive without a pension showed poverty rates four times higher than the recorded average. Furthermore, the poverty risk increases with the number of children. Finally, the poverty rates for Roma households are seven times the average (Gábos/Szivós 2004: 102-103). In terms of material deprivation, the figures are again similar in regards to the socio-demographic groups experiencing poverty. Assessing the deprivation of the poor make it possible to indicate the level of social exclusion in the country. Those experiencing deprivation lack the resources to participate in every day life. In a materialist, consumerist society, the materially deprived are in danger of stigmatisation and ultimately face greater hurdles to escaping poverty (Spéder 2002: 150). According to Spéder’s (2002: 158) estimations, those classified as poor and suffering from deprivation was four percent for the those below the 50% poverty line and seven percent for those at risk of poverty with incomes below the 60% median 150 Average earnings are taken from Havasi (2002: 56). The figures in Euros are based on MNB yearly average exchange rates. 189 income threshold in 1999. Again, unskilled workers, single mother households, families with more than two children and members of the Roma minority are those socio-demographic groups deemed to be both poor and deprived (Spéder 2002: 158- 159).151 4.3.4 Inequality, Poverty and the Hungarian State As shown, the Hungarian development strategy had distinct social effects. However, it would be short-sighted to place the full responsibility on the inflows of FDI. The social implications of the strategy are also the result of the state’s educational, social and income policies. In the case of Hungary, the role of the state in these policy areas has contributed to deepening and to sustaining market-generated inequalities and poverty. Attempts by the state to reduce the individual costs of transition created social costs in the form of higher contributions to the social insurance system and rising state indebtedness. However, the social costs have not been evenly distributed. Middle and higher income groups have been able to complement their marketgenerated incomes by benefiting from wide ranging tax allowances leading to a regressive tax system. Transformation of the Hungarian Welfare State The process of transition fundamentally changed the country’s social and welfare regime. On the one hand, social welfare institutions were created and social citizenship rights strengthened. On the other hand, social policy repeatedly has been affected by expenditure cuts. These have affected the redistributive nature of the implemented instruments. In similarity to other transition countries, the transformation of the Hungarian welfare since 1990 is seen to have entailed three overlapping elements. The first element was characterised by the response to the economic decline of the SOE sector. The SOEs were the main element of the socialist welfare system (Wagener 2002: 156). The liberalisation and privatisation of the economy relieved the enterprise sector from its previous social tasks. Furthermore, the liberalisation of prices resulted in the abolition of price subsidies, which had constituted a major egalitarian instrument of the former regime (Wagener 2002: 157). The state introduced an insurance-based social welfare system with the establishment of health and social security funds financed by employer and employee contributions. The chosen models followed the “the conservative-corporatist Bismarckian model” (Wagener 2002: 160). The introduction of a mutually financed and 151 For the exact proportions and material deprivation definitions, see Spéder (2002: 162-165).