Abstract

This paper is a first attempt to answer the question what are the predictors for party support for the Social Democratic Union of Macedonia (SDSM) on local level. While there are general theoretical guidelines and similar research from other countries, this issue has not been tackled in Macedonia. Using multiple regression analysis, I test the influence of five sets of factors (geography, demography, socioeconomic conditions, cultural factors and electoral opportunities) on the vote share of SDSM in three local electoral cycles (2005, 2009 and 2013). The results show that SDSM's support is positively related to municipalities with a greater number of households, a higher average age of citizens, a higher number of ethnic Macedonians and a higher number of economically active citizens; where electoral opportunities are higher. Education levels show mixed results, while a higher proportion of religious citizens shows a negative relationship with support for SDSM. Whether municipalities are urban or rural, rich or poor, close to the capital or in the capital itself, have high levels of social welfare spending and greater representation of  women, are all factors that are not related to support for SDSM.

 

Key words: party support, local elections, Social Democratic Union of Macedonia, left political base

 

Introduction

In this paper I try to answer the question what predicts party support for the Social Democratic Union of Macedonia (SDSM) on local level? While there are some general theoretical guidelines to build expectations for party support and some specific studies for party support and voter behaviour at local level, the research so far undertaken about political parties in Macedonia hasn't tackled this question.
From the research work done so far, in general and in Macedonia, one can identify five factors that influence support for a political party: geography, demography, socioeconomic conditions, cultural factors and electoral opportunity structures. To measure the independent variables and to build the dataset I used several sources: data from the census conducted in 2002, a study on social inclusion at local level, and data from the State Electoral Commission. I set the level of analysis at the local level (i.e. sub-national level). The units of analysis are the municipalities. Vote share of SDSM in three cycles of local elections (2005, 2009 and 2013) is the dependent variable.
To test the influence of the independent variables, I use multiple regression analysis. I build five different models and run tests for each cycle of local elections. The results confirm some expectations. In some instances, SDSM support comes from the expected social democratic base. However, other results call for rethinking our understanding of what constitutes the left political base in Macedonia. This has significant implications for academic research and for political party strategists.
Some might be tempted to point out one weakness of this approach. I test support only for one party and do not make comparisons with other parties. In all fairness, no research work is immune from personal biases. My bias lies with the politics of the left. On the other hand, I compare and analyse support for SDSM over time. Comparisons with other parties would be useful for gaining a better insight into the dynamics of the party system but would not serve to better answer the research question of this paper.
The paper proceeds as follows: I first briefly summarize the literature for analysing party support. I then elaborate my research design, to explain my method of analysis and to build the expectations. The results are then discussed, followed by a conclusion which summarizes the findings.

Literature for analysing party support?

Literature on political cleavages is usually used as theoretical guidance for analysing political party support. This literature has two main strands of thought. The first is that party support is based on social characteristics, while the second is that party support is based on attitudinal values. Recent research in post-communist countries shows that political differences with regard to the communist past also play a role in party support. At the risk of omitting many significant elements, I will shortly summarize this literature, pointing to the main tenets which I will use to build my approach in this paper.

 

The relevance of social characteristics for party support

The study of Seymour M. Lippset and Stein Rokkan (1967) convincingly showed that the national and industrial revolutions in Europe created four different types of political cleavages: centre – periphery; urban – rural; church – state; and owners – workers. Political parties were seen as superstructures that represented these cleavage structures (Bartolini and Mair 1990). Belonging to a class or religious group “structured the content of political and electoral discourse in Western democracies, and thereby partisan alignments and voting choices” (Dalton 1996: 319). The main tenet of this approach is that modernization and socio-cultural factors are most important in the creation of the party system (Birnir 2007; Neto and Cox 1997; Ordeshook and Shvetova 1994).
This approach puts forward the hypothesis that the party system and party support in Western Europe became “frozen” due to social characteristics. Earlier research corroborated these findings in the new Central and East European democracies (Markowski 1997) such as the Czech Republic (Mateju et al. 1999), Hungary (Evans and Whitefield 1995) and Poland (Shabad and Slomczynski 1999). The work of McAllister and White (2007) shows that even though parties are less effective in representing social cleavages in emerging democracies, religion and class are predominant social cleavages in both established and emerging democracies.
However, recent research has shown that social characteristics are not directly associated with party support in Central and Eastern Europe (Casal Bértoa 2012) and that, over time, social structures tend to matter less for party support in Western Europe (Franklin et al, 1992; Dalton, 1996), for example when the standard of living improves (Norris 2004). This has led to a thesis about the “decline of social structures” emanating from sectoral change, mediatization, individualization and secularization among other things (Enyedi 2008). The usage of social cleavage politics as a concept in new democracies has also been questioned (Mainwaring 1999; Mainwaring and Zoco 2007), not least because of the role of political parties as agents that structure the political divisions (Eneydi 2005) along with the impact of critical junctures (Bornschier 2009), the absence of strong social differences and the strong influence of ethnic identity on party support in post-communist countries (Moser 2005; Barany 2005).

 

The relevance of attitudinal values for party support

Attitudinal differences, also referred to as the opposition between authoritarian and libertarian values (Kitschelt 1994, 1995b), are not necessarily embedded in social structures. They are based on matters of lifestyle and moral guidelines that voters share with a  political party. Cultural components thus influence party support (Knutsen and Scarbrough 1995; Bartolini 2000). The main tenet of this approach is that values and the “cultural left-right” determines party support. Post-material attitudes (e.g support for environmentalism, euthanasia, LGBT) are juxtaposed to nationalistic and moral conservative values. Libertarian values are placed in the left camp, while the politics of the right are associated with authoritarian values (Enyedi 2008).
Some research showed that party support across Central and Eastern Europe can be explained by value-based voting (Toka 1998; Kitschelt et al. 1999), which has been corroborated by recent findings in Hungary (Evans 2006) and Poland (Tworzecki 2003). However, notwithstanding the relevance of attitudinal values, it seems that instability and unpredictability are the main characteristic of party politics in Eastern Europe (Casal Bértoa 2013).

 

The relevance of differences over the communist past

Herbert Kitschelt (1995a) hypothesized that in post-communist democracies the nature of the regime, along with the pathways to democracy and individual experiences, would shape political divisions. The work of Joshua Tucker and Grigorie Pop-Eleches (in progress) claims that communist legacies are important for understanding political behaviour in post-communist countries. Research from Hungary (Körösényi 1999) shows the relevance of anti-communist conflict for political divisions described as “members and descendants of the communist ruling class versus the principal losers of the communist takeover” (Enyedi 2005: 298). In Poland, party support was dominated by the rivalry between the post-Solidarity and post-communist camps up to 2006 (Jasiewics 2009).
Research from Croatia shows that family lineage in World War II is a crucial variable for an individual's choice of party support (Šiber 2001). This division coincides with a centre-periphery cleavage (Zakošek 2002). And even though this influence is weakening (Zakošek 2000), authoritarian preferences are still stronger among right supporters (Čular 2005). The communist and anti-communist cleavage appeared as a strong line of political division in Slovenia in the early 1990s (Fink Hafner and Krašovec 2006). And it seems that perceptions over the communist past are highly relevant for political divisions in Macedonia (Panov and Taleski unpublished).

 

Analysing SDSM's support in Macedonia: research design and methods

After political pluralism was introduced, authors wrote about the social and political profile of parties in Macedonia in the early 1990s (Mojanoski 1996). Others wrote about party system developments in relation to various electoral cycles (Škarić 2005), about the ideological profiles of the parties (Siljanovska-Davkova 2006) and minority parties (Taleski 2008). Hossein Shahla (2005) showed the overriding importance of ethnic identity for party support in Macedonia, while Lidija Petkovska-Hristova (2010) showed that expectations about party support that are valid in other post-communist countries are only partially found in Macedonia.
Notwithstanding the previous research undertaken in Macedonia, to the best of my knowledge this is the first attempt to test voter support for a single party using results from three cycles of local elections as units of analysis. However, there is similar research from other countries. For example, research from the United States shows that geography and demography are potential variables in explaining voter behaviour at municipal level (Klos 2008). Shifts in party support can be explained by determining geo-statistical correlations between demographic variables and voting behaviour. Other authors from the US find that ethnic identity is an important variable for voter turnout in local elections (Barretto et al. 2005). On the other hand, research from Norway shows that institutional and party system variables (e.g., district magnitude and voter turnout) have higher explanatory power for party support than socio-economic conditions in local elections (Kestilä and Söderlund 2007).
When results from local elections are the units of analysis, one should bear in mind that institutional challenges may influence voter turnout (e.g. reforms in public services or local elections coinciding with national elections) (Hajnal and Lewis, 2003). In Macedonia, the institutional variables were under control for the whole period. Decentralization and changes of municipal boundaries happened before the local election of 2005. The rules and procedures for local elections and the district magnitude were constant in all three cycles of local elections. In 2009, the local elections coincided with the presidential election. However, the analysis for 2005 and 2013, when the local elections were held alone, should correct for this.
Taking into consideration the theoretical guidance for analysing party support, the research using municipalities as units of analysis and previous research conducted in Macedonia, I decided to take five different sets of independent variables: geographic factors, demography, socioeconomic conditions, cultural factors and electoral opportunities. I build five different models where I test SDSM's party support in local elections. Each model comprises a set of independent variables. For geographic factors, I take as independent variables whether the municipality is urban or rural, the number of households as an indicator of the size of the municipality, the distance from the capital Skopje measured in kilometres, and the percentage of the population living in villages of up to 50 people. For demographic factors I take as independent variables the average age and levels of education, gender, and the number of illiterate people. For socio-economic conditions, I take as independent variables the number of employed and unemployed persons, the size of the municipal budget, the percentage of social welfare recipients in the total population, the percentage of the municipal budget spent on social welfare and the number of economically inactive citizens. For cultural factors, I take as independent variables ethnic identity, religion and the representation of women in municipal councils. And for electoral opportunities I take district magnitude and voter turnout.
For measuring the independent variables I mainly take data from the census in 2002,ⁱ data from the “Index of Social Inclusion at Local Level” (Karajkov and Simovski 2011) and from the State Electoral Commission.ⁱⁱ The measurement of some independent variables are based on my own calculations. The data for measurement has some limits: for example, the census data is ten years old, while other data is from more recent years. However, on the one hand there is a lack of better and more available data; while on the other, most of the independent variables are of such a character that they should be durable over time (e.g. education levels, number of households, ethnic identity and religion, etc.).
The dependent variable is the vote share for SDSM in the local elections held in 2005, 2009 and 2013. The votes cast for the party's list for municipal councillors were taken as data for measurement. Some limitations need to be pointed out here. In 2005 and 2013, SDSM headed a multi-party coalition list of councillors, while in 2009 the SDSM list was made in coalition only with the Liberal Democratic Party (LDP). This can potentially influence the electoral opportunities structure. However, the type of electoral coalition is a constant condition in all municipalities for one electoral cycle. The main analysis of variance is across municipalities in an electoral cycle. Therefore, the type of electoral coalitions was omitted from the analysis, notwithstanding that it probably had a significant effect on the total votes.

The models, independent variables and expectations are summarized as follows:

 

 

Model 1. Geography

Independent variables	Expectations
Urban/rural split c	SDSM support is stronger in urban municipalities.
Number of households in the municipality a	More households mean greater support for SDSM.
Distance from capital in kilometres c	More distance from capital means less support for SDSM.
Percentage of population living in villages up to 50 people b	More population in villages means  less support for SDSM.

a. Census data, b. Karajkov and Simovski (2011), c. author's calculations

 

Model 2. Demography

Independent variables	Expectations
Gender a	Support for SDSM is not gender-based.
Illiterate persons b	More illiterate persons means less support for SDSM.
Average age in the municipality b	Higher average age means more support for SDSM.
Education levels a	Higher education levels means more support for SDSM

a. Census data, b. Karajkov and Simovski (2011)

 

Model 3. Socio-economic conditions

Independent variables	Expectations
Number of employed a	More employed means higher support for SDSM.
Number of unemployed a	More unemployed means less support for SDSM.
Municipal budget in 2010 b	Bigger budget means more support for SDSM.
Social welfare recipients b	More social welfare recipients means less support for SDSM.
Percentage of municipal budget spent on social welfare b	Higher percentage for social welfare means more support for SDSM.
Economically inactive citizens a	More economically inactive citizens means less support for SDSM.

a. Census data, b. Karajkov and Simovski (2011)

 

Model 4. Cultural factors

Independent variables	Expectations
Ethnic identity a	More Macedonians means higher support for SDSM.
Religion a	More religious citizens means lower support for SDSM.
Women representation in municipal council b	More women represented means higher support for SDSM.

a. Census data, b. Karajkov and Simovski (2011)

Model 5. Electoral opportunities

Independent variables	Expectations
District magnitude a	Bigger district magnitude means higher support for SDSM
Voter turnout a	Higher voter turnout means higher support for SDSM.

a. State Electoral Commission

 

The expectations are in line with the literature and research and with the assumption that SDSM is a centre-left party. Therefore one would expect that it would have higher support in urban areas, in bigger municipalities and in municipalities in the city of Skopje. Gender-based association for party support are rarely reported. In some cases, women can be more left; however, this usually relates to issue-based voting and therefore gender should not play a strong role in Macedonia. Older people are expected to support SDSM more, along with more educated voters. SDSM is expected to have more support in municipalities that have better standards of living (i.e., with less unemployed, more economically active citizens, and a larger budget) and less support where people live in poorer conditions (i.e. with a high number of unemployed and welfare recipients). However, where there is a bigger percentage of the municipal budget dedicated for social welfare one would expect an increase in support for SDSM. Due to ethnic divisions, one would expect ethnic Macedonians to be more likely to vote for SDSM. Religion is expected to have a negative impact on support for SDSM, but a higher representation of women should have a positive impact. More electoral opportunities (i.e., higher district magnitude and voter turnout) should serve to increase support for SDSM.
To test the expectations, I ran multiple regressions for the geographic, demographic, socioeconomic conditions and cultural factors. The effect of the independent variables in each model was tested in the three electoral cycles to check whether inferences were valid over time. The electoral opportunity model was run two times because voter turnout data was available only for the local elections held in 2009 and 2013.
Multiple regression models show the explanatory power of an independent variable in one model. I opted for a discriminatory approach because this is a first attempt to map SDSM's local voter support. In the theory there are only general guidelines. Not enough research has been done in Macedonia. I wanted to avoid random grouping of independent variables. Also, since there are only 85 municipalities in Macedonia, which is my total population of cases, I wanted to avoid too many predictors in one model. Losing the interaction effects between independent variables from one or more models is the obvious limitation with this approach. However, I hope these initial results will point to pathways for future research.

 

Discussion of results

The results of model 1 are shown in Table 1 below. The model has a very good fit. The R2 is 0.752 for SDSM's votes in 2005 (adjusted R2 0.739), 0.705 (adjusted R2 0.690) for SDSM's vote share in 2009 and 0.780 (adjusted R2 0.769) for SDSM's vote share in 2013. The tests are statistically significant with p < 0.01.

 

Table 1. Multiple regression analysis for geographical factors: Votes for SDSM in local elections

Model 1. Geography
Variables	SDSM2005	SDSM2009	SDSM2013
Urban	0.133* (502,027)	0.023 (518,914)	0.36 (716,833)
Households	0.796*** (0.037)	0.843*** (0.038)	0.880*** (0.053)
Distance from Skopje in km	0.201** (5.44)	0.218** (5.643)	0.205** (7.795)
Villages up to 50 people	0.19 (9.014)	0.005 (9.317)	0.028 (12.870)
R2	0.752***	0.705***	0.780***
Adjusted R2	0.739***	0.690***	0.769***
N	82	83	83

* p < 0.1, ** p < 0.05, *** p < 0.01, standard errors in parentheses

 

The impact of the independent variables shows some surprising results. Only in 2005 did SDSM's support have a positive relation in urban municipalities, with weak statistical significance. This means that the urban/rural split was only weakly relevant in the elections of 2005, but not in 2009 and 2013. On the other hand, the number of households is strongly and positively connected to the vote share of SDSM, and this relation is increasing. This means that the size of municipalities is more important than whether they are urban or rural. In bigger municipalities, SDSM gets more support. This is further corroborated with the impact from the distance of municipalities from the capital. There is a positive and statistically significant relationship between distance from Skopje and voting for SDSM. On the other hand, the percentage of citizens living in villages of over 50 people is not relevant. In short, the model based on geographic factors shows support for SDSM increases as the size of municipality increases, and that this is applicable both to urban and rural municipalities and not only to the city of Skopje.
The results from Model 2 are shown in Table 2. The first version of this model is based on four independent variables: males, females, the number of illiterate people and the average age of citizens in the municipality. All tests in the model have a very good fit and are statistically significant with p < 0.001.

 

Table 2. Multiple regression analysis for demographic factors: Votes for SDSM in local elections

Model 2.1 Demography: Gender, illiterates and average age
Variables	SDSM2005	SDSM2009	SDSM2013
Male	-1.661 (0.583)	-2.968 (0.577)	-1.288 (0.853)
Female	2.486 (0.545)	3.603* (0.539)	2.155 (0.798)
Number of illiterates	-0.508 (0.806)	0.152 (0.800)	-0.067 (1.183)
Average age in December 2009	0.301*** (63.882)	0.283*** (63.446)	0.230** (93.876)
R2	0.693***	0.661***	0.710***
Adjusted R2	0.677***	0.664***	0.695***
N	83	84	83

* p < 0.1, ** p < 0.05, *** p < 0.01

The results show that, in line with the expectations, gender did not play a strong role in support for SDSM in local elections. The elections in 2009 were an exception, when there was a statistically weak positive relationship between the number of females in the population and support for SDSM. The number of illiterate people per municipality is not relevant. On the other hand, age is a highly relevant demographic variable. The strength of this predictor has been decreasing over the years; however, a statistically significant positive relationship remains. In municipalities where the average age of citizens is higher, support for SDSM is more likely to be higher. And this is the main point of the tests in Model 2.1.
The second part of the demographic model is shown in Table 3. Education levels are used as independent variables in the tests. The high number of education levels is due to the census categories. The model has a very strong fit and statistical significance with p < 0.01, but the results are mixed.

 

Table 3. Multiple regression analysis for demographic factors: Votes for SDSM in local elections

	Model 2.2 Demography: Education levels
Variables	SDSM2005	SDSM2009	SDSM2013
No education	0.007 (0.494)	0.099 (0.515)	-0.071 (0.878)
Incomplete primary	0.582*** (0.157)	0.680*** (0.164)	0.427** (0.280)
Primary school	-0.445*** (0.040)	-0.475*** (0.042)	-0.236* (0.071)
High school	-0.344* (0.078)	-0.607** (0.081)	-0.117 (0.138)
Vocational school	-0.585** (0.681)	-0.735*** (0.711)	0.132 (1.212)
University education	3.828*** (0.981)	4.351*** (1.024)	1.996* (1.745)
Postgraduate studies	-2.182*** (16.522)	-2.439** (17.248)	-1.420** (29.401)
Doctoral school	-0.282 (7.184)	-0.305 (7.496)	0.041 (12.778)
R2	0.896***	0.875***	0.858***
Adjusted R2	0.885***	0.862***	0.843***
N	83	84	84

* p < 0.1, ** p < 0.05, *** p < 0.01

 

 

The results show that there are statistically significant and positive relationships between incomplete primary education and university education and voting for SDSM. University education is the strongest predictor for support for SDSM, though one should not neglect the fact that incomplete primary education is also a predictor for support. On the other hand, primary school, high school and vocational education showed a negative relationship with support for SDSM in the local elections held in 2005 and 2009, though this does not apply for the local elections held in 2013. Postgraduate education shows a constant statistically significant and negative relationship with support for SDSM, and doctoral level is not relevant. The tests show that university education remains the strongest predictor for SDSM support, while incomplete primary education is a weaker predictor. On the other hand, postgraduate studies remains a negative predictor, along with primary school as a weaker negative predictor.
The results from Model 3, based on socio-economic conditions, are given in Table 4. The model has a very strong fit and all tests are statistically significant with p < 0.01. The results are mostly in line with the expectations, though with a surprising twist.

 

Table 4. Multiple regression analysis for socioeconomic factors: Votes for SDSM in local elections

	Model 3. Socioeconomic conditions
Variables	SDSM2005	SDSM2009	SDSM2013
Number of employed	0.944*** (0.054)	0.904*** (0.060)	0.823*** (0.089)
Number of unemployed	0.519*** (0.101)	0.543*** (0.111)	0.557*** (0.166)
Municipal budget in 2010	0.050 (0.000)	0.035 (0.000)	0.008 (0.000)
Social welfare recipients	-0.024 (34.968)	0.016 (38.540)	0.000 (57.552)
Municipal budget spent on social welfare	-0.016 (57.192)	-0.015 (63.114)	-0.014 (94.247)
Economically inactive citizens	-0.618*** (0.038)	-0.599*** (0.042)	-0.477*** (0.062)
R2	0.861***	0.813***	0.837***
Adjusted R2	0.850***	0.798***	0.824***
N	83	84	84

* p < 0.1, ** p < 0.05, *** p < 0.01
The number of social welfare recipients, the size of the budget of municipality and the percentage spent on social welfare are not relevant to support for SDSM in local elections. This means that one of the main tenets of social policy is not related to support for SDSM as a social democratic party. The number of employed and unemployed (i.e. economically active citizens) both have a statistically significant and positive relationship with support for SDSM. The predictor for the number of employed is stronger, although declining from the elections in 2005 to 2009. On the other hand, the effect of economically inactive citizens (e.g. housewives, pensioners, students) is significant and in a negative relationship with support for SDSM, although the strength of the predictor is similarly in decline.
The results from Model 4, based on cultural factors, is given in Table 5. The tests have a very strong fit and high statistical significance with p < 0.001. The results are also in line with the expectations.

 

Table 5. Multiple regression analysis for cultural factors: Votes for SDSM in local elections

	Model 4. Cultural factors
Variables	SDSM2005	SDSM2009	SDSM2013
Macedonians	1.145*** (0.019)	0.975*** (0.018)	1.123*** (0.036)
Albanians	-0.005 (0.027)	-0.083 (0.026)	-0.044 (0.050)
Others	-0.134 (0.015)	-0.402*** (0.015)	-0.105 (0.029)
Christian Orthodox	-0.570*** (0.019)	-0.819*** (0.018)	-0.386* (0.036)
Muslim	-0.174* (0.033)	-0.273** (0.031)	0.017 (0.061)
Other	0.340*** (0.395)	0.636*** (0.375)	0.120 (0.733)
Women in municipal council	0.048 (15.229)	0.062 (13.819)	0.011 (27.034)
R2	0.907***	0.906***	0.862***
Adjusted R2	0.898***	0.897***	0.848***
N	80	81	81

* p < 0.1, ** p < 0.05, *** p < 0.01

 

The number of ethnic Macedonians is the strongest predictor for support of SDSM, with a strong statistical significance and positive relationship. The more ethnic Macedonians in a municipality, the more likely it is that support for SDSM will be higher. In 2009, other ethnic groups showed a positive and significant relationship with SDSM. On the other hand, membership of the Christian Orthodox religion had a significant negative relationship with support for SDSM, increasing from local elections in 2005 to 2009. However, the strength and significance of this relationship had significantly decreased by 2013. It is interesting to note that in 2005 and 2009 other religious groups (Catholics, Protestants, etc) had a positive and significant relationship with SDSM’s share of votes. It seems that SDSM is mainly supported by ethnic Macedonians, negatively associated with mainstream religions, and open for support from other religious groups. Also it is worth pointing out that women’s representation in councils, a recognizable policy and achievement of SDSM, is not significant for party support.
The results of the last model, based on electoral opportunities, are presented in Table 6. The tests are statistically significant with p < 0.01. The fit is better for the test for 2013, explaining over 70 per cent of the variance, than the test for 2009, which explains over 50 per cent of the variance in results.

 

Table 6. Multiple regression analysis for electoral opportunity factors: Votes for SDSM in local elections

	Model 5. Electoral opportunities
Variables	SDSM2009	SDSM2013
District magnitude	0.891*** (45.069)	0.940*** (55.698)
Voter turnout	0.423*** (23.500)	0.440*** (37.033)
R2	0.585***	0.712***
Adjusted R2	0.574***	0.704***
N	79	79

* p < 0.1, ** p < 0.05, *** p < 0.01
District magnitude is constant across municipalities in the two electoral cycles. It is positively related and significant in both cases. Interestingly the predictor is also increasing. This again corroborates the view that support for SDSM is more likely to be higher in bigger municipalities, which by default have bigger district magnitude (e.g. award more seats in local councils). Voter turnout also has a positive and significant relationship with support for SDSM. When the voter turnout increases, it is more likely that SDSM will have higher support.

 

Conclusion

The paper aims to identify the predictors for SDSM's party support at local level. Research in Macedonia has seldom touched upon this issue. Ethnic identity was found to be the main predictor for party support (Shahla 2005), and some expectations for party support from post-communist countries were partially confirmed (Petkova-Hristova 2010). This is a first attempt to test support for SDSM at local level.
Using the standard literature for party support and similar work from other countries, I derived five sets of independent variables. These are labelled as follows: geography, demography, socio-economic conditions, cultural factors, and electoral opportunity structures. I built a database using available data from the State Statistical Office, the State Electoral Commission, and a study on social inclusion at local level. The level of analysis is set at local level with municipalities as the units of analysis. The vote share for SDSM's list for councillors is the dependent variable. It was measured in three cycles of local elections: 2005, 2009 and 2013. Multiple regression analysis was chosen as the appropriate method to test the influence of the independent variables.
Some of the findings confirm the expectations. SDSM's support is not gender-based and it is higher in bigger municipalities where there are more households. A higher proportion of older people is likely to indicate greater support for SDSM. Support for SDSM comes mainly from ethnic Macedonians. In 2005 and 2009, a higher number of religious citizens meant lower support for SDSM; however, this weakened in 2013. A greater number of electoral opportunities results in higher support for SDSM.
On the other hand, some of the findings are surprising and contrary to the expectations. Urban municipalities and proximity to Skopje, as well as municipalities in Skopje, do not produce more support for SDSM. This means that SDSM has equal chances in different municipalities regardless of their location and of whether they are urban or rural. SDSM's support is also not related to the budget of the municipality, meaning it does not matter whether the municipality is rich or poor. The education levels of citizens show mixed results. University education is a solid predictor for support for SDSM, but so is incomplete primary education. This means that SDSM has support both from highly educated and from uneducated people. In 2005 and 2009, support for SDSM was negatively related to other educational levels (e.g. completed primary and high school education), though this changed in 2013.
High numbers of economically active citizens, both employed and unemployed, i.e., those working and those trying to find work, indicate support for SDSM, while high numbers of economically inactive citizens indicate lower support. The first group are directly concerned with job market opportunities and protection of workers’ rights. Even though SDSM's policy track record in these areas is mixed, it seems that job market opportunities and protection of workers’ rights are key policy areas for the political base of SDSM. On the other hand, social welfare spending and increasing gender balance, as recognizable social democracies policy areas, do not matter for the political base of SDSM.
The results can be useful as directions for developing party strategy. The results can also show pathways for future academic research. For example, one can compare the findings with parliamentary elections to see whether the inferences hold in both types of elections. Most of the independent variables would be the same, although electoral opportunity structures would be different. It would also be interesting to compare support for SDSM with support for other political parties. This would shift the focus of analysis from support for one party to develop a more detailed picture of party system dynamics. To keep the present focus and go further in detail, one can juxtapose municipalities where SDSM has strong support with municipalities where the party has weak support. This approach would keep the same focus of interest, but case-oriented research methods would be more appropriate. As a first attempt to map SDSM's party support at local level, the results presented here cannot be definite. Rather, this work should open new issues and raise new questions relevant for academics and other interested members of the public.

------------------

ⁱ For more details, see State Statistical Office of the Republic of Macedonia (www.stat.gov.mk).

ⁱⁱ For more details, see State Electoral Commission of the Republic of Macedonia (www.sec.mk).

 

 

References

Barany Z (2005) Ethnic Mobilzation in the Postcommunist context: Albanians in Macedonia and the East European Roma. In: Barany Z and Moser RG (eds) Ethnic Politics after Communism. Ithaca and London: Cornell University Press, pp. 78-108.

Barany Z and Moser RG (2005) Ethnic Politics after Communism. Ithaca and London: Cornell University Press.

Barreto M, Villarreal M and Woods ND (2005) Metropolitan Latino Political Behavior: Voter Turnout and Candidate Preferences in Los Angeles. Journal of Urban Affairs, 27 (1): 71 – 91.

Bartolini S (2000) The Political Mobilization of the European Left, 1860 – 1980. The Class Cleavage. Cambridge: Cambridge University Press.

Bartolini S and Mair P (1990) Identity, competition and electoral availability. The stabilization of European electorates 1885 – 1985. Cambridge, New York: Cambridge University Press.

Birnir JK (2007) Ethnicity and Electoral Politics. Cambridge: Cambridge University Press.

Bornschier S (2009) Cleavage Politics in Old and New Democracies. A review of literature and avenues for future research. EUI Working Papers MWP 2009/07.

Casal Bértoa F (2013) Post-Communist Politics: On the Divergence (and/or Convergence) of East and West. Government and Opposition 48 (3): 398 – 433.

Casal Bértoa F (2014) Party systems and cleavage structures revisited: A sociological explanation of party system institutionalization in East Central Europe. Party Politics, Epub ahead of print, DOI: 10.1177/1354068811436042.

Čular G (2005) Political Parties and Support for Democracy. In: Čular G (ed) Elections and Consolidation of Democracy in Croatia. Zagreb: Faculty of Political Science, University in Zagreb, pp. 123 – 174. (In Croatian).

Dalton RJ (1996) Comparative Politics: Micro-Behavioral Perspectives. In: Robin G and Klingelmann HD (eds) A New Handbook of Political Science. Oxford: Oxford University Press,pp. 336 – 352.

Enyedi Z (2005) The role of agency in cleavage formation. European Journal of Political Research 44 (5): 697 – 720.

Enyedi Z (2008) The Social and Attitudinal Basis of Political Parties: Cleavage Politics Revisited. European Review 16 (3): 287 – 304.

Evans G (2006) The social basis of Political Division is Post-Communist Eastern Europe. Annual review of Sociology (32): 245-270.

Fink Hafner D and Krašovec A (2006) Ideological – political cleavages in Slovenian party arena after 1989 year. In: Fink Hafner D and Pejanović M (eds) Development of Political Pluralism in Slovenia and Bosnia and Herzegovina. Sarajevo/Ljubljana: Promocult, Faculty of Social Sciences Ljubljana, Foundation FDV, Faculty of Political Science in Sarajevo, pp. 71 – 105.

Franklin M et al. (1992). Electoral Change: Responses to Evolving Social and Attitudinal Structures in Western Countries. Cambridge: Cambridge University Press.

Hajnal ZL and Lewis PG (2003) Municipal Institutions And Voter Turnout in Local Elections. Urban Affairs Review. 38(5): 645 – 668.

Jasiewicz K (2009) Poland. In: Sagar DJ (ed) Political Parties of the World, Seventh Edition. London: John Harper Publishing.

Karajkov R and Simovski A (2011) Index Of Social Inclusion On Local Level. Skopje: Evro-Balkan Press (available on line at http://www.euba.edu.mk/tl_files/Publikacii%20EUBA/Indeks%20na%20socijalna%20inkluzija%20-%20WEB.pdf

Kestilä E and Söderlund P (2007) Local Determinants of Radical Right Wing Voting: The Case of the Norwegian Progress Party. West European Politics 30(3): 549 – 572.

Kitschelt H (1994) The Transformation of European Social Democracy. Cambridge: Cambridge University Press.

Kitschelt H with McGann AJ (1995b) The Radical Right in Western Europe. A Comparative Analysis. Ann Arbor: University of Michigan Press.

Kitschelt H, Mansfeldova Z, Markowski R and Toka G (1999). Post-Communist Party Systems: Competition, Representation and Inter-Party Cooperation. Cambridge: Cambridge University Press.

Kitschelt H (1995a) Formation of Party Cleavages in Post-Communist Democracies. Party Politics 1 (4): 447 – 472.

Klos J (2008) Determining Spatial Correlations Between Voting Behavior and Selected Demographic Variables in a Changing Electorate. Middle State Geographer 41: 19 – 26.

Knutsen O and Scarbrough E (1995) Cleavage Politics. In: Van Deth JW and Scarbrough E (eds) The Impact of Values. Oxford and New York: Oxford University Press, pp. 492 – 523.

Körösényi A (1999) Government and Politics in Hungary. Budapest: CEU Press-Osiris.

Lipset SM and Rokkan S (eds) (1967) Party Systems and Voter Alignments: Cross-National Perspectives. New York: Free Press.

Mainwaring SP (1999) Rethinking Party Systems in the Third Wave of Democratization, The Case of Brazil. Stanford: Stanford University Press.

Mainwaring SP and Zoco E (2007) Political Sequences and the Stabilization of Interparty Competition, Electoral Volatility in Old and New Democracies. Party Politics 13(2): 155 – 178.

Markowski R (1997) Political parties and ideological spaces in East Central European societies. Communist and Post-Communist Studies (30): 221 – 254.

Mateju P, Rehakova B and Evans G (1999) The politics of interests and class realignment in the Czech Republic, 1992–96. In: Evan G (ed) The End of Class Politics?: Class Voting in Comparative Context. Oxford: Oxford University Press.

McAllister I and White S (2007) Political Parties and Democratic Consolidation in Post-communist societies. Party Politics 13(2): 197 – 216.

Mojanoski C (1996) Social and Political Profile of the Political Parties in Macedonia. Skopje: Liber (In Macedonian).

Moser RG (2005) Ethnicity, Elections, and Party Systems in Postcommunist States. In: Barany Z and . Moser RG (eds) Ethnic Politics after Communism. Ithaca and London: Cornell University Press, pp. 108-140.

Neto OA and Cox GW (1997) Electoral Institutions, Cleavage Structures and the Number of Parties. American Journal of Political Science. 41(1): 149-174.

Norris P (2004) Electoral engineering: voting rules and political behavior. Cambridge: Cambridge University Press.

Ordeshook P and Shvetsova OV (1994) Ethnic Heterogeneity, District Magnitude and the Number of Parties. American Journal of Political Science. 38 (1): 100-123.

Panov T and Taleski, D (2014) What Drives Political Divisions in Macedonia? Social Structures, Attitudes Or Perceptions of Communist Past As Basis For Political Divisions. Manuscript in preparation for publishing.

Petkovska–Hristova L (2010) Left and Right in the Post Communist Countries. New Balkan Politics.12

Shabad G and Slomczynski KM (1999) Political Identities in the Initial Phase of Systemic Transformation in Poland: A Test of the Tabula Rasa Hypothesis. Comparative Political Studies 32 (6):  690-723.

Shahla H (2005) Voting Behavior in Macedonia. In: Škarić S Democratic Elections in Macedonia, 1990 – 2002. Analyses, Documents and Data. Berlin: WZB, Edition Sigma. pp. 119 – 137.

Šiber I (2001) Political behavior of voters in elections 1990-2000. In: Kasapović M (ed) Croatian Politics 1990 – 2000.Zagreb: Faculty of Political Science, University in Zagreb, pp.  65 – 99. (In Croatian)

Siljanovska Davkova G (2006) Macedonian Political Partis Through the Prism of Ideology. In: Lutovac Z (ed) Political Parties and Voters in the Countries of former Yugoslavia. Belgrade: Friedrich Ebert Foundation and Institute for Social Sciences, pp. 205 – 237 (In Serbian).

Škarić S (2005) Democratic Elections in Macedonia, 1990 – 2002. Analyses, Documents and Data. Berlin: WZB, Edition Sigma.

Taleski D (2008) Minorities and Political Parties in Macedonia. In: Biber F (ed) Political Parties and Minority Participation. Skopje: Friedrich Ebert Foundation, pp. 127 – 153.

Tucker JA and Pop-Eleches G (In Progress) Communism's Shadows: Historical Legacies, and Political Values and Behavior. Available at https://files.nyu.edu/jat7/public/pubs.html (accessed on September 25, 2013).

Tworzecki H (2003) Learning to Choose: ElectoralPolitics in East-Central Europe. Stanford: Stanford University Press.

Zakošek N (2002) Polical System of Croatia. Zagreb: Faculty of Political Science, University in Zagreb (In Croatian).