In those days, every vital vote in the Sejm [the lower house of the Polish legislature] [L.P.2] was a lottery. You could not even rely on intuition to predict the outcome. I remember sitting opposite the dashboard displaying the voting results and thinking will we lose or will we win?
This is how Hanna Suchocka remembers voting in the legislature on government proposals in 1993 when she served as the third prime minister since Poland’s democratic transition in 1989. A year later, after waiting in suspense for the results of a vote of no confidence proposed by members of parties making up the governing coalition, her cabinet fell. Suchocka recalls: ‘When the final tally was announced, it appeared that the number of aye and nay votes was precisely the same. Someone from the back benches of the Sejm exclaimed: “The cabinet has survived!”’ However, the speaker explained that for the no-confidence vote to fail, an additional vote was needed. Twenty-one years later, Suchocka remains sceptical as to whether his decision was as per the Sejm’s rules of procedure. She is certain, however, that no such doubts or uncertainties about the results of parliamentary voting vex Poland’s party leaders today.
She is quite right. The current government is led by the Law and Justice Party (PiS), which has been in power for the past four years and has a roll rate (the proportion of bills that the majority of the party supported but that lost on the floor) of almost zero. Its predecessor – Civic Platform (PO) – was in power for eight years and under its rule, no bill lacking a majority of the ruling party failed to pass. These two parties, while in government, have voted together on almost all bills considered in the legislature (Carroll and Nalepa 2019a). This is how a senior PO staffer, Agata Krośnicka, describes PO’s preparation for voting in the Sejm:
The decision on how to vote is made by the caucus Presidium, which is a smaller body than the entire caucus: the [caucus] chairman, deputy chairpersons, secretary, treasurer, and one MP representing each region, as well as three to four senators. This body decides whether discipline should be extended over a particular vote. It is not uncommon for MPs to have different ideas about bills we are working on to maintain political continuity of our policy programme; we have to introduce discipline. Discipline is rarely broken, but when it is, the party sanctions insubordination with anything from ‘reprimands from the party chairman’ to financial penalties for breaking discipline. The exact penalties are determined ex-post by an inner circle made up of Presidium members: the chairman, his deputy, the secretary and the treasurer of the caucus. Although the sanctioned MP can appeal the decision, the appeals are resolved by the very same leadership circle. A typical financial sanction is 1,000 zloty, [L.P.3] with each case considered individually in the context of an MP’s overall activity. An MP’s absence during a voting session is also considered a violation of discipline, as a ‘technical failure to vote properly’. (interview with Krosnicka)
In the second part of the interview, Krośnicka revealed how the caucus compiles statistical data: ‘So that we know who and how often votes diﬀerently from the caucus.’ She and other staffers insisted that other parties prepare for votes in similar ways.
The glue that keeps Polish parties voting together is more than sheer discipline. A revealing interview with the PiS whip (referred to as the ‘ombudsman for party discipline’), Wiesław Tchórzewski, clarifies how frequently voting discipline is used:
Only in 5 per cent of cases does the Ombudsman mandate discipline and it is typically just an eﬃciency measure adopted when there is not enough time to hold a caucus meeting. Given limited plenary time and the sheer volume of bills, the so-called ‘bill coordinator’ prepares a cheat sheet instructing MPs how, in the party experts’ opinion, they should vote. Ninety-five per cent of the time, the coordination of the cheat sheet is suﬃcient to ensure unity. Discipline is necessary for compliance the remaining 5 per cent of the time. (interview with Tchorzewski)
Later, in the same interview, Tchorzewski explained that ‘agreeing with the party 60 per cent of the time is enough for him to vote with it 100 per cent of the time’.
For two decades, Polish parties have attained record-breaking levels of voting unity thanks not only to discipline but also ideological cohesion. Moreover, as the example of PO illustrates, even when discipline is used, it is employed for the sake of long-term programmatic consistency. Documents of the post-communist Democratic Left Alliance (SLD) – the third largest party at the time – showed that it kept tabs on its members and distributed parliamentary and cabinet perks according to the loyalty rankings. Among other documents, I discovered an internal document explaining sanctions meted out to two MPs who violated voting discipline. Their violations consisted not of casting a ‘nay’ vote, but merely unexcused absences from the vote.
The aim of this paper is to explain how Poland’s party system and institutions regulating legislative–executive relations evolved from a chaotic, yet individually representative, system with abundant opportunities for opposition voices to a system dominated by a consolidated majority, where the opposition has but token opportunities to influence the political process and instead must resort to extra-institutional forms of dissent, such as street protests.
Parties ascendant: electoral vehicles to representative agents
Political parties evolve in almost every political system to overcome the collective action problems of legislatures, to form executives and to mobilize voters in elections, but for parties to be agents of representation or to be programmatic, they must form a coherent perspective on public policy that voters can use in their decision-making. Otherwise, parties – even those that appear strong – will be relegated primarily to organizational functions. They will become merely vehicles of electoral politics. Assuming that parties are necessary to democracy, what determines the kind of parties that emerge in post-communist-rule Europe such as Poland?
One of the trade-offs facing party leaders in new democracies is between the long-term goal of building a party brand and winning enough seats to create a government (Müller and Strøm 1999). In other words, leaders can either recruit members to build cohesive parties capable of taking consistent policy positions over the long term or they can create different parties that are large enough to win power but do little beyond that. Leaders of this latter type of parties, which recruit members purely to win more votes, may end up having trouble controlling the voting behaviour of their diverse membership, while parties focused on developing their long-term policy reputations in parliament – that is, electoral ‘brands’ – should avoid these costs of reining in a diverse membership by concentrating on recruiting members whose preferences are aligned with those of their leadership.
The success of such branding in the aftermath of the democratic transition relies on two critical factors, which are also the exogenous variables in our theory:
- The presence of ideological cleavages
- The nature of the electoral system
First, the electoral value of developing brands is diminished in societies that lack evident ideological cleavages, either because authoritarian legacies or ethnic identities are too prominent. In the absence of ideological cleavages, parties can more easily form linkages based on pre-existing identities, whether these are ethnically based or formed by an affiliation with the ruling versus dissident part of the authoritarian society. The second source of effective brand building is the electoral system, which regulates the costs of controlling members of parliament. In candidate-centered systems – systems where voters vote for individual candidates instead of or in addition to parties – MPs can use their reputations to win party votes.
I argue that in candidate-centered contexts dominated by non-ideological cleavages, the high cost of controlling members puts pressure on parties to develop ideologically cohesive organizations. Meanwhile, party-centered systems, that is systems that allow leaders to control members more efficiently, reduce incentives to organize ideologically cohesive parties.
These tendencies inhibit the representation of ideological cleavages even after the prominence of post-transition cleavages (based on ethnicity and the regime divide) declines (Grzymala-Busse 2001).
I will use these insights to shed light on the development of the Polish party system. Critical for my understanding is the value of voting unity and the pathways to attaining it.
Voting unity: two conceptions
Voting unity is the extent to which parties vote together on issues placed on the parliamentary agenda. But voting unity also has intrinsic value. Parties may value voting unity because it allows them to present themselves to voters as ideologically coherent groups that can be trusted to pursue particular sets of policies. For governing parties, unity leads to legislative success, which in turn produces a record of achievement and allows voters to hold parties in government accountable. For parties in the opposition, the intrinsic value of unity is particularly stark. Since legislative success is out of reach – due to small numbers and a lack of control over the agenda – opposition parties promote their brands by demonstrating to voters that they vote in concert on policy positions that are consistent over time. Voting unity allows voters to attribute party labels not only to programs but also to actions consistent with those electoral promises.
The distinction between instrumental and intrinsic unity helps us understand the seemingly irrational behaviour of parties in post-communist Europe, such as the decision of Poland’s SLD leadership to indefinitely suspend the party membership of MPs who were habitual discipline violators.
2[CBJ4] The SLD caucus formalized this decision by including a section on disciplinary sanctions towards maverick MPs who ignored caucus voting decisions. The most severe sanction of last resort was expulsion from the caucus. Why would leaders take actions that reduce their legislative seat share? Consider the following comments made by the SLD chairman following a vote on the selection of two justices of the Constitutional Tribunal that the SLD leadership decided to boycott: ‘The participation in this vote, against earlier caucus agreements, indicates that some MPs interpret decisions of the caucus too liberally, risking compromising the reputation of the party not only on the Sejm floor but also in the eyes of the broader public.’ The participation of the few SLD MPs did not make any difference to the outcome of the vote in question. The value in voting (or rather, non-voting) unity was purely intrinsic. The intrinsic value of voting unity is also the most plausible reason why caucus meetings whip more votes than is necessary to pass their preferred legislation or ensure the failure of legislation they wish to avoid.
Another dilemma facing leaders preparing their parties for elections to the assembly is between recruitment aimed at maximizing legislative seat share and recruitment aimed at maximizing the realization of policy goals. The former strategy ostensibly aims to gain executive control. Parties can maximize their odds of winning government offices by recruiting politicians with clout, such as popular local and national notables. Recruiting such members carries risks however; once in the legislature, they may prove hard to control by party leaders.
The second strategy involves parties recruiting for policy-seeking. In this latter case, leaders need voting unity to pursue their goals. Note that policy-seeking itself can be oriented to the short or long term. Short-term policy-seeking is focused on legislative achievements (such as passing bills). This calls for voting unity, but of the instrumental type discussed above. In contrast, long-term policy-seeking goes hand in hand with brand building, for which the value of unity is intrinsic.
Considering voting unity helps us understand why some parties, but not all, form with policy-seeking goals in mind. When ideological cleavages are too weak for parties to organize around them, and non-ideological ones such as those based on ethnicity or authoritarian legacies are strong, using the party organization as an electoral vehicle may be more advantageous than building programmatic parties. Under such circumstances, a focus on personalities and short-term performance tends to dominate the electoral process. There is also an informational advantage in focusing on such non-policy related issues. Namely, in the early stages of the transition, it can be difficult to predict what the dominant cleavages in society will be. In electoral systems that incorporate elements of the personal vote (that is, allow for the possibility of voting for specific persons, not just parties), creating parties made up of independently popular politicians may be a new party’s swiftest strategy for boosting seat share. This strategy may backfire if the independently popular MPs refuse to vote with the party leadership, at the expense of voting unity.
For this reason, candidate-centered systems – that is, ones that allow for the personal vote, seem to perpetuate non-ideological cleavages and have been considered by some scholars to impede programmatic party development (Kitschelt 2000; Kitschelt et al. 1999). I argue, however, that this is a transient feature of candidate-centered systems. While legacies of the regime continue to taint legislative achievements, eventually they give way to genuine substantive cleavages in society. Moreover, once ideological cleavages have stabilized, personal-vote electoral systems begin to curb incentives for the recruitment of powerful and independent notables who stifle the prospects for the development of programmatic parties (Carroll and Nalepa 2019b). This happens because getting members with clout to toe the party line is so costly that it forces leaders to recruit for ideological cohesion if they want to attain voting unity. Whether parties value unity for its instrumental significance in producing short-term policy success or for its intrinsic worth associated with long-term brand building, they must adjust their recruitment strategy accordingly. While leaders in party-centered electoral systems can afford to ignore the ideological diversity of their membership and still vote in unison (because discipline is cheap), leaders in candidate-centered systems cannot afford to. In short, the personal vote forces parties into ideological cohesion.
Figure 1: Three pathways to voting unity and the utility of unity (figure created by author)
Figure 1 shows the three pathways to voting unity: (1) ideological cohesion, which calls for uniform preferences of party members; (2) discipline, which requires access to resources to make party membership valuable; and finally, (3) agenda setting. I elaborate on these distinctions in the remainder of this section.
Pathway to voting unity I: cohesion
The first pathway to voting unity is through ideological cohesion. This pathway is best understood by contrasting it with the discipline pathway, particularly since strategies of recruitment influence both. The following quote from Hanna Suchocka illustrates that a party composed of like-minded members will vote in unison regardless of how well the party organization functions:
In the very beginning, as part of the Civic Committee (KO), we voted together because we believed in the idea of reform. We shared the same goals and were convinced that we needed to change the system. The ‘contractual’ Sejm of 1989 played a more important role than any other Polish assembly. It instituted more changes because of the KO’s shared convictions and psychological momentum. Confronted with KO’s cohesion, the former communists could not even muster the courage to vote against the reforms we were proposing, even though we only held 35 per cent of the seats and our proposals were very far-reaching.
The voting unity accompanying regime change was driven, however, only by a desire to pass reforms making regime change irreversible. Once this legislation was in place, ideological divisions within the KO became apparent as the caucus fragmented into a multiplicity of parliamentary ‘circles’. By the second parliamentary term, the legislature had descended into the chaotic voting patterns described in the opening quote to this paper.
The recruitment for KO well illustrates the origins of parties with cohesive potential. Namely, to produce a party composed of like-minded politicians, leaders have to select members based on their ideological attachments. In the early aftermath of transition, illustrated by the case of KO, these attachments were based on anti-communism and support for regime-changing reforms. In contrast, when leaders recruit members whose loyalty can be bought by perks of office, they are using discipline to produce unity.
Note that even as parties try to recruit for ideological cohesion, institutionalizing this requires coherence in voters’ policy preferences. Voters in new democracies may lack experience with acquiring political knowledge, and exogenous sources of uncertainty associated with the transition might cause these preferences to shift a great deal. However, as long as independent-minded MPs are in the party’s minority, a well-organized party can achieve with discipline what it lacks in cohesion, which leads us to the second pathway through which parties can attain voting unity.
Pathway to voting unity II: discipline
What a party lacks in ideological cohesion can be corrected with a high degree of party discipline, provided the party has the requisite organizational strength. Party discipline, explicitly distinguishing discipline from ideological cohesion, is the subject of Carroll and Nalepa (2019b). In open-list proportional representation systems (OLPR), that is, systems where in addition to casting ballots for parties, voters select specific party members, MPs have incentives to appeal both to their leaders (who control candidate nominations) and voters (whose votes determine which candidates win seats). Carroll and Nalepa propose a game-theoretic model that highlights the mediating influence of legislators’ preferences and popularity with voters on party leaders’ ability to enforce discipline (2019a). Conditional on how MPs’ preferences diverge from those of the party leadership, variation in compliance depends on parties’ organizational costs and resources for enforcing discipline. The model provides practical implications at the party level – for differences in voting unity – and at the individual level – for differences in voting loyalty. Most importantly, this model also explains the large-scale changes in the pathways to party unity across time as parties institutionalize their organization outside of parliament.
Carroll and Nalepa (2019a) use a formal model to assess the role played by control over rank in channelling incentives for party organization. Electoral systems differ in how much information about a party member’s clout is available to the party leadership as well as to the member himself. Under OLPR, it is transparent precisely how many votes each member brings to the party list. This information can give members leverage against the leadership in asking for exceptions to disciplined votes. In closed-list proportional representation systems (CLPR) – systems where voters cast ballots for parties and not specific candidates (party leaders, in turn, determine the order in which their members gain seats) – although the gains to the party leadership from putting popular members on the list are proportional to their popularity, there is no way to measure accurately how much of the list’s vote share can be attributed to any specific member’s popularity. Since his or her contribution to the list is obscured, no member can use it as leverage against party discipline. We show that because discipline is costlier in OLPR if leaders value voting unity (as we argue they do). They are forced to recruit for cohesion. Meanwhile, in CLPR, leaders can achieve the same level of voting unity by relying on sheer discipline. To the extent that programmatic parties are more likely to form based on cohesive parties, OLPR offers better prospects for programmatic party development than CLPR.
Carroll and Nalepa also use Polish data on legislation co-sponsorship to measure party members’ preferences. They demonstrate that shared preferences are a sufficient but unnecessary condition for parties to vote in unison. Parties with strong, if temporary, organizational advantages can use discipline to overcome even severe divisions. This is a novel approach because until recently, scholars have often used the concept of discipline interchangeably with cohesion, suggesting that they are observationally equivalent (Kam 2014). Carroll and Nalepa propose a way of operationalizing this difference with measures that do not always go together (Carroll and Nalepa 2019a). They argue that parties have a harder time disciplining members who are independently popular and that, regardless of party organization, parties that are about to lose viability will find it difficult to get their members to toe the party line.
Producing unity via discipline alone requires that leaders maintain party viability. A party that recruits opportunists and appeals to voters by adjusting its policy platform accordingly can rely on discipline to produce unity only in times of electoral success – when it captures voter sentiment and obtains access to power. Without that success, its members will flee to more viable parties. In short, in order to rely on discipline as opposed to cohesion, leaders must be able to correctly anticipate changing voter preferences and be able to instantly adjust policy to the platform that will appeal to voters. Unsurprisingly, a tension arises between the opportunism seen in emerging democracies and the role of parties as representative agents (programmatic parties). Because they do not have precise information about prominent policy cleavages that will become stable, parties are initially easily reduced mainly to their electoral role as vehicles for electoral success. As cleavages based on policy conflict emerge and stabilize, it becomes possible for parties to not only win elections and establish party records in government but also create brand names that are resilient enough to meet prospective demands even while these parties are out of power.
Pathway to voting unity III: agenda setting
Disciplined parties are capable of more than merely effecting policy change. Once large enough – or in cooperation with other unified parties with similar aims – they can change the institutional organization of the legislature (Nalepa 2016 and 2017). A final pathway to voting unity is, therefore, agenda setting: those that agree on policies they want to avoid can delegate power to agenda setters and limit prospects for the division. While these practices increase voting unity, partisan representation implies entirely subordinating party membership to a party organization’s long-term programmatic goals. This means institutionalizing the recruitment of members who support party programmes when discipline is absent. To reconcile the programmatic goals of parties with the policy preferences of their members in the long term, programmatic party appeals must converge with the dominant cleavages in society.
One of the most significant changes that a party controlling a parliamentary majority can implement is modifying the rules of parliamentary decision-making from open to closed. Under open rules, any MP can propose bills for consideration. Open rules are consistent with extensive rights granted to individual MPs as opposed to party caucuses. Under closed rules, proposal power is restricted to the speaker, a special committee or another small subset of all members of the legislature. Nalepa uses a series of simple game-theoretic models to show that amending rules of the legislature from open to closed gives governing parties an opportunity to keep off the agenda issues they do not want to see passed and helps change existing policies into ones that are more to their liking (Nalepa 2017). Furthermore, governing parties can maintain this negative agenda control even if they lack cohesion or lose their majority coalition in the legislature, provided they have a trusted speaker who is capable of setting the agenda in this fashion. In other words, disciplining members is not the only way to ensure voting unity in parties that lack comprehensive cohesion on all policy dimensions that could potentially be considered in the assembly. Parties may still vote in unity as long as they have some ideological cohesion and the items placed on the agenda are restricted to the dimensions on which members share preferences. When a governing party can delegate agenda-setting powers to a trusted agent, such as the speaker of the House, that agent can ensure that issues that could splinter the party vote never appear on the agenda. Such control over plenary time relies on a high degree of party organization, which is then translated into the legislative organization.
This use of agenda control was first described by US Congressional scholars (and later also beyond the US context) (Cox et al. 2001, Cox and McCubbins 2002; Cox et al. 2008). As my fieldwork indicates, this strategy was taken up by quite a few Polish speakers. According to Marek Borowski, who served as Sejm speaker between 2002 and 2004, ‘Beginning in 1998, the speaker pretty much holds a veto over the agenda. He is the gatekeeper. If he decides to go forward with an agenda, MPs may protest, and he has to consider their opposition, but as a gatekeeper he can delay considering legislation for up to six months or even longer. Formally, the speaker has less power than he has in reality. In reality, he can skew decisions in his favour.’ Marek Jurek, who served as speaker under the Law and Justice (PiS) government was described by Józef Oleksy as a ‘true emissary of the PiS’. Under his speakership term, journalists and pundits coined the term ‘zamrażarka’ (freezer) to describe the speaker’s power to ‘shelve a proposal completely and not do anything with it’. Another former speaker interviewed for this project said, ‘There is nothing that can force the speaker to put an item on the agenda for a first reading. This is why ruling parties typically reserve this position for one of their most loyal members’.
The agenda-setting pathway to voting unity is only available to parties in the government and also presupposes majoritarianism in executive–legislative relations. It is easier to control the flow of bills by a single actor, such as the senior party in the cabinet coalition. That is why, for agenda control to be a feasible pathway to unity, the system must already have made the shift from one of legislative dominance to one of executive dominance, which I describe in the next section. To facilitate such change, at least the parties in the majority must be organized well enough to implement it. A party that has reached an agreement to delegate agenda control to a trusted member once in government can obtain through legislative organization and agenda setting what it lacked in cohesion and discipline.
From party institutionalization to legislative dominance
As illustrated in Figure 2, success in agenda setting hinges on the legislative institutions, but underscoring every pathway – discipline and agenda setting in particular – is party organization or a party’s degree of ‘institutionalization’. I call a party ‘institutionalized’, when it has developed long-term concerns about building a reputation with voters and with other parties, consistent programmatic policy goals and the organizational capacity to maintain a stable majority coalition. I believe that this understanding adapts to the level of a single party the criteria developed by Mainwaring and Scully (1995) to describe the institutionalization of party systems: stability in interparty competition, roots in society, electoral legitimacy and overcoming personalism.
Party institutionalization affects recruitment for cohesion, affects discipline enforcement, and – at the level of the party system – makes agenda setting possible. The degree to which parties become institutionalized can mitigate the trade-offs faced by parties that are too ideologically anchored or too opportunistically oriented. Nalepa shows that despite strong incentives for disloyalty among high vote-getters in candidate-centered electoral systems, institutionalized parties can prevent disunity by recruiting candidates that share the leaders’ views and are accountable to constituencies (Nalepa 2017).
This is easier to achieve as parties develop clearer reputations, focus more on long-term goals and formalize their recruitment organizations. In the absence of recruitment for ideological cohesion, as parties increase in size, developing clear reputations becomes more difficult. Political parties are ubiquitous in electoral competition. They evolve in almost every political system to overcome the collective action problems of legislatures, to form executives and to mobilize voters in elections. Whether they are weak or strong, political parties and democracy go hand in hand. However, the mere presence of parties does not mean they are the primary means of democratic representation. For parties to be agents of representation, they must form a coherent perspective on public policy that voters can use in their decision making. Otherwise, parties – even those that appear strong – will be relegated primarily to organizational functions. They will become vehicles of electoral or legislative politics, but only individual politicians will determine how democratic demands for representation are fulfilled. In short, the existence of parties in newly democratic legislatures – almost always a natural by-product of the electoral process – is not a sufficient condition for parties to serve as vehicles of democratic representation.
At the same time, legislatures empowering individual members to initiate laws on behalf of their electors, who return them to office as long as they serve their voters’ interests, lie at the heart of the democratic ideal of representation. Even if parties do not mediate this relationship, the ideals of both accountability and representation may still be fulfilled.
In the context of parliamentary democracies, representation without parties seems contradictory because, in Western Europe, parliaments and highly representative party systems developed concurrently. Thus, the relationship between strong parliamentary systems and disciplined or cohesive parties is largely taken for granted, even today. If the universe of cases is expanded to include Eastern Europe, the contrast between representation through parties and individual members of parliament (MPs) can cut across rather than overlap with the distinction between executive and legislative dominance. Figure 2 illustrates this idea:
|Consensus democracy (e.g. Netherlands, Belgium)
|Majoritarian democracy (e.g. modern UK)
|Empowered individual legislators (e.g. US Congress, Victorian England)
|Executive–legislative fusion (e.g. LDP-era Japan, 20th-century Colombia)
Figure 2: Executive–legislative power in partisan and non-partisan modes of representation (figure created by author)
The top two cells of Figure 2 show the two possible effects of resolving the executive–legislative balance in an environment in which parties are the primary vehicle of representation. On the left is what Arend Lijphart and scholars of comparative politics call the ‘consensus’ model of government. In this model, decisions are reached through considerable bargaining involving many or even most legislative parties. The bargaining occurs in the government formation process as well as in the legislative process within governing periods. The consensus model is considered to be broadly representative of a diverse array of groups in society. This model is often thought to sacrifice accountability because voters are less sure which party is responsible for which policy. In this partisan mode of representation, if the balance of power were to tip in favour of the executive, a majoritarian model emerges. This model, shown in the top right cell of Figure 2, is potentially threatening to the idea of broad representation. This is because when a strong majority party in government takes control over legislative activity by staffing all key legislative positions with its trusted agents, supporters of parties remaining in the opposition are left with little representation. As a result, this governing party is the focal point of democratic accountability.
The bottom part of Figure 2 relaxes the conventional assumption that parties are the main vehicles of representation. Instead, it shows two alternative possibilities for resolving the legislative–executive balance in an environment where representation takes place through individual legislators. In the lower left, the balance again favours legislatures, but now in a context in which parties are subordinated to the elected politicians with whom they affiliate. As in the consensus model, the executive has limited power. However, individual legislators are not only strong but also maintain tight connections with their regional electorates. Such legislators may be elected in single-member districts – a system that predates modern parties and allows candidates to be relatively independent. However, unlike in the Westminster model, the role of parties in the candidate nomination process is minimal. Hence, legislators can remain independent of their party labels. The lower right cell also represents an individualized environment, but one in which the executive plays a prominent role, analogous to the majoritarian model described above. In some cases, this dynamic emerges because the individual legislators are overshadowed by a separately elected and constitutionally powerful president. This results in executives dominating the legislative process and interacting with individual legislators that are focused more on representing local constituencies and interests.
The trade-oﬀ between representation through parties and representation through individuals can be summarized as follows. Individuals ensure that local interests are strongly represented and that candidates are held accountable for their performance. However, if MPs are responsible for getting re-elected, those MPs attract local electorates by representing their particularistic demands (government actions beneficial only to a geographically exclusive group of citizens). To enact such policies, legislators must ‘logroll’ (exchange favours) in the assembly to procure goods for their own constituents. Such narrow relationships can produce global inefficiency by sacrificing broader goals associated with policies and programmes that are desirable nationally.
Parties play contrasting roles in the two distinct modes of representation: in the top part of Figure 2 they are exogenous to individual politicians, and in the lower part they are endogenous to them. When parties are exogenous, members are compelled to join pre-existing party organizations in order to take advantage of established programmatic brands. In the endogenous case, individual members shape and form parties with a purpose in mind that is a function of their personal goals.
Parties in transitional environments
Where do Eastern European parties and legislatures fit in the framework presented above? The answer to this question allows us to explain the transformation that has occurred in legislative–executive relations and how this transformation has gone hand-in-hand with changes in party institutionalization. I answer this question with a particular focus on Poland. Post-communist Europe presents us with transitional systems that make it particularly difficult for parties to emerge as representative agents. Rapid transitions to party representation are more difficult to complete if the key decision makers who design the constitutional regimes are hostile to having parties serve as vehicles of representation and if the primary sources of political identity are rooted in the transition itself. This was precisely what happened with new post-authoritarian states in Central and Eastern Europe.
This was prevalent in many (if not most) post-communist countries after the Soviet bloc collapsed. In light of the monopoly that the Soviet-style communist parties had held over political institutions after communism fell, East Europeans were wary of political parties. Thus, with varying degrees of success, designers of the first institutions tried to delay the emergence of parties. At the same time, reformist legislatures across the region had played a trailblazing role in the transitions by, for instance, allowing anti-authoritarian dissidents to assume legislative seats.
Norton and Olson (2013) note that ‘although constitutions and party systems are the logical initial steps to understanding a stable political system, they were not the first features to emerge at the end of the authoritarian rule [in post-communist Europe]. Parliamentary elections came first, and the widespread scepticism toward political parties shaped the first parliamentary institutions.’ Many of the first post-communist legislatures were elected through systems that reward personal characteristics of MPs rather than their attachments to party labels. So these first parliaments created rules of procedure that extended rights to individual MPs. Among these institutions were electoral systems containing elements of the personal vote, directly elected presidents and decentralized legislative institutions. Eventually, once they were created, many of the new constitutions extended considerable powers to legislatures constraining the roles of cabinets and, indirectly, parties.
Poland contains a range of formal institutions that can potentially undermine the incentives to form strongly organized parties, including a legislature that empowered individual MPs over parties and empowered the legislature relative to the executive. Despite emerging from 50 years of authoritarian rule, the opposition had an important voice in negotiating the terms of the transition. But both the former dissidents and the outgoing communists opposed an electoral system that would enable party representation. Solidarity’s dissident organization, crippled by the Martial Law crackdown, lagged far behind the communists in party organization. A proportional electoral system adopted for the first election could have dramatically reduced their parliamentary representation following the first elections in 1989. Likewise, the communists opposed any electoral system based on party lists (such as PR) because it would have forced them to recognize Solidarity as an official party. As a result, a first-past-the-post electoral law with single-member districts became the consensus choice in the first elections in which non-communist candidates were allowed to run.
Although an electoral system with similar or even more personal vote potential would have been first past the post with single-member districts, this would have resulted in regional representation. In a context where the primary cleavage – the communist government versus opposition – did not coincide with geography, at least one of the players negotiating the terms of the first election should have vetoed first past the post as a permanent electoral solution. Although it cost the communists executive power in the short run, this move offered them an organizational advantage over competing parties that had yet to be created. Thus, the transitional election strategy preferred by all major players counteracted the emergence of centralized parties, mainly based on anti-communist dissident groups.
Going beyond the electoral law, the standing rules of the Sejm strongly favour the rights of the individual MPs and small factions by offering them opportunities to introduce bills, powers to amend legislation and influence in committees. Because these institutions typically have the effect of granting MPs from the opposition opportunities to influence the legislative process, the early scholarship of post-communist politics characterized Poland’s legislature as particularly strong vis-à-vis its executive and the political parties as weak and underinstitutionalized.
In sum, Poland contains a range of formal institutions that could potentially undermine the incentives to form strongly organized parties, including a legislature that empowered individual MPs over parties and empowered the legislature relative to the executive. Despite emerging from 50 years of authoritarian rule, the opposition had an important voice in negotiating the terms of the transition. Both the former dissidents and the outgoing communists opposed an electoral system that would enable party representation.
Tracing the evolution of legislative institutions leads to the discovery that Poland’s parliament has become transformed: the broad rights that had initially been afforded to individual MPs were later limited as parties in the government established control over legislative institutions and converted the legislature into an extension of their majoritarian dominance. Parties have become the dominant vehicle of not only electoral but also legislative politics in a way that closely resembles cartel models of legislatures. How did Poland’s system move so quickly from the mode of representation based on individual MPs and strong legislatures to a system favouring party representation and strong executives? My explanation hinges on how long-standing deep political cleavages have interacted with subtle institutional shifts that took place over the last two decades in Poland, most notably in executive–legislative relations, which recently allowed the ruling party to implement sweeping illiberal reforms rapidly. These institutional factors can be summarized as follows: 1) recorded voting has allowed leaders to exercise party discipline on the floor of the legislature and effectively put an end to free mandates; 2) the open-list proportional representation electoral system gave parties an incentive to emphasize ideological cohesion as they recruited candidates, contributing to ‘strong’ parties that can sustain temporary swings in voter volatility; 3) reforms to standing rules of the Sejm have concentrated agenda-setting power in the hands of a representative of the ruling party, allowing whoever is in government to use the legislature as an extension of its executive power.
This has allowed the Polish system to shift from the bottom right cell to the top right cell of Figure 2. With these institutions in place, it is practically impossible for the opposition to have any influence on the legislative process. In light of this shift toward majoritiarian institutions, centralized parties and the polarized context in which they emerged, what we are observing in Poland is not surprising. The institutions allowing for such rapid recent authoritarian backsliding have been in place since 2007 and were exploited by the previously ruling PO, who happened to use this power for a liberal agenda. In short, institutions put in place as the party system institutionalized failed to provide adequate representation for the forces behind long-standing political divides.
The evolution of the Polish party system has gone hand-in-hand with the transformation of Poland’s parliament: the full rights that had been initially afforded individual MPs were later limited as parties in the government established control over legislative institutions and converted the legislature into an extension of their majoritarian dominance. Parties have become the dominant vehicle of not only electoral but also legislative politics in a way that closely resembles cartel models of legislatures. In terms of the framework presented in Figure 2, why did Poland’s system move from the mode of representation based on individual MPs and strong legislatures to a system with party representation and strong executives so quickly? Why have other post-communist nations and other emerging democracies taken different paths? This puzzle can only be addressed by careful analysis of empirical data from legislative roll calls, electoral returns and voter surveys. The theory presented here suggests that the combination of an electoral system incentivising ideological recruitment, the dissipation of regime divide cleavages and a majoritarian shift in legislative institutions all led to the erosion of the individual mandate and gradually shut the opposition out of influence over democratic politics.
 1,000 zloty is the equivalent of about $330.
 Archival research for this project was conducted at the Polish Sejm in Warsaw between 2009 and 2011.
 Chairman Leszek Miller, according to a transcript from an SLD meeting on 27 November 2001.
 The lower house of the Polish assembly in 1989–91 was referred to as the ‘contractual Sejm’ to reflect that its composition – non-communist candidates were allowed to compete for only 35% of seats – was part of the Round Table agreement.
 See Nalepa (2017) for more details on changes in the legislative organization that took place in the Sejm to allow for such inflation of speaker powers.
 The US Congress at the low points of partisan polarization is the classic example of striking the executive–legislative balance this way in the context of individual representation. Another historical example comes from the so-called golden age of the private MP, described by Cox (1986).
 An example of this is Japan under the long reign of the Liberal Democratic Party. Membership was a prerequisite to obtaining legislative power, which was centralized in the cabinet. However, MPs had to cater to their constituents to maintain office individually. That is, it was individual legislators who served as vehicles of representation, rather than the party, which was an umbrella organization incapable of offering a consistent policy brand.
 Examples where a powerful president entirely substitutes the role of parties include Latin American systems such as historical Brazil and Colombia.
Carroll, Royce and Monika Nalepa (2019a) ‘When Do Open Lists Matter? The Consequences of the Personal Vote for Party Loyalty’. Parliamentary Affairs.
— (2019b) ‘The Personal Vote and Party Cohesion: Modeling the Eﬀects of Electoral Rules on Intraparty Politics’. Journal of Theoretical Politics. 32 (1), 36–69.
Cox, Gary W., William B. Heller and Mathew D. McCubbins (2008) ‘Agenda Power in the Italian Chamber of Deputies, 1988–2000,’ Legislative Studies Quarterly 33 (2), 171–98.
Cox, Gary W., Mikitaka Masuyama and Mathew D. McCubbins (2001) ‘Agenda Power in the Japanese House of Representatives’. Japanese Journal of Political Science 1 (1), 1–21.
Cox, Gary W. and Mathew D. McCubbins (2002) Agenda Power in the US House of Representatives, 1877 to 1986. New York, NY, Cambridge University Press.
Grzymala-Busse, Anna (2001) ‘Coalition formation and the regime divide in new democracies: East Central Europe’. Comparative Politics 85–104.
Kam, Christopher (2014) ‘Party Discipline’. The Oxford Handbook of Legislative Studies,
Kitschelt, Herbert (2000) ‘Linkages between Citizens and Politicians in Democratic Polities’. Comparative Political Studies 33 (6–7), 845–79.
Kitschelt, Herbert, Zdenka Mansfeldova, Radosław Markowski and Gabor Toka (1999) Post-Communist Party Systems: Competition, Representation, and Inter-Party Cooperation. Cambridge: Cambridge University Press, 1999.
Mainwaring, Scott and Timothy Scully (1995) Building Democratic Institutions: Party Systems in Latin America. Stanford, CA: Stanford University Press.
Müller, Wolfgang C. and Kaare Strøm (1999) Policy, Oﬃce, or Votes? How Political Parties in Western Europe Make Hard Decisions. Cambridge: Cambridge University Press.
Nalepa, Monika (2016) ‘Party Institutionalization and Legislative Organization: The Evolution of Agenda Power in the Polish Parliament’. Comparative Politics 48 (3), 353–72.
— (2017) ‘Adapting Legislative Agenda Setting Models to Parliamentary Regimes: Evidence from the Polish Parliament’. Studies in Logic, Grammar, and Rhetoric 50 (1), 181–203.
Norton, Philip and David M. Olson (2013) The New Parliaments of Central and Eastern Europe. London: Routledge.
List of interviews (full texts are on file with the author)
- Hanna Suchocka, interviewed by Monika Nalepa in Chicago, 30 October 2014.
- Marek Borowski, interviewed by Monika Nalepa in Warsaw, 8 November 2009.
- Józef Oleksy, interviewed by Monika Nalepa in Warsaw, 23 May 2011.
- Włodzimierz Cimoszewicz, interviewed by Monika Nalepa in Warsaw, 10 June 2011.
- Agata Krośnicka, interviewed by Monika Nalepa in Warsaw, November 2011.
- Wiesław Tchórzewski, interviewed by Monika Nalepa in Warsaw, November 2011
The author is grateful to Royce Carroll for extensive help in writing this paper and to Irina Khmelko and Michael Mezey for comments. All mistakes are the author’s responsibility.
Abstract: This paper focuses on the case of Poland, which has joined a growing number of countries plagued by democratic backsliding as opposition parties have failed to challenge the ruling party’s infringement on the rule of law. Opposition parties, though allowed to run in elections and attend the legislature, have been gradually shut out of any institutionalized forms of protest. Pundits, policy makers and even scholars have been quick to blame this state of affairs on the weakness of formal institutions. This paper takes a different view and argues that institutional features – such as agenda control and party discipline alongside legislative institutionalization – have been blocking the opposition from effectively influencing policy. Moreover, these majoritarian institutions that have shifted executive–legislative relations in the executive’s favour are not the doing of the new populist government but have been in place for at least a decade. It is, therefore, too strong rather than not-strong-enough institutions – particularly those governing the executive–legislative relations – that have contributed to the recent phenomenon of democratic backsliding in this country.