Twenty years after the fall of communism and the rebirth of multiparty politics in Central and Eastern Europe, the party systems evolved from high levels of fragmentation and volatility characterizing the 1990’s (Mair 1997; Rose et al. 1998; Lewis 2000; Birch 2001; van Biezen 2003) to a general continuity and stability visible in the most recent decade (Kolarova 2002; Enyedi 2006; Rose and Munro 2009). The Hungarian and the Czech party system appear to be the least electoral volatile (Sikk 2005; Tavits 2005), whereas in longitudinal terms starting with the third elections in the Czech Republic, Hungary, Poland, and Slovakia (Sitter 2002), and with the fourth in Romania (Gherghina 2009). In this entire process of stabilization, the ideological stances of political parties and their positioning in the political space play relevant roles. Kitschelt (1992) argues that the dynamics of post-communist politics indicates, from the beginning, a clear orientation towards the spatial dimension of party competition. The major explanations provided by previous studies for the electoral instability of post-communist parties partially or totally touch upon the role played by spatial positions in achieving specific electoral outcome: the initial institutional choices (Elster et al. 1998 ; Kostelecky 2002; Birch 2003), the structure of cleavages and the incumbency effect to be reflected in the socio-economic conditions (Evans and Whitefield 1993; Bielasiak 1997; Tworzecki 2003), and new party entries (Spirova 2007; Tavits 2008; Rose and Munro 2009). On the other hand, there is evidence that the citizens of these countries use pure value-based voting to attach party loyalties (Tóka, 1998). Whitefield (2002) finds evidence for the social and ideological divisions in the post-communist societies, with socio-demographic factors making a difference in/when choosing ideologies.

The Romanian party system witnessed a relatively slow but solid stabilization, being positioned between the extreme cases of fast (e.g Czech Republic and Hungary) and slow and tedious institutionalization (Bulgaria and Poland). Therefore, its investigation provides relevant information on the dynamics of party politics within moderated cases. Three further particular elements of the Romanian party system make it relevant for scrutiny and highlight an apparent paradox. First, it displays a continuous decrease in the number of successful political competitors. The May 1990 elections allowed 16 political parties to get into Parliament (second Chamber), the 1992 elections reduced this number to seven, in the 1996 elections there were six parties, one less in 2000, whereas for the 2004 and 2008 this number stabilized to four. Second, the number of entries in the Romanian party system is very small. Three out of four parties from the current legislature are in Parliament from their initial creation, whereas the last entry in the range of successful competitors was registered in 2000 (due to a coalition with the social-democrats). With newly created parties having little if any chances to enter the system, we can expect a sharp ideological separation between the existing actors, clear positioning in the ideological space (to maximize their votes). Third, the contemporary development of the party system partly contradicts such an expectation and reveals the necessity of continuous ideological clarifications among the political parties. The Democrat Liberal Party, one of the three main actors in the current party system undergoes ideological changes following its shift from social-democracy to conservatism in 2005. Moreover, the system currently faces the creation of a new leftist party, based on a group of independent MP’s, who left the liberal and social-democratic opposition and now support the government. Whether we speak of major ideological shifts of parties or individual transfers from one party with a particular ideology to another, the poor politicians’ attachment to particular ideologies impedes the creation of electoral loyalties among voters on the long-run (Mair 1997). The direct consequence of this situation is a volatile situation at the level of representation where political identities and organizational loyalties are recomposed every electoral cycle (Roberts and Wibbels 1999, 575).

Summing up, a puzzle is visible: the Romanian party system gains stability although there is an apparent quest of some political parties for their ideological identity. Is there any stability in terms of ideological competition going on in Romanian politics? This article seeks an answer to this question, by revealing the evolution of the Romanian party system from an ideological perspective. In doing so, we aim to determine the degree of ideological institutionalization during the two decades of post-communism. Our enquiry is theoretically driven and we use process-tracing as major methodological tool. The unit of analysis is the ideological family (i.e. a concept situated between parties – legally “palpable” structures – and a party system – purely theoretical structure). In an ideological family there can be one or several parties and in any viable party system there should be at least two ideological families. We analyze the entire post-communist period, starting with the first free elections in 1990 and ending with the most recent 2008 elections.

The article begins with a review of the concept of institutionalization, identifying the position of the current article in the existing theoretical debates. The second section presents the methodological framework of the analysis, with an emphasis on the used tools. The analysis from the third section uses process-tracing to explain the Romanian electoral results. Finally, the conclusions summarize the main findings and open the floor for further research.

The Theoretical and Conceptual Frameworks
The institutionalization is considered either a dichotomous phenomenon that may appear or not (Sartori 1976) or a process through which parties and party systems transform and make steps towards becoming institutionalized or the reverse (Mainwaring and Torcall 2006, 206). Huntington (1965; 1968) defines institutionalization as “a process through which organizations acquire value and stability”. Following this line of argument, institutions can be regarded as stability conducive mechanisms, having the capacity to derive stable outcomes (Shepsle 1989). stability represents one key assumption within the current study and can refer both to internal aspects, to be found in the formal institutionalization of parties, and to external aspects related with the effect that party institutionalization has on voters. Three pillars reflect diverse party features that are relevant in the formal institutionalization process. Thus, the regulative pillar refers to constraints and regulations coming from the institution towards its members. Through these, the party establishes rules, monitors activities, and sanctions behavior. The normative pillar includes both values and norms on which the party was built and wants to be perpetuated or adjusted. When norms and values are attributed to certain specific actors: these institutionalized positions become roles (Scott 2001, 55). The cultural-cognitive pillar implies “the shared conceptions that constitute the nature of social reality and the frames through which meaning is made” (Scott 2001, 57).
Mainwaring and Scully (1995) argue that institutionalization is “the process by which a practice or organization becomes well established and widely known, if not universally accepted”. With respect to party institutionalization, Huntington (1968) identifies four dimensions: adaptability, complexity, autonomy and coherence. Adaptability refers to the capacity of a party to resist in time and to survive its leaders. Organizational complexity refers to the number of subunits. Autonomy deals with the degree of differentiation between the party and the behavior of other social subgroups. Coherence regards the capacity to reach consensus and to resolve the disputes inside the party.

Panebianco (1988) sees institutionalization as a “solidification” of a political organization, which becomes “valuable in itself”, and its survival becomes the “goal” of its followers. Institutionalization should be understood as the process by which the party becomes established in terms both of integrated patterns of behavior and of attitudes, or culture. We suggest further that it is helpful to distinguish between internal and externally related aspects of this process. Internal aspects refer to developments within the party itself; external aspects have to do with the party’s relationship with the society in which it is embedded, including other institutions. Within each of these aspects there will be a structural and an attitudinal component.” (Randall and Svasand 2002, 12) In the life of an organization there are three phases: genesis, institutionalization and maturity (Panebianco 1988, 19). The particular combination of organizational factors from the genesis phase influences both the degree of institutionalization and the forms of institutionalization. Thus, some parties become strong institutions, others hardly institutionalize at all. (Panebianco 1988, 19).

Randall and Svasand (2002) believe the term systemness, in Panebianco’s understanding, overlaps with the concepts of complexity and coherence, as understood by Huntington, wile autonomy is regarded in the same manner by both authors. Janda (1980) does not consider autonomy to be an essential feature of institutionalization of parties, bringing as an example the Labour Party in the UK, which exists in strong connection with the unions. The alternative concept put forward by Janda is “external institutionalization”, by which a party becomes a point of reference in the actions of all other social and political actors. Levitsky (1998) puts forward the concept of “value infusion”, while Randall and Svasand speak of the attitudinal dimension of institutionalization, both expressing the idea that a party becomes institutionalized, with a stable and loyal electorate, if the values it promotes are constant and coherent from an ideological perspective, thus making the party a predictable one.

We notice that these definitions and operationalizations of institutionalization emphasize not only the persistence over time (Riker and Ordershook, 1973; Scott, 2001), but also touch on the idea of ideological stability. Transposing them to the intention of determining the degree of ideological institutionalization at the level of a party system, we can say it deals with the capacity of a particular ideological stream to acquire stability and intrinsic value. However, there is a difference of nuance to be pointed out. While we do not discuss directly the institutionalization of parties, they remain at the heart of our conceptualization as the vehicles through which ideological streams become institutionalized or not. An ideological stream does not have an own will, it is not “a living body” that can make decisions as it is the case of political parties. This is why we speak of ideological institutionalization solely based on the actions and results of the parties that reflect a particular ideological stream and on the reactions of the citizens, expressed through the votes they cast in elections.

The organizational component plays an important part in the discussions about party institutionalization, but for the purpose of this study it is less important what happens inside the parties that form a particular ideological stream; the actions of the party as a whole matter. Hence, we approach institutionalization from two perspectives. First, it is the process by which parties behave coherently from an ideological perspective (Przeworski 1975; Mainwaring and Scully 1995; Mainwaring 1998). In other words, the actions of a party must be predictable and stable from the point of view of its ideology. Second, it is the process through which the electorate affiliates and stabilizes from an ideological perspective, through which ideological links between parties and citizens are formed, meaning that citizens constantly vote for the party or the parties that represent a particular ideology. In other words, we can speak of ideological institutionalization when parties become stable by creating ideological roots in the society.

One further theoretical clarification is necessary. There is an extensive theoretical debate regarding the overlap between the institutionalization of parties and the institutionalization of party systems. Authors slowly agree that one does not necessarily imply the other (Casal Bertoa 2007). In this article, we refer mainly to the ideological institutionalization using analogies with the institutionalization of parties (having in mind the clarifications made above). As the core of a party system is represented by the patterns of interaction between and within its subunits (i.e. political parties), the stability in the rules and nature of inter-party competition are the crucial components of party system institutionalization. The fundamental precondition raised by Sartori (1976) when defining a party system has to be recalled: in order to speak of a “system” we must have at least two parties, because a system implies the existence of at least two components interacting and mutually influencing their actions. The more stable the system (i.e., structured inter-interaction), the more institutionalized it is (Przeworski, 1975; Mair, 2000).

Research Design: Operationalization, Data, and Method
The institutionalization is operationalized in this study as low electoral volatility and lack of fragmentation. Our unit of analysis (i.e. the party family) corresponds to Bartolini and Mair’s (1990) concept of block as the intermediate level of representation between the party system and party. A block of parties can be either a party family or a group of parties occupying a common place on the left-right continuum or old vs. new and opposition vs. government continuum (Pennings et al., 1998). At block level, we can distinguish between intra- and inter-block dynamics with the first involving the changes between parties from the same bloc, whereas the second catches the behavior between blocks. For instance, when calculating the electoral volatility and the effective number of parties within a party family, we use the mathematical formula applicable to parties. The ideological stream/family is regarded as a miniature party system. As we will see, there are instances when inside an ideological stream there is only one party representing it, in such cases the effective number of parties being 1.

The concept of ideological family is used by Klaus von Beyme (1984), who identified ten ideological families , but his categorization combines ideological and theoretical criteria with aspects related more to the practical actions of the parties. Another perspective, used even more widely in the analysis of parties, as well as in journalistic analysis, is the one based on the left-right continuum. This is somewhat an exaggerated simplification of Lipset and Rokkan’s (1967) theory on social cleavages, although many authors, most notably Kitschelt (1995), created spatial models for the distribution of parties and party families inside the party system, based on criteria relating to the actions of the parties and the social contexts in which they are active.

Our study considers three major ideological families: social-democracy, conservatism, and liberalism. These families are common all major European party systems and have strong historical and philosophical roots. It is not the goal of this article to establish whether they represent the left or the right or on which side of particular cleavages they are situated in Romania. Also, the three families have been selected for this study because they are the main ideological streams within the European Parliament and hold the most numerous three parliamentary groups. Their presence is continuous within the Romanian party system from its re-emergence: within the system there has always been at least one party claiming to represent each of these families.

Hence, this study does not use the classical understanding, particularly valid for Romania, that social-democrats represent the left, conservatives the center-right and liberals the right. For the same reasons, we do not take into consideration the relative ambiguity on the identity of the liberals, considered to represent the left in Western Europe and the right in post-communist Europe. We consider the three ideological families to be disposed in a triangle, as reflected in Figure 1, which helps us to better understand the interactions among the three families, each family having in common a particular feature with each of the other two. Social-democrats and liberals share their opening towards social freedoms (as opposed to the conservative traditionalism), liberals and conservatives advocate the minimal state in economic affairs (while social-democrats support the state’s involvement in economy), and conservatives and social-democrats emphasize the community as a primordial structure of the society (while liberals consider the individual to be the core component of the society).

In order to determine the degree of institutionalization of ideological families in Romania, we analyze the electoral results obtained by the parties composing each family in each of the parliamentary elections taking place in Romania after the fall of communism: 1990, 1992, 1996, 2000, 2004, and 2008. For simplicity, we account only for the results obtained by parties in the elections for the Chamber of Deputies. Also, in all six rounds of elections, we consider only those parties obtaining over 5% of the votes and the electoral alliances that passed the specific threshold set for them as established since the 2000 elections. The reason is that we must use the same criteria for selection in all six elections. The electoral threshold varied from 0 in 1990 to 3% in 1992 and 1996 and then to 5% since 2000. The number of parliamentary parties in 1990 was high and including in the analysis all the parties that gain seats by obtaining only a few thousand votes would have a minimal impact and complicate matters without any reason.

In order to determine the parties that belong to each family for each round of elections, we use four elements. The first element is the name of the party. It is self-speaking many times and offers a clear indication on the party’s orientation. Thus, the Social-Democratic Party, the National Liberal Party, the National Peasant Party – Christian-Democrat indicate through their name the ideological family they belong to. However, the name can sometimes be misleading or may not offer clear indications. Such examples are the Democratic Party (today the Democratic-Liberal Party), the Greater Romania Party, and the National Salvation Front. Thus, the second element used in establishing the ideological identity is the party’s internal documents where the ideological orientation is specified or at least where certain indications are given based on the stated principles regarding the economy or the social matters. A third element is the membership in international associations of parties, be them European parties or the so called “internationals”. Even if in the ’90’s this element is not particularly useful, because the Romanian parties were not fulfilling the conditions to become members of these associations. After 2000, their belonging to these organizations certified their closeness to particular ideological families. There is also some ambiguity here concerning the Conservative Party (formerly known as the Romanian Humanist Party), which oscillated regarding its international belonging. It tried to become a member of the European People’s Party, but it was rejected, so the Conservatives reoriented towards the European Liberal-Democrat, so the first MEP’s of this party joined the ALDE group in the EP. However, the program of the Conservative Party recommends it as a member of the homonym family, even if in the 2000, 2004 and 2008 elections it ran in an alliance with the social-democrats and at one point the name of the party included the words “social-liberal”. The final element that helps the ideological classification of parties in a particular electoral moment is the past and the future of the party. For instance, we include the National Salvation Front in the family of social-democrats, having in mind that it evolved in the social-democratic parties. While none of the four elements indicates by itself without doubt the ideological family of a party, taken together they provide an accurate image.

There are three methodological clarifications to be made before the analysis. First, we do not take into consideration the Democratic Alliance of Hungarians in Romania (UDMR). It holds a special position in the Romanian party system, having its own, highly stable, electorate. Although it is a member of the EPP, being recognized as a formation with a conservative orientation (despite the fact that inside it there are several ideological streams), we do not include it in the analysis because it does not effectively compete against the other parties for votes. A good mobilization of its own electorate – the Hungarian minority – is sufficient for passing the threshold. Second, the problem of alliances needs to be clarified. This analysis is based on the percentages of votes obtained by parties in elections, but in the case of alliances we speak of a cumulated score of several parties. For the parties winning seats as part of an alliance, we account for the number of seats gained by the party following the distribution and we proportionally calculate, based on the electoral results of the alliance, the percentage corresponding to the number of seats for each party. The result is the electoral score of the party in the respective round of elections. This approach towards alliances minimizes miscalculations, even if there is always a difference between the percentage of seats and the percentage of votes obtained by a party. The difference, however, is relatively small and it does not influence significantly the results of the analysis. Furthermore, we consider the cumulated score of parties belonging to the same ideological family running in elections as part of the same alliance. This is the case of the liberals in the Romanian Democratic Convention in 1992 and 1996, for instance.

The third aspect that should be clarified here is how the two nationalistic parties appearing in the analysis will be classified: the Romanian National Unity Party (PUNR) and the Greater Romania Party (PRM). PUNR is relevant for the 1992 elections, when it gained 7.72% of the votes. This is a party that was active initially only in Transylvania, being in close connection with the nationalistic organization Romanian Cradle, later merging with other smaller parties and gaining national representation. The main objective of this party was the promotion of “national policies” and in the 90’s it was in the sphere of influence of the Party for Social-Democracy in Romania (today the Social-Democratic Party), collaborating in government with it before the 1996 elections. However, it is difficult to place PUNR in one of the three ideological families included in the study as the nationalistic discourse was its main feature without a specific economic program or clear political identity that would help us to include it in one of the three main families. Moreover, following its collaboration with the social-democrats in the ’90s, it eventually merged with the conservatives. On the other hand, PRM is also a party with a strong nationalistic discourse, but it is included in the family of social-democratic (or rather socialist) parties. Besides its collaboration with the social-democrats in the ’90s, the political program of PRM speaks of “the equality of chances for all citizens, human solidarity and family values” and of “the development of national economy as a condition for a decent living for all citizens”. These elements and the public interventions of its president in favor of the state’s intervention in economy or against wealthy businessmen place PRM in the social-democrat corner of our triangle. Table 1 includes the electoral results taken into consideration in the analysis.

Source: The election results, available at Essex Database (last accessed 21 May 2010) and on the Website of the Central Electoral Bureau (last accessed 22 May 2010).

For each round of elections, we add the results of all the parties that obtain seats in the Parliament, regardless if they ran on their own or as part of an alliance. Based on this number we calculate the level of electoral volatility for each ideological family from one round of elections to the next. Birch (2001) calculates the volatility in a relative manner, reporting the difference of vote shares to the total of votes received by the party in both elections. This measurement accurately describes party fluctuations in support relatively to its share of votes. This is the reason for which we use the same formula for our calculations:

Vj = party j electoral volatility,
Vt0 = the share of votes obtained by party j at moment/election t0 (the initial elections).
Vt1 = the share of votes obtained by party j at moment/election t1 (most recent elections).

We consider each ideological family as representing a party system and calculate the effective number of parties belonging to each family in order to determine the degree of fragmentation in each family. The formula used to determine the effective number of parties is the following:

ef = the effective number of parties within a family,
N = the mathematical number of parties within the family,
Vi= the percentage of votes obtained by party i relative to the total number of votes obtained by the ideological family.

How Institutionalized are the Romanian Party Families?
This section presents the results obtained after calculating the electoral volatility and effective number of parties for the Romanian ideological party families. After a general comparison of the observed trends, we delve into details and explain them in a narrative manner. In doing so, we point out the key elements in the evolution of each family following the fall of communism.
Figure 2 graphically depicts the levels of electoral volatility for each analyzed ideological family. Three features are observable. First, the levels of volatility greatly differ between the three families: the conservative family is the most volatile, followed by the social-democrats and liberals. In fact, the latter two have quite similar averages, with alternative peaks. For example, the liberal family is more volatile than the social-democratic block in 1992-1996 and 2000-2004, whereas the social-democrats register higher levels of volatility in the 1990-1992, 1996-2000, and 2004-2008 electoral cycles. Second, all three families register oscillations, there are no consecutive electoral cycles in which the volatility increases or decreases. Third, the conservative and social-democratic families follow similar patterns of increase and decrease. Their dynamics coincides in all the post-communist electoral cycles. Furthermore, the figure shows that whenever one of the two families wins votes, from one election to the next, the other one loses and viceversa, as pointed out by the “+” and “-“ signs that indicate increases and decreases in the electoral strength of a family within a cycle. Thus, we see that there has been a constant exchange of votes between the social-democrats and the conservatives, which may in fact represent a bulk of average voters which are regularly the target of catch-all parties and that switch back and forth between the two. However, this study does not bring data on this segment of the electorate.

Source: Compiled on the basis of the electoral results from Table 1 and the formula used by Birch (2001).

Table 2 includes the effective number of parties for each ideological family, calculated for the electoral cycles. We can easily notice that liberalism has one representative for all the elections (the issue will be extensively discussed in the following lines), whereas a similar pattern is visible for conservatism with the exception of the most recent elections when the number increased from 1 to 1.08. The social-democratic family appeared to be the most contested with one effective party in the first and most recent post-communist elections. Apart from those two instances, the number varies from 1.88 in 1996 to 2.63 in 2004.

Source: Compiled based on the electoral results in Table 1.

The Conservatives
In the case of the conservative family, we can notice that the effective number of parties is constantly 1, with the exception of the 1990 elections, when no party of this family gained over 5% of the votes, and the slight variation after the 2008 elections. However, the volatility scores are much higher than those of the liberals. The results, again, may seem surprising, given that the conservative ideology and the membership in the European People’s Party have been among the main political targets on the political stage in the last decade. The high scores for the electoral volatility are explained by the very poor results of the conservative parties in three of the six rounds of elections. In May 1990 no conservative party gained over 5% of the votes. In 1992, the National Peasant Party – Christian-Democrat (PNTCD) became the main voice of the opposition, as part of CDR.

The electoral result of the PNTCD, still part of CDR and still the only conservative party in Parliament, doubled in 1996. Thus, the score of the conservative family went from 0 in 1990 to over 20% in 1996, hence the high volatility. However, the 1996-2000 cycle, when PNTCD led the government as part of a grand coalition, had unfortunate consequences for this party. In 2000, PNTCD decided to continue the CDR tradition, without having any significant ally. The alliance gained only slightly over 5% of the votes, insufficient for gaining seats. The conservative family continued to be represented in Parliament by the Romanian Humanist Party – Social-Liberal (PUR-SL), which gained representation for the first time as part of an alliance with the Party for Social-Democracy in Romania (PDSR, today PSD), but gaining just 1,41%. This party strengthened its position in the alliance with PSD in the 2004 elections, when it gained 5% of the mandates in the Chamber of Deputies, despite the fact that after the elections it left its allies and joined a right-wing government. In 2005 it changes its name in Conservative, thus clearly defining its ideological orientation. In 2008, they retied their alliance with PSD, but gained only 1% of the mandates.

However, in 2008 the Democrat-Liberal Party appears in elections for the first time as a conservative party. Formerly a social-democratic party, it changed its ideology in 2005-2006. Noticing the space left after PNTCD lost its electoral power, insufficiently represented by PC, PDL emerged as the main conservative party, gaining 32% of the votes and taking over government. Thus, as opposed to the liberals, always represented by PNL, by factions of PNL or parties that eventually merged with PNL, the Romanian conservatives were represented after 1990 by three different parties. One of them disappeared in 2000, another one gained seats only as part of alliances with the social-democrats and the third one claimed its conservative identity after 15 years as a social-democratic party. Hence, we notice the mechanism which explains why the effective number of parties inside the family is constantly 1, but the variations in the electoral scores of the family is very high.

The Social-Democrats
For the social-democratic family, we notice a relatively low electoral volatility, but also the highest effective number of parties inside the family among the three families included in the analysis. The only rounds of elections when the effective number of parties is 1 are the first, in 1990, and the last, in 2008. In 1990, social-democracy was represented by the National Salvation Front (FSN), who dominated the political scene. FSN was a conglomerate of political groups and personalities, some coming from the former Romanian Communist Party, while others were former anti-communist fighters. It is rather difficult to fit such a large political formation in a classical political family, however it is easier to do so given that FSN evolved following the 1990 elections in two openly social-democratic parties – PDSR and PD. The split was a consequence of the rift appeared between the president, Ion Iliescu, and the prime-minister, Petre Roman. Iliescu’s followers formed the Democratic National Salvation Front (FDSN), which became PDSR after the 1992 elections. Roman’s followers continued under the label of FSN and renamed the party PD after the 1992 elections. Another social-democratic party emerging as an important actor in 1992 was the Romanian Social-Democratic Party (PSDR), the successor of the inter-war, historical, social-democrats.

However, PSDR gained seats after running in elections as part of CDR together with parties coming from other political families, such as PNTCD and PNL. In the 1996 elections, the same three parties continued to represent social-democracy, but PD and PSDR ran together in an alliance and later joined CDR to form a governmental alliance. In 2000 and 2004 a fourth party considered here to be social-democratic had a very strong showing – PRM. Also, in 2000, PDSR and PSDR ran together in an alliance and later merged to form PSD. In 2004, PD ran together with PNL in an alliance which had as its main goal to remove PSD from power. In fact, what characterizes the relations between the parties forming the social-democratic family is the permanent state of conflict.

Since the split of FSN in 1991, PD and PDSR were in a permanent dispute for supremacy on the social-democratic lane of the party system. PD gained international acceptance, becoming a member of the Socialist International and of the Party of European Socialists before PDSR, which was regarded as the unreformed successor of the Communist Party. PRM has always been considered “the black sheep” of Romanian politics and, except a brief period before the 1996 elections, when PDSR used its support to maintain its position in the government, no party was ever willing to associate in any form of alliance with PRM. The competition within the social-democratic family was resolved after PD switched its ideological affiliation from social-democratic to conservative, thus leaving PSD to be the only social-democratic party winning seats in the 2008 elections. Thus, we notice that, with the exception of 1990 and 2008, the competition inside the social-democratic family was the highest among the three families, but the electorate was rather stable throughout time, despite the fact that the social-democrats suffered the biggest quantitative “loss”, with an entire party leaving the family.

The Liberals
The data presented above show that the liberals represent the most stable political family in Romania in the past 20 years. This conclusion may be paradoxical to those familiar with the troubled history of the liberals after the fall of communism. After its rebirth in the days of the 1989 revolution, PNL gained 7,32% in the first elections. The liberals showed they were the second political force, after FSN, the main actor of Romanian politics at that time. However, right after the 1990 elections, the liberals went through the first split. PNL – Young Wing (PNL-AT), let by prominent young liberals, was formed. When the Romanian Democratic Convention (CDR) was formed in 1991, PNL was one of the founders. Another liberal party part of CDR was the Party of Civic Alliance (PAC). Although it remained a parti of CDR for the local elections in 1992, it separated and ran on its own in the general elections later the same year, which led to the second split. A fraction decided to remain part of CDR and thus PNL – Democratic Convention (PNL-CD) was formed. PNL’s decision to leave CDR proved to be wrong, because it failed to gain any seats in 1992. However, the liberals continued to be represented in Parliament by the two liberal partis inside CDR – PAC and PNL-CD. Following the loss suffered in elections, the leadership of PNL changed and the former leader, Radu Campeanu, left the party and formed PNL-Campeanu (PNL-C). In 1993, PNL-AT becomes the Liberal Party ’93 (PL ’93). PNL returns in CDR in 1994 and runs as part of it, together with PNL-CD and the Romania’s Alternative Party (PAR, later called the Union of Democratic Forces), another newly formed liberal party, in the 1996 elections. In the same time, PL ’93 and PAC leave CDR and form their own alliance and fail to gain seats. They later form the Liberal Party (PL). In 1998, the liberal family begins to regroup. PNL absorbs PL and PNL-CD and merges with PAC. In the 2000 elections, PNL leaves CDR again, this time an inspired move, because CDR 2000 failed to gain any seats. In 2001, the liberals merge for the first time with a party coming from outside the liberal family, the Alliance for Romania (APR) being the result of a split in the main social-democratic party and failing to gain seats in the 2000 elections. In 2003, PNL merged with PAR and later the same year the liberal reunification was completed after the absorption of PNL-C.

In 2008, PNL absorbed two smaller parties: the Popular Action, founded by former president Emil Constantinescu, and the Democratic Force, a small social-democratic party, led by former prime-minister Petre Roman. However, there has been another split inside PNL in 2006-2007, when a large faction of the party left, formed the Liberal-Democrat Party (PLD), which later joined the PD and formed PDL . A quick look over this troubled history shows us that the political liberal family went through four splits and nine mergers. Despite these numbers, tables 1 and 2 reveal that the liberals had the most stable electorate throughout the last 20 years. The effective number of parties was constantly 1, because the failure of one party meant the success of other ones. Moreover, the liberals have the lowest level of electoral volatility. The highest variation is noticed after the 2004 elections, the first in which the liberals ran under only one party label, although as part of an alliance with the Democratic Party. However, the liberals managed to maintain their score in 2008 elections, when they ran on their own.

FULL ARTICLE AVAILABLE ON DEMAND