European Welfare Regimes : Political Orientations versus Poverty

This inquiry analyzes how political orientations shape welfare states and labour market institutions when seeking to reduce poverty. In order to identify effects of these two key variables, we conduct a panel regression analysis that includes two poverty measures: poverty rates before and after social spending. This inquiry considers 14 EU countries, and in the period from 1995 to 2008, which are grouped according to welfare state regimes. We consider Social Democratic, Corporatist, Mediterranean and Liberal welfare state regimes. Panel regression results indicate that political orientation engenders no significant statistically measurable effects on poverty rates before social spending. Effects register, however, as significant when considering poverty rates after social spending. With respect to the first set of results, we advance two key explanations. First, we note a longer period of time is necessary in order to observe actual effects of political orientation on market generated poverty. Second, political parties with their respective programs do not register as influential enough to solve social problems related to income distribution when taken alone. Influences register as indirect and are expressed through changes in employment rates and social spending. The second set of results support the hypothesis that a selected political regime does indeed contribute to poverty reduction. In sum, political orientation and political regime does indeed affect poverty through welfare state institutions, as well as through labour market institutions.

Connections between policies of left and right-oriented political parties, income distribution and levels of extant poverty has attracted attention of numerous researchers, and especially economists and sociologists. Curiously, exertion of political influence registers as more indirect rather than direct because actions of political parties shape market institutions and also welfare states: key mechanisms through which society mitigates distributional challenges. Influences exerted by market institutions on poverty rates are manifested primarily through changes in employment rates. In economic theory, there is almost a universal agreement that unemployment serves as an underlying source of poverty. (De)regulation of labour markets could indeed cause increases in employment rates for groups faced with falling or remaining under poverty lines. It is the active approach, representing a transition for the "poor" from a category as beneficiaries of social aid to the category of net taxpayer, which not only PANOECONOMICUS, 2011, 5, Special Issue, pp. 651-674 reduces poverty but also by increasing national income, and also creates conditions for a sustainable implementation of passive policies. Welfare state institutions deal with levels of poverty primarily through social spending. By changing the extent and structure of social transfers, there are also the changes in the level of income of the poor population, affecting their positions in the distributional matrix relative to set poverty lines. We consider distribution in this manner as essentially "passive" approach as the social problems of and related to poverty are dealt with through the redistribution of national income, regardless of influences on levels of actual or potential national income.
In this sense, we could qualify that a main objective of the welfare state is to reduce poverty while also dealing with economic inequality. Hence, theories considering welfare states provide an adequate framework for the understanding of levels of measurable effects engendered by political orientations, and especially the key relationships between market institutions and social spending and their effects on levels of poverty. However, theories of welfare states should not be considered unique with respect to importance attributed to relationships between political variables and poverty, on the one hand, and then labour markets and welfare state institutions, on the other. In general, it is possible to distinguish between two approaches: (i) structuralan approach that emphasizes the importance of structural-functional factors, and (ii) a political-economic approach that emphasizes roles played by political factors. The political-economic approach emphasizes the key role of political orientation in determining the level and dynamics of poverty in developed democracies. Market and welfare state institutions are strongly influenced by the political processes. The dominance of the left-wing parties is associated with the strengthening of the welfare state institutions while the right-wing parties give emphasis on the deregulation i.e. the implementation of market mechanisms in solving social problems.
On these bases, it is possible to further distinguish between two basic approaches in the implementation of social policy: (i) liberal -primarily favoured by conservative and right wing parties and (ii) the redistributive -closer to social democratic and leftist parties. The division of the liberal welfare into the state and redistributive models represents a rough theoretical framework within which specific types of welfare states are profiled, also recognizable in the specific practices of individual countries. Having in mind the diversities present in the sources, the EU is characterized by four main models: the Social Democratic, Corporatist, Mediterranean and the Liberal welfare state regime.
The aim of this paper is to determine to what extent is the impact of political orientations to poverty compatible with the given theoretical framework of the welfare state, starting from the structural relations vs. social-economic, liberal vs. redistributive approach, as expressed through the practice of the Social Democratic, Corporatist, the Mediterranean and the Liberal welfare state regime.

Literature Survey and the Methodological Framework
In the literature, a greater number of both quantitative and comparative historical approaches to the settings of the political-economic approach to welfare state were accessed. In accordance with their significance, the following are most prominent PANOECONOMICUS, 2011, 5, Special Issue, pp. 651-674 (John D. Stephens 1979Stephens , 1995Stephens , 2001John Myles 1984, 2002Gösta Esping-Andersen 1985, 2007Francis G. Castles 1996, 2002. Most research focuses on the influence of the left-wing parties on the institutions of the welfare state and the consequent poverty and economic inequality (Donald Sassoon 1996;Erik Olin Wright 2000;David Brady 2003;Lane Kenworthy 2004;Andrew S. Fullerton and Jennifer Moren-Cross 2009). However, in recent years, there has been substantial research of the impact of right-wing parties (Clarence Y. H. Lo and Michael Schwartz 1998; James P. Brady and Kevin T. Leicht 2008). Most of the findings are in compliance with the conventional view of the welfare state by which the differences in social spending, as well as in the scope of the welfare state, can explain the success of the left-wing parties, especially those of the social-democratic orientation of along with strong unions.
In selecting the research method, we began with the assumption of four possible relationships between political orientation, poverty and the welfare state: First, the political orientation affects the poverty reduction regardless of the welfare state program. Greater representation of the left-wing parties in the government reduces the poverty rate before social transfers, by which the importance of redistribution in solving social problems decreases. The empirical confirmation of these settings is in line with the political-economic approach and within it, the Liberal welfare state model. Second, the political orientation has an impact on poverty reduction through welfare state programs. The dominance of the left-wing parties reduces the poverty rate after social transfers, which suggests that political orientation is primarily manifested through policies and mechanisms of the welfare state. Social transfers have the key role in poverty reduction, whose size and structure depends on the distribution of political power between parties of left and right. Such understanding is in line with political-economic approach of the redistributive welfare state model. Third, political orientation and institutionalized welfare state have an impact on poverty reduction. Social spending reduces poverty after social transfers, but the impact of political orientation on poverty does not consist only of the social transfers. Activities of political parties on poverty manifest themselves through social spending and employment. From a theoretical point of view, it is the political-economic explanation of the welfare state, which implies a combination of liberal and redistributive model in politics.
Fourth, in addition to the mentioned three possibilities, a case according to which political orientation has no significant effect on poverty is possible. The regulation of labor markets and /or redistribution in favour of the poor depends on the political importance that society attaches to poverty, as decided by the social groups which are not at risk of poverty and on which the poor who can not have any significant effect. The poor belong to a politically vulnerable the population since they cannot directly represent their interests through the institutions of government. Even if they are members of parties or trade unions, their voice is very weak due to lack of influence on legislative activity, both through their elected representatives and through lobbying. Labour market and redistribution policy, as a party of both the left and right orientation, is primarily driven by the interests of the most numerous and PANOECONOMICUS, 2011, 5, Special Issue, pp. 651-674 most politically relevant social groups. Consequently, the change of power between the right and the left orientation does not significantly change the position at the lower segment of the population income distribution. If this hypothesis should be confirmed it would imply the rejection of the political-economic and the acceptance of structural-functional approach to the welfare state.
From the assumptions above, four hypotheses about the possible impact of the political orientation on poverty are defined.
H1. The influence of political orientation on the level and dynamics of poverty is primarily manifested through employment.
H2. The influence of political orientation on the level and dynamics of poverty is manifested through social spending.
H3. The influence of political orientation on the level and dynamics of poverty is manifested through social spending but also through employment.
H4. Political orientation has no significant effect on the level and dynamics of poverty.
In order for the aforementioned hypotheses to be tested, it is necessary to analyze the impact of political orientation on the two indicators of poverty: (i) poverty before social transfers and (ii) poverty after social transfers. The poverty rate before social transfers reflects a market-generated poverty. This indicator is meant to compare the observed risk of poverty with a hypothetical measure of a risk of poverty in absence of all social transfers (other than pensions) all things being kept equal. The key influence on the level and dynamics of the poverty rate before social transfers is consisted in the activities of the labor market. High employment rate -low marketgenerated poverty rate and vice versa, low employment rate -high market-generated poverty.
The poverty rate after social transfers reflects poverty as a result of the application of redistributive policies. The level and movement rate of poverty is a function of the generosity of the welfare state expressed through the social spending of the welfare-state. High social spending is associated with low poverty rates, while low social consumption leads to high poverty rates corresponding to market-generated poverty.
The study was conducted with the use of a balanced panel model where the unit of analysis is the country-year with a total of 196 observations (N=14, T=14). During the selection of the approach in research of the links between the efficiency of the welfare state and poverty, several previous studies on the variations of poverty and inequality at the macro-comparative level served as a starting point (Arthur S. François Nielsen 1999, 2002;Bjorn Gustafsson and Mats Johansson 1999;Brady 2003Brady , 2005Brady , 2006Brady , 2008Scruggs and James 2006;Brady and Leicht 2008). We analyzed the time period from 1995 to 2008. The data was taken from Klaus Armingeon et al. (2010), Comparative Political Data Set (1960, Institute of Political Science, University of Berne (2010) and the Eurostat database 1 . PANOECONOMICUS, 2011, 5, Special Issue, pp. 651-674 The paper examines the impact of poverty on the political orientation by using country-year as well as country-year grouped according to the regimes of the welfare state level of analysis. The starting point of welfare state typology is Esping-Andersen's classification of capitalist societies or the welfare states. He proposes the division of modern capitalist societies into three regimes: (i) Liberal, (ii) Social Democratic and (iii) Conservative-corporatist welfare state regime. The classification is based on two criteria: (i) decommodification, in the sense of independence of the individual from market outcomes, and (ii) social stratification, which indicates the possibility of movement of individuals between the existing social classes. In the paper, it is applied the extended Esping-Andersen's classification by introducing the Mediterranean welfare state regime with sufficiently differentiated characteristics in relation to the liberal, social democratic and corporatist welfare state regimes.
The main characteristic of the Liberal welfare state regime (UK and Ireland) is market-orientation in the distribution of resources and social protection benefits. Public provisions are modest, flat-rate, and means tested, producing a residual and stigmatised group of beneficiaries. The state encourages market solution by private welfare schemes. The decommodification effect is limited and the distribution of living conditions is closer related to the stratification created by market forces. The Conservative welfare state regime (Belgium, Germany, France, and Austria), rooted in corporatism and Catholic social policy, is identified by status-preserving intervention, supporting the market logic of distribution. Family and motherhood are supported, women encouraged staying at home, and the one-breadwinner family is supported by benefits and taxation. In the Social Democratic welfare state regime (Sweden, Denmark, Finland, and the Netherlands), social policy is encompassing with high standards of material living conditions guaranteed, high levels of transfers as well as subsidised public service; hence also high levels of tax rates. The individual is the favoured unit instead of the family as in the conservative regime. The Mediterenian welfare state regime (Greece, Spain, Portugal, and Italy) is characterise by limited history of full employment; but a recent history of authoritarian politics, where welfare politics is minimal and left to the household subsistence economy and a large informal sector. Poverty and inequality are higher compare with other welfare state regimes (Joachim Vogel 1999).
By using the E-view 6 program, two specifications were evaluated: (i) model with fixed effects, and (ii) model with stochastic effects. The choice between fixed and stochastic specification was made based on the nature of the data, the objective of the research, and the correctness of the decision would later be verified with the application of the Hausman test.
All three indicators pointed to a fixed specification. Firstly, the testing was conducted on a sample of 14 countries belonging to certain modes of the welfare state. The selection of countries was not carried out from a larger set in a stochastic approach suggesting a fixed specification. In the model, the number of countries (14) is equal to the time period (14), which is another argument in favour of fixed specification. Secondly, from the standpoint of economic theory the assumption that the individual effects of the selected set have a decisive influence on the observed variation in the poverty rate can also be accepted. Thirdly, in all models the result of the PANOECONOMICUS, 2011, 5, Special Issue, pp. 651-674 Hausman Chi-square test and p values indicate that the acceptance of the fixed and the rejection of the stochastic specifications.
Model selection is based on the nature of the relationship between the dependent and independent variables. As the most adequate, the semi-logarithmic model was selected, which, when compared to other alternative forms, most closely reflects the nature of the connection between the rate of relative poverty and the given explanatory variables. The semi-logarithmic models describe the relations connected with slower growth of the dependent variable at higher values of the explanatory variables.
In the model, the dependent variable is the rate of relative poverty while the key independent variables are measures for poverty reduction: social spending and employment rate. Such a relationship is best described with the semi-logarithmic models. There are two key explanations. First, in contrast to absolute poverty, relative poverty can not be reduced indefinitely as a result of the implementation of the redistributive and labor market policies. The relative poverty is calculated as the share of people having an equivalised disposable income that is below the at-risk-ofpoverty threshold. In society, there will always be a certain percentage of the population whose income is below the average disposable income. Second, the poverty reduction expenditure per capita increase with reducing the number of the poor. Namely, in the first phase, increasing social spending reduce the poverty rate in the part of the population whose income is closest to the poverty line. At the same time they are also the group with the lowest depth of poverty and consequently the poverty reduction expenditure are lowest per capita. The increase in social spending or employment rates is associated with the reduction of relative poverty by decreasing rate.
In the first iteration, two models were evaluated. In the first model, the dependent variable is the rate before social transfers, and in the second the poverty rate after social transfers. The at-risk-of-poverty rate before social transfers is calculated as the share of people having an equivalised disposable income before social transfers that is below the at-risk-of-poverty threshold calculated after social transfers. The same at-risk-of-poverty threshold is used as for the at-risk-of-poverty rate after social transfers, set at 60% of the national median equivalised disposable income. Independent variables were divided into key and control variables.
Key variables include the impact of the employment rate in the first model, and both employment and social spending, in the second model, the trend rate of poverty before i.e. after social transfers. As a key variable the political orientation indicator is involved. For this purpose, the Schmidt index of political concentration (GOVPARTY) was used, whose values, on a scale of 1 to 5, reflect the ruling political ideology. Value 1 -the hegemony of right-wing parties -0% share of the leftwing parties in power; the value 2 -dominance of the right-wing and centre partiespart of the left-wing parties less than 33.33%; value 3 -uniform power of right and left-wing parties -the left-wing party participation in the government ranging between 33.33% and 66.66%; value 4 -the dominance of social-democratic and other leftist parties -the left-wing party participation in government higher than 66.66% and the value 5 -the hegemony of social-democratic and other parties of the leftwing -100% participation of the parties of the left-wing in the government. The dominance of the right-wing and left-wing parties may have a different impact on poverty rate depending on whether it is observed the poverty rate before or after social transfers. The dominance of right wing political parties has a greater influence on poverty reduction if we analyze the poverty rate before social spending whereas the dominance of left wing political parties has a greater influence on poverty reduction if we analyze the poverty rate after social spending. As a result, if we change the cabinet composition toward the left wing political parties (expressed by increasing the GOVPARTY variable in the model) the poverty rate before social spending might increase but after social spending will decrease.
The control variables in the first model are: the percentage of population over 65 (ELDERLY), the Gini coefficient (GINI), the rate of economic growth (GDPGR), debt (DEBT), the public and mandatory private expenditures for training programs at work, expressed as a percentage of the GDP (TRAINING_PMP), the index of capital account openness (KAOPEN) and the level of union organization (UD). The index for the extent of openness in capital account transactions incorporates a variable indicating the presence of multiple exchange rates, a second variable indicating restrictions on current account transactions and a third variable indicating the requirement of the surrender of export proceeds. The higher the value the more open is the country to cross-border capital transactions. In the second model, control variables include: percentage of population over 65 (ELDERLY), public debt (DEBT) and the rate of economic growth (GDPGR). Control variables reflect economic and demographic determinants of poverty.
The EMPRATIO* variable shows the impact of employment rate on the poverty rate in the welfare state regimes compare to the EU-14 average. The EMPRA-TIO* variable is calculated by taking to the starting model dummy variables for the countries belonging to the particular welfare state regimes and multiplying them with the series EMPRATIO. If the obtained value of the EMPRATIO* variable is statistically significant, the impact of employment on poverty rate, in the analysed welfare state regime, is different from the EU-14 average. The gap size is calculated by adding the value of variables EMPRATIO and EMPRATIO* for the analysed welfare state regimes. Depending on the sign of EMPRATIO* variables, there are two possible forms of deviation: (i) upwards deviation (negative sign) and (ii) downwards deviation (the positive sign). The upwards deviation indicates that the impact of employment on poverty reduction, in the observed model of the welfare state regime, is above the EU-14 average, and vice versa, the downwards deviation -the impact of employment on poverty reduction is below the EU-14 average.
From economic perspective, the EMPRATIO* variable is a rough approximation of the effects of structural factors, typical for certain welfare state regimes, on poverty rate. Namely, unemployment is a key cause of poverty in all countries and welfare state regimes in the EU. It follows that the most effective way to reduce poverty is to increase employment. However, the impact of employment on poverty reduction is not the same in all countries and welfare state regimes, which is explained by the influence of structural factors on poverty, labor market and welfare state. The structural factors include a number of factors which can not be controlled by the market and welfare state mechanisms for poverty reduction in the short term, such as: the culture of poverty, institutional features of labor markets and welfare states. As the result of structural factors, the determinants of poverty differ between the countries and the welfare state regimes.

Results
By using the explained methodological framework, whether and how political orientation affects poverty in the EU-14 countries, classified by welfare state regimes was tested. For each independent variable, the value of the regression coefficient with associated probability p was shown. The level of significance of evaluation is moving within the range from 1 to 10% with the use of the double-sided test. Because of the potential problem of multicollinearity, the same indicator of political orientation is applied in every model -the Schmidt index of political concentration.
When it comes to the impact of the degree of union organization on the poverty rate, it is a positive relation. When other factors are constant, the increase in the degree of union organization increases the rate of market-generated poverty. In economic theory, a great number of explanations for the trade-off between the activities of trade unions and poverty are accessible. One of the most famous is the insideroutsider theory. Union members are the insiders and the unemployed are the outsiders. Unions focus the interests of its members in terms of requirements for stability of earnings and position. Consequently, rigid structures of the labor market which reduce the possibility of obtaining work for the unemployed are created. Because the unemployed at risk of poverty are the most vulnerable social group, a greater degree of union organization leads to higher unemployment and poverty. C 3,695 0,000*** 3,929 0,000*** 3,378 0,000*** 3,843 0,000*** 3,715 0,000*** SOCEXP_T_PMP -0,023 0,000*** -0,025 0,000*** -0,015 0,033** -0,023 0,000*** -0,021 0,001*** EMPRATIO -0,009 0,017** -0,013 0,009*** -0,010 0,010** -0,009 0,012** -0,007 0,073* DEBT -0,005 0,000*** -0,005 0,000*** -0,004 0,000*** -0,005 0,000*** -0,005 0,000*** ELDERLY 0,029 0,011** 0,032 0,006*** 0,024 0,040** 0,040 0,001*** 0,023 0,057* However, the link between trade unions and poverty is not significant. In practice of the European countries significant variation in how the organization and goals of the labor union are present. Poverty reduction is not an explicit goal and the influence of unions on the degree of labor market rigidities is smaller than is commonly thought. Thus, in countries of the Social Democratic welfare regimes the highest degree of trade union organization is present, 63 to 41 which is the average of the EU-14. At the same time, the rate of market-generated poverty in these countries is almost identical to the average poverty rate in the EU-14, 27% to 26%. This is explained by the use of special concepts of social policy -the flexsecurity which combines a flexible labor market, social security and active labor market policy. As the name suggests it is a combination flexibility and security. Consequently, a high level of union organization in social-democratic countries is not followed by a rigid labor market and on this basis significant poverty of the unemployed.
In the second model for all countries, economic growth rate has the expected negative impact on the poverty rate before social transfers, but this effect is not statistically significant. Since within the first model of economic growth was a significant factor, the explanation for the insignificance in the second model should be sought in the mutual relationship between the economic growth and social transfers in combating poverty. Overall, GDP growth may act toward strengthening, but also toward the weakening the influence of social spending on poverty. If the society becomes richer due to dynamic economic growth, then the state, through a system of taxes and contributions, collects more funds earmarked for social spending. Higher social spending is the most important mechanism for reducing poverty. The impact of economic growth on poverty is expected in terms of sign, but is not significant because it is manifested to a greater extent through social spending rather than through increasing employment.

Political Orientation
Political orientation in terms of greater participation of the left-wing party in power has the expected positive sign but that effect is not significant. In the analysis of market-generated poverty, it was presumed that employment is a key mechanism for poverty reduction. High employment rates are associated with the policy of promoting economic growth which is closer to right than to left-wing parties. Generally, the right-wing parties prefer the supply side economic policy whose major components include tax reduction, deregulation of labor markets, reduce public spending, dismantling the welfare state, weakening the legal rights of unions, including the cooperation with big business in shaping the economic and political environment. All these measures are aimed towards the deepening of economic inequality and poverty. However, the negative impact of poverty can be cancelled by the positive influence on economic growth and employment. The results show that in the case of marketgenerated poverty, for the given group of countries in the observed period, the first PANOECONOMICUS, 2011, 5, Special Issue, pp. 651-674 hypothesis on the primary impact of political orientations to poverty through labor market institutions is rejected.
By further analysis of market-generated poverty rates, employment rates and political orientation has shown that countries with social-democratic orientation did not achieve significantly different results in the alleviation of poverty in the countries where the authorities belonged to the centre and right-wing parties. If we take as a criterion the above average employment rate (65%) and below the average poverty rate (25%) in 9 observations countries in which the authorities belonged to the rightwing parties showed better result and in 8 observations the countries where the government was constituted out of left-wing parties.
Such results can be interpreted in two different ways. First, it takes a long period of time to understand the real effects of a political option on social welfare. Poverty reduction as a result of stimulating economic growth and employment presumes the authority belonging to the right-wing parties in multiple terms in order to gain the power of effective control over the work of government. This power is accumulated over time because it is unlikely that institutionalized history of egalitarianism and free market may turn in a different direction immediately after the elections (Esping-Andersen 1990).
Secondly, it is possible that the effects of the measures of previous governments, oriented either right or left, came to the forefront after the government changes, having in mind that each newly elected government inherits institutions, legislation, administration, as well as all activities of the previous government. Many of the programs and reforms, including changes in the labor market take several years and their effects occur at the time that does not have to coincide with the rule of the political party that initiated the change. Therefore, caution is needed in interpreting the results in the field of social policy in countries where the government often changes.

Employment and Active Labor Market Policy
Employment rate and expenditure for active labor market measures have the expected negative sign and are significant determinants of poverty. The increases of the employment rates and higher expenditures for active measures of employment, alongside all other constant factors, have a strong impact on poverty reduction. Such findings favour the thesis that solving social problems through employment is the key determinant of market-generated poverty.
The impact of employment on poverty is not the same in all regimes of the welfare state. Compared to the average EU-14, the above-average employment rate of significant impact on poverty reduction is present in the social-democratic group and below average in the Mediterranean group. In the remaining two modes, the corporatist and liberal, the impact of employment on poverty corresponds to the average of the group. Social democratic countries have the best results in terms of employment. In the period from 1995 till 2008, the average employment rate in the socialdemocratic group was 10 percentage points above the average of the EU-14. At the same time, the Mediterranean countries had almost identical results only with the opposite sign. The stronger impact of employment in poverty reduction in social-democratic countries should be interpreted, first of all, through a high rate of female employment, higher expenditures for active labor market policies, including a high share of public services in the GDP. In all these areas, the social-democratic countries have the highest values in the EU-14. Thus, the employment rate of women in social democratic countries is 10 percentage points higher than in the EU-14, expenditures for active labor market measures are 60% higher than in the EU-14, the share of public services in GDP is 40% higher than the average EU-14. The data on the movement of the same indicators in the Mediterranean countries, where the deviations from the average of the EU-14 is almost identical to the social-democratic countries, only with a negative sign, speak in favour of the correctness of this approach.

Economic Inequalities
Economic inequality and poverty are different and the same time and related concepts. Poverty is related to the standard of living of the population whose income is below the poverty line, while economic inequality is related to the standard of living of the entire population. Hence, it is possible to reduce or increase economic inequality and for the poverty to remain unchanged. Such a situation is apparent when the income is allocated in groups with no risk of poverty, while the position of the poor remains unchanged. At the same time, economic inequalities and poverty are related categories, as increased poverty leads to rising inequality. As a result, social policy usually states as a goal both the reduction of poverty and economic inequality.
Panel regression results show that economic inequality increase marketgenerated poverty. The increase in the Gini coefficient, alongside other constant factors, causes the growth of the poverty rate. Referring to the previously described relationship between poverty and economic inequality, we can conclude that the benefits of income growth and its market distribution are enjoyed to a greater extent by groups which are not at risk of poverty. In most EU-14 countries, in the period from 1995 to 2008, a decline in economic inequality was recorded. If we bring the data on the movement of economic inequalities in connection with the movement of the poverty rate before social transfers, in most countries changes in inequality are more prominent than changes in poverty. With this in mind, it can be concluded that not all social groups have the same benefits of economic growth and employment growth. In general, the absolute benefit of the non-poor population is higher when compared to the poor population. The explanation should be sought in the low level of human capital at their disposal which makes it difficult for the poor to access the labor market, even when under the influence of economic growth, new positions are created.
With regard to the political orientation, economic inequality and marketgenerated poverty, the data corresponds to the expectations. If the parties of left and right-wing are ranked in relation to the below average poverty rate before social transfers and the below average value of the Gini coefficient, the number of observations with below average values for both indicators are twice higher where the parties in power belonged to the right-wing orientation rather than the left-wing orientation (14 vs. 6). This result is well expected since the impact on poverty before redistributing is being analyzed. Policies that encourage economic growth and employment influence to a greater extent the economic inequality and poverty before social transfers than policies that advocate more prominent redistribution of income in society.

Population Aging
Panel regression results show that when other factors are constant, the increase in the participation of the elderly in total population results in the growth of marketgenerated poverty. Such a result was expected bearing in mind that the older population is in the group with above average risk of poverty. If we limit our analysis to the poverty rate before social transfers, there are two main explanations for the increased risk of poverty of the old population.
First, the income of the old population is lower than the income of workingage population, which increases the risk of poverty. Second, the ability to compensate the missing additional income through involvement in the labor market is very limited with the old population. Hence, the market mechanism provides very little room for solving the problem of poverty of the old population. If we exclude social transfers from the analysis, the main available channel is help within the family based on the generational solidarity. Since the problem of an aging population and poverty in all countries has fiscal character, we devote more space to this issue in part of this article which comments on the results in terms of poverty rate after social transfers.

Public Debt
Public debt has a dual role in determining the level and dynamics of poverty. On one hand, the public debt affects the quality of business environment and investor expectations. The high level of public debt is usually associated with the fall in investments because investors fear the growth of taxes as the ultimate means to balance public finances. This problem is particularly present in terms of economic integration, when barriers for free movement of capital are removed. In the final instance, lower tax rate can lead to the relocation of capacities to other countries and a rise in unemployment and poverty. Manoeuvring capabilities of the government to reduce public debt without raising taxes are limited due to lower credit solvency of the country and the resistance of unions and other interest groups to the limitation of public spending. The final result is the growth of poverty caused by the reduction in investments and employment.
On the other hand, debt can have a negative impact on poverty levels. The largest share of public expenditure, which causes the growth of public debt, is the social spending, i.e. the measures with the ultimate goal of preventing and resolving social problems. Hence, a large public debt may be due to the efforts that the government seems to undertake with non-market solutions to combat poverty. In addition to social spending, public debt in itself may contain components that directly or indirectly -through the market, affect the level of poverty. In terms of market mechanisms, we refer to before all subsidies and other form of fiscal incentives for the economy and population. Employment in the public sector should also be added, PANOECONOMICUS, 2011, 5, Special Issue, pp. 651-674 which falls into the realm of public spending that causes the growth of public debt due to the generally lower productivity of the public in comparison to the private sector. In both cases, the growth of public debt is associated with falling poverty rates.
Panel regression results indicate a significant negative relationship between public debt and the rate of market-generated poverty. The growth of public debt, when other factors are constant, leads to a decrease in market-generated poverty. The explanation for this situation should be sought in greater negative impact of public debt through fiscal incentives for the economy and population on the poverty rate than the positive impact, through the fall in investments. This is corroborated by a small degree of fiscal harmonization in the EU. In addition to restrictions on the functioning of the euro zone, EU member states have almost full fiscal freedom, which is manifested through the adjustment of public finances and economic interests of the population taking into account the investment and employment. With the exception of few countries (mainly Greece), public debt has not led to significant decline of credit solvency of the country or in investments.

Political Orientation
At first, it is necessary to note that political orientation, while keeping in mind the poverty rate after social transfers has a negative and, unlike the market-generated poverty, significant impact on poverty trends in the EU-14. The strengthening of the left-wing parties, for the given group of countries in the period from 1995 to 2008, was accompanied by a reduction in poverty rate.
In the context of the set hypotheses, we are talking about the acceptance of hypothesis 3 according to which the impact of political orientation on the level and dynamics of poverty is manifested through the institutions of the welfare state but also independently from the welfare state. In other words, it is the combined impact where institutions and mechanisms and the welfare state and labor market mechanisms reduce poverty with the support of the left-wing parties.
Social spending, as an indicator of the generosity of the welfare state, when other factors are constant, has the strongest effect on poverty reduction. However, it is not about channelling effects in the sense that the impact of political orientation is solely manifested through social spending. At the same time, the variable rate of employment is also significant, which indicates that the effect of political orientation on poverty is manifested through the labor market.
In this regard, it is logical to ask why in the case of market-generated poverty political orientation is not significant, while in poverty after social transfers it is a significant factor in the level and dynamics of poverty. There are two key explanations. First, the effects of changes in government, in relation to right -left, on welfare, are manifested fastest through changes in social spending. If the parties of the left-wing orientation want voters to show their commitment to solving social problems, the fastest reaction is the change in level of social transfers. Change course in social policy through social spending in many areas is only the change in the level PANOECONOMICUS, 2011, 5, Special Issue, pp. 651-674 and requires no change i.e. institutional reform. Moreover, obtaining political consensus to change the abundance of social programs without interfering in their functioning is less uncertain in terms of coalition governments in regards to the measures that are easy to alter with the change of government.
Second, changes in levels of social spending are a politically less controversial mechanism in comparison to the reforms of institutions, of the welfare state and labor markets as well. The reform of pension and health systems, or the (de)regulation of the labor market is contained in the agenda of political parties but they are still politically very risky projects. Any reform requires a waiver in the present to achieve greater benefits in the future. From the perspective of the ruling party, the reforms are a less attractive instrument because of a possible change of government, because of the price of reforms that voters are paying, so that the final political gain can benefit a government that did not even participate in initiating and implementing change.
The results of econometric analysis have showed that the impact of political orientation on the poverty rate after social transfers in the EU-14 is expected, in terms of the sign, and statistically significant. However, in the case of welfare state models, this impact is not statistically significant for the Social Democratic and the Corporatist Welfare state regime. This deference is primarily explained by structural factors that diminish the impact of changing political orientation on poverty.
The EMPRATIO* analysis shows that the influence of structural factors in the the Social Democratic and the Corporatist welfare state regime is higher than the EU-14 average. The Social Democratic and the Corporatist Welfare state regime have more developed institutions of the welfare state and labor market compared with the Liberal and the Mediterranean Welfare state regime. In addition, they have more stable economies and a lower average poverty rate. In such circumstances, changes in political orientation is less expressed through poverty reduction compared with economies that are characterized by relatively unstable institutions of the welfare state and labor market, economic situation and higher poverty rates.

Social Spending
Panel regression results showed that, when other factors are constant, social spending is a key determinant of poverty rate after social transfers. This result was expected since the impact of social spending on poverty reduction is being considered. In all countries, and models of the welfare state there is a very noticeable negative relationship between social spending and poverty. It is not necessary to outline the way in which social spending reduces poverty. Social spending is the main mechanism by which redistribution of the welfare state affects poverty and economic inequality. The analysis of the efficiency of social spending to combat poverty is much more interesting is a much more interesting issue.
The countries of EU-14 are characterized by relative uniformity of social spending and the relative disparity in poverty rates. On one hand, the average social spending in the EU-14, in 2008 amounted to about 27% of GDP with little variation between countries. The ratio of social expenditure between countries with the most abundant social spending, Denmark, and the country with the least generous social spending, Ireland, is less than 1.5. On the other hand, the average rate of relative PANOECONOMICUS, 2011, 5, Special Issue, pp. 651-674 poverty after social transfers amounted to about 15% but with significant variations between countries. Ratio of the rates of relative poverty of the country with the highest poverty rate (Greece) and lowest poverty rate (the Netherlands) is over 2. Different levels and dynamics of social spending and poverty, point to differences in the effectiveness of the redistribution system, as observed in countries and welfare state regimes.
Study of the efficiency of social spending to combat poverty mostly show high efficiency of the Social Democratic and Liberal, and low efficiency of the Mediterranean and Corporatist model of the welfare state (José Adelantado and Eduardo Cuevas Calderón 2006;Luis Moreno 2006;Andre Sapir 2006;Jingjing Huo, Moira Nelson, and John D. Stephens 2008;Kosta Josifidis, Novica Supić, and Emilija Beker Pucar 2010). If the efficiency is an assumption of the sustainability of social spending, it is realistic to expect a reform of the inefficient, the Mediterranean and Corporatist regimes of the welfare state and in the direction of two possible alternatives -the Anglo-Saxon deregulation and Scandinavian flexicurity. The current crisis of the neoliberal concept leaves little room for expectations that Europe, especially European countries with traditions and developed national consciousness, will begin to move toward greater deregulation. Scandinavian flexicurity is a more likely option. It is also important to note that in all modes the trend of rising social expenditures is more noticeable, which is, moreover, most prominent in the least effective, the Mediterranean group. Thereby, the solutions for the increase in the efficiency should be sought primarily in the structural rather than in fiscal characteristics by which the welfare state regimes are differentiated. The influence of the political orientation on the rate and direction of change is of great, if not decisive, importance.
As for the relation between political orientation and social spending, by reviewing mutual relations it is evident that there was a higher level of social spending in countries where the authorities belonged to the parties of the left-wing rather than in countries where authorities belonged to the right-wing parties. In the period from 1995 to 2008, a range of social spending in countries where the dominant parties belonged to the left-wing ranged from 32.5% of GDP (Sweden in 1995) to 17.34% of GDP (Greece 1995). At the same time, the range in countries where the authorities belonged to the right-wing parties went from 29.5% of GDP (France 2005) to 13% of GDP (Ireland 1998). Higher social spending in the first group is attributed to the greater commitment of left-wing parties to solve social problems by means of redistribution.

Employment
Panel regression results showed that when other factors are constant, the increase of employment rates reduces the poverty rate after social transfers. This result was expected because along with the social spending, the growth in employment is a key mechanism for fighting poverty.
In the first model, political orientation, in terms of the domination of the rightwing party, was not a significant factor in curbing market-generated poverty, but the sign was in line with expectations. The positive influence of right-wing parties on poverty is interpreted by the means of supply side economic policies that are more PANOECONOMICUS, 2011, 5, Special Issue, pp. 651-674 useful to the capital rather than the workers. The absence significant effects were explained by spill over effects and short term analysis horizon.
In the second model, which considers poverty rate after social transfers, the dominance of the left-wing parties is a significant factor in reducing poverty. At the same time, this effect is also shown by social spending, as an indicator of the welfare state, and the employment rate as an indicator of the labor market. This result suggests that the left-wing parties do not only manifest their impact on poverty solely through social spending but also through the labor market in terms of institutionalized social policy.
In connection with the impact of the employment rate on poverty after social transfers, it is interesting to consider the extent to which social spending and employment are mutually channelled or that is not mutually channelled mechanisms in combating poverty. In the first case, it is a synchronous impact while in the latter case, the impact is ambivalent. In a theoretical sense, we talk about complementarity vs. exclusivity. Exclusivity refers to situations where generous social spending reduced incentives to work, in terms of labor supply, and reduces the number of new jobs created, in terms of demand for labor. Of course, in reality, the relationship between social spending and employment is regulated by more flexible instruments to reduce this sort of extreme exclusivity to a minimum. Thus, in most countries, the use of unemployment aid is conditioned by active measures to find jobs. Hence, it is more correct to speak of channelling vs. non-channelling of social spending and employment than about complementarity vs. exclusion.
In order to analyze (non)-channelling, a decomposition of the employment rate according to the welfare state models was conducted. In short, the impact of employment on the poverty rate after social transfers is below average in socialdemocratic countries and above the average for countries of the corporatist welfare state model. Below average effect implies channelled and above average a nonchannelled employment impact on poverty. The observed differences in the manifestation of social spending and employment can be explained by different approaches in the organization of social protection systems and labor market regulation.
Social Democratic countries, as previously mentioned, have, over time developed a special approach to social policy -flexicurity which involves a combination of a flexible labor market, social security and active labor market policy. Labor market flexibility is achieved by a high degree of freedom of employers in hiring and firing workers. Social security is provided by generous unemployment benefits. It is a combination of efficiency and fairness. However, the operationalization of the idea is not feasible without the use of active labor market policies which, on one hand, allows retraining of workers to the needs of the market and on the other hand, prevents the abuse of the unemployed status. The link between employment and social spending is indirect, through active labor market measures. The link between employment and social spending is indirect, through active labor market measures. The key source for financing the training and retraining is the social spending. As a result, we confirm the thesis of channelling effects as the effects of employment growth on poverty is comprised in social spending. In addition, high employment rate, especially of groups with above average poverty risk, in the public sector should also be PANOECONOMICUS, 2011, 5, Special Issue, pp. 651-674 added. Employment in the public sector is the most obvious example of channelled effects when employment affects the poverty rate through social spending.
In order to understand the low channelling of social spending and the employment of the corporatist model, it is necessary to point out the specifics of the organization of social protection systems in the corporatist model in comparison to the social democratic welfare state model. The social-democratic model, in its essence, represents a combination of universalism with the generosity of the social program on one, and the comprehensive socialization of social risk on the other. In this way, social spending largely comprises and channels the effects of labor market on poverty. What is the situation with the corporatist model?
The basis of the corporatist regimes of the welfare state is the Bismarckian tradition according to which the work merit, achievements and productivity, are a prerequisite for meeting the needs through the social security system. The concept of mandatory social insurance is applied where working contributions are the conditions for using the services of social protection systems. People without a work history can participate in the residual social assistance programs, but their role is marginal in comparison with the obligatory social insurance. Because of strict conditionality of employment and participation in programs of social protection, social spending has to a lesser extent the effects of labor markets in poverty reduction. Social problems are primarily addressed through the recruitment and use of social assistance on the basis of previous employment.

Public Debt
Panel regression results show that when other factors are constant, public debt has a significant negative impact on the poverty rate after social transfers. This result was expected given the ratio of public debt and social spending.
In the structure of public expenditure, the most abundant component is the social consumption. Primarily, public debt is due to the cumulative faster growth of public expenditure over public revenues. Thus, the increase in public debt, in most cases, is based precisely on the excessive social spending. Since social spending is the key mechanism of reducing poverty, a negative relationship between the public debt and poverty assumed, as evidenced in the survey.
From the perspective of political orientation, it is interesting to analyze how different political orientations influence the level of public debt. On the basis of the relationship between the political orientation and public debt, for the EU-14 countries during the period from 1995 to 2008, public debt over 60% of GDP in 47 observations (year and country) was present in countries where the dominant parties belonged to the left-wing orientation (political concentration index greater than 3) and 45 cases in which the countries dominant parties belonged to the left-wing orientation (political concentration index is less than 3). This even score can be explained by the fact that a change of government does not automatically mean a reduction in public debt, since the new government inherits the previous state of public finances that can not be repaired within the time limits which coincide with election cycles. Therefore, the impact of political orientation on the budget deficit is a better indicator. However, the data is uniform. If we take a budget deficit of 3% of GDP as a cri-PANOECONOMICUS, 2011, 5, Special Issue, pp. 651-674 terion, in three observations, the budget deficit higher than 3% of GDP was in the country in which the parties in power belonged to the right-wing orientation and in 4 observations where the government was formed by right-wing parties. Therefore, the analysis of the impact of political orientation on the level of public debt and budget deficit requires long term series analysis and the inclusion of a broader institutional context.

Population Ageing
The results of the panel regression show that with other factors being constant, the increase of the share of elderly people in the total population after social transfers causes poverty growth. The obtained result was expected since ageing population represents a group which is exposed to the risk of poverty beyond average. Therefore, the increase of the ageing population increases the number of people dependent on the social welfare. With constant social spending, the same amount of transfers is directed to the larger number of users, which deepens social problems.
European Commission report from 2007 shows what kind of challenge is put forward for the European welfare state model. The report states that one quarter of the EU-15 adult population is older than 65. If current ageing trend continues, by 2050 the share of ageing people in the total EU population will be 50%. The above mentioned trends seriously endanger the existing social models in Europe, especially in the field of regulating and financing social and health insurance.
In order to obtain a better insight into the ageing and poverty problems in Europe, it is necessary to simultaneously observe the relation between the ageing population -dependant population and young one -working population. Dependency rate indicator of the elderly population is used for this purpose. In the last few decades, dependency growth rate of the elderly population has immensely started to endanger the fiscal situation in most European countries. According to the European Commission report from 2007, it is estimated that the dependency growth rate of the elderly EU-15 population will have resulted in the increase of the social security expenditure between 5% and 7% GDP by 2020. With such trend, by 2040, health security and pension expenditure will have doubled or even tripled in most of the EU-15 countries.
However, such predictions should be cautiously taken into account having in mind a high level of uncertainty in the working population projections. While monitoring the movement of the elderly population, it is relatively possible to determine precisely the dynamics of its reaching the retirement age, which makes these projections relatively reliable, together with the previously obtained mortality rate. With the working population, the situation is significantly more complex since it is quite difficult to predict with certainty if the birth rate will grow, fall or stagnate in the future. Migration flows also have a significant role in determining number of the members of the working population. In the circumstances of large influx of emigrants, it is possible to keep status quo at the employment market despite the decrease in the fertility rate. EU-15 countries might be the best example of this because the total amount of population, not only working one, would be by far in decline if there was not for the migrations in them. PANOECONOMICUS, 2011, 5, Special Issue, pp. 651-674 When the relation between political orientation, ageing process of the population and poverty is concerned, this influence is indirect and primarily manifested through changes in regulations regarding the exercise of rights to have pensions and health security. On top of this, there are also opinions on the migrations and encouragement of programmes aiming to boost human capital of young population. Since changes in this field suggest reforms, whose results do not coincide with political cycles, it is not possible to notice in the given example more evident deviations between the left and right wing parties. The same goes for observing the particular welfare state models.

Conclusion
This paper explores the impact of political orientation on poverty in the European welfare state models. Starting from a given theoretical framework: politicaleconomic vs. structural or liberal vs. redistributive welfare state approach, four hypotheses about the possible impact of political orientation on poverty have been defined: (i) the impact of political orientation on the level and dynamics of poverty is primarily manifested through employment; (ii) the impact of political orientation on the level and dynamics of poverty is primarily manifested through social spending; (iii) the impact of political orientation on the level and dynamics of poverty manifests itself through social spending but also through employment; and (iv) political orientation has no significant effect on the level and dynamics of poverty. Hypothesis testing was performed using a balanced panel models, the EU-15 countries without Luxembourg, in the period between 1995 and 2008.
Two models have been evaluated: the model with market-generated poverty rate and model with welfare state mediated poverty. In the first model, the influence of employment and political orientation, as key variables, on the poverty rate before social transfers have been analysed, using the control variables: the share of ageing population in the total amount of the population, economic inequality, economic growth rate, public debt, active labor market measures, the degree of union organization and openness of capital transactions. In the second model, the analysis also include social spending as a key variable while the control variables were reduced to the share of ageing population in the total amount of the population , economic growth rate and public debt.
With market-generated poverty, the results showed that greater participation of the right wing parties in running the country lowers the poverty rate before social transfers, but that this influence is not significant. With other factors being constant, a key mechanism in reducing the poverty market is high employment rate. A more detailed analysis of market-generated poverty rates, employment rates and political orientation has shown that countries with social democratic orientation parties in power did not achieve significantly different results in the alleviation of poverty in comparison to the countries where right wing parties were in power. This is explained by the necessity for an analysis of a longer period in order to understand fully the real effects of a political option on the social welfare. Also, it is possible that the measures of the previous governments become more conspicuous with the change of PANOECONOMICUS, 2011, 5, Special Issue, pp. 651-674 government, since each newly elected government inherits the institutions as well as the activities commenced by the previous government.
For welfare state mediated poverty, the opposite results were obtained. Political orientation has both a negative and significant impact on the poverty rate after social transfers. Such findings are consistent with the hypothesis on the combined impact of political orientation on poverty through social spending and employment. With regard to this, this paper promotes the thesis according to which the observed differences in the impact of political orientation on market-determined poverty and welfare state mediated poverty are primary the consequence of the determination of political parties to manifest their influence in the area of social policy primarily through the use of short-term measures, such as the change of levels in the social spending, rather than through the use of long-term measures, such as institutional reforms of the welfare state and labor market. Such approach not only provides faster results but it is politically less risky since the reforms take a long time and the final political gain can be obtained by a government that did not participate in the initiated implementation of changes.