Introduction
Party pledges have a crucial role in the party mandate fulfilment process. First, parties formulate them in manifestos because they hope to attract voters with them. Second, after the elections, the parties take the pledges as the policy base for bargaining on the coalition agreements’ content. Third, the whole coalition is responsible for the real fulfilment of preferably those pledges adopted into agreements. Finally, voters are expected to evaluate pledge fulfilment and vote according to their evaluation (APSA 1950; Downs Reference Downs1957; Klingemann et al. Reference Klingemann, Hofferbert and Budge1994; Royed Reference Royed1996; Thomson Reference Thomson2001). This article focuses on the intermediary part of the process, which is almost unattended in party mandate fulfilment research, i.e. adopting pledges into coalition agreements. Coalition agreements are usually seen as to-do lists of to-be coalition actions (Timmermans Reference Timmermans2003; Quinn Reference Quinn2014; Müller and Strøm Reference Müller, Strøm, Strøm, Müller and Bergman2010). It is likely that each party bargaining for coalition agreement aims to be electorally rewarded and should try to adopt as many of its narrow pledges as possible in the emerging document (Schermann and Ennser-Jedenastik Reference Schermann and Ennser-Jedenastik2014a). The success of individual coalition parties is not guaranteed, and different parties can adopt a different number of their pledges.
This article applies the pledge approach to the party mandate model (Royed Reference Royed1996; Thomson Reference Thomson2001; Mansergh and Thomson Reference Mansergh and Thomson2007; Naurin et al. Reference Naurin, Royed and Thomson2019; Schermann and Ennser-Jedenastik Reference Schermann and Ennser-Jedenastik2014b; Thomson et al. Reference Thomson, Royed, Naurin, Ferguson, Kostadinova, Moury, Pétry and Praprotnik2017). While coalition agreements are often used as the independent variable explaining final pledge fulfilment, this article uses the adoption of pledges into coalition agreements as the dependent variable and theorizes about what factors influence pledge adoption. I distinguish between two kinds of adoption: full adoption and partial adoption, which means that a limited version of a pledge is inscribed into a coalition agreement (Schermann and Ennser-Jedenastik Reference Schermann and Ennser-Jedenastik2014a). In this article, I argue that some features on pledge and issue level impact on the parties’ ability to adopt their pledges fully, and to force them to compromise on a form of partial adoption. Dummies for individual parties are also included because party characteristics such as size, organization or experience can influence the ability to adopt pledges fully or partially.
The analysis rests on a unique dataset of 1394 narrow Czech coalition party pledges in three governments established after elections in 2006, 2010 and 2013. Although the Czech Republic was chosen because of the author’s language skills (being a native speaker is crucial for this kind of pledge analysis), it also serves as a good case for study. First, the empirical research on party mandate fulfilment is underdeveloped in Central-Eastern Europe. Compared with established Western European democracies, there are only a few new democracies where pledges have been examined (Roberts Reference Robertsn.d.; Kostadinova Reference Kostadinova2013; Svačinová Reference Svačinová2016; Škvrňák Reference Škvrňák2015). Deeper studies into the pre-fulfilment work with pledges are non-existent for this area. Second, the way of writing coalition agreements is still developing and undergoing change. It is also important to capture this process in young democracies. Finally, the Czech case is typical with the presence of new (entrepreneurial) parties in coalitions, and it is important to look at the way these parties work on fulfilling their mandate. As the trend of new parties in coalitions is typical for more countries in Central-Eastern Europe, the Czech Republic is a relevant case for pledge adoption study here.
The paper is structured in the following way: in the theoretical part, I set the study into the pledge approach to party mandate model and discuss party strategies regarding the adoption of pledges into coalition agreements. Then, I formulate the hypotheses about the full and partial adoption of pledges influenced by the pledge and issue-level variables. In the empirical part, I introduce the Czech case and the new dataset, and test the hypotheses with the use of multinomial logit regression. The interpretation of the results is followed by a conclusion. Generally, the results show that the consensus among parties about the pledges increases the probability of partial adoption. Contrary to that, issue level variables defining the salience of the issue do not work as expected. Additionally, party size and experience seem to impact on full adoption of pledges.
Party Mandate Model and the Role of Coalition Agreement
The party mandate model (APSA 1950) describes how the links between citizens and parties should work in a democracy. As a normative theory, it defends democracy as a good type of government. As a descriptive theory, it provides an account of how high-quality democracy should work and establishes the framework for empirical testing of the expectations (Hornsteiner Reference Hornsteiner, D’Ottavio and Saalfeld2015, 113). The model, in its essence, assumes that parties create manifestos, including pledge packages, about future actions or results if they get into government. Voters compare the policy offers from the parties and vote for the party that best corresponds to their preferences. Once elected, parties get the role of government or opposition and are expected to work to implement their election promises. Voters are informed about enacted policies, evaluate government performance, and finally either reward or punish parties for non/fulfilment of the mandate (for example, Klingemann et al. Reference Klingemann, Hofferbert and Budge1994; Louwerse Reference Louwerse2011; Royed Reference Royed1996; Thomson Reference Thomson2001). The fear of punishment motivates parties to fulfil promises. Additionally, if a governing party is not able to fulfil its promises, then the pledges in its following campaign seem less credible (Costello and Thomson Reference Costello and Thomson2008: 19). The established relationship between voters and parties thus leads to a translation of voters’ preferences into real government policies.
The pledge approach empirically measures the fulfilment of party mandates, which is operationalized as the number of narrow pledges in a manifesto (for the original definition of broad and narrow pledges, see Royed et al. Reference Royed, Naurin, Thomson, Naurin, Royed and Thomson2019, 24–30). More fulfilled pledges at the end of the government period mean better mandate fulfilment. It was also found that parties are more likely to fulfil pledges that obtain majority support in government and parliament, represent the status quo, and where coalition parties are positionally closer together on the respective issue, etc. (Royed Reference Royed1996; Thomson Reference Thomson2001; Mansergh and Thomson Reference Mansergh and Thomson2007; Costello and Thomson Reference Costello and Thomson2008; Schermann and Ennser-Jedenastik Reference Schermann and Ennser-Jedenastik2014b; Thomson et al. Reference Thomson, Royed, Naurin, Ferguson, Kostadinova, Moury, Pétry and Praprotnik2017).
Adopting pledges into coalition agreements serves as a stable explaining variable of pledge fulfilment. This is because coalition agreement serves as the whole coalition government’s policy agenda instead of manifestos and increases the salience of common coalition pledges. The coalition agreement is also a conflict prevention device, where parties pre-define the ministerial policy and facilitate the future coalition life (Timmermans Reference Timmermans2003; Quinn Reference Quinn2014). If the party forces its coalition partners to agree with the pledges adopted into the coalition agreement, it is probable that the deal will be followed in the future, even if some party tries to ignore it.
Empirical studies repeatedly show that parties take pledges in coalition agreements seriously. Pledges adopted into coalition agreements were found to have more than twice the probability of being fulfilled compared with pledges that are just mentioned in a party manifesto (Mansergh and Thomson Reference Mansergh and Thomson2007; Costello and Thomson Reference Costello and Thomson2008; Schermann and Ennser-Jedenastik Reference Schermann and Ennser-Jedenastik2014b; Praprotnik Reference Praprotnik2017). Schermann and Ennser-Jedenastik (Reference Schermann and Ennser-Jedenastik2014a) even equate the relationship between pledge adoption into the agreement and its real fulfilment. Adopting pledges into coalition agreements is seen as the first step in their fulfilment.
Until now, we do not know much about what pledges are preferentially adopted and what influences their adoption. There is only one article that investigates the way pledge adoption is taken into coalition agreements in Austria (Schermann and Ennser-Jedenastik Reference Schermann and Ennser-Jedenastik2014a). The authors found that adoption is influenced by the same variables as fulfilment. Alas, they did not look into the differences between partial and full adoption and did not think about the parties’ strategies regarding full and partial adoption. This article tries to fill this research gap. Parties negotiating the coalition agreement have a clear goal – to get the support of their coalition partners for their pledges. Of course, not all pledges can be adopted because the document has a limited length. As such, coalition parties must bargain on what pledges will be adopted. They can use various skills and force their coalition partners to compromise.
Schermann and Ennser-Jedenastik (Reference Schermann and Ennser-Jedenastik2014a) define three possible pledge adoption results. The worst result is non-fulfilment, meaning that the action or the result intended in the pledge is not translated into the coalition agreement. It still can be fulfilled, but the probability decreases. The best possible result is full adoption, meaning that ‘the action outlined in the pledge needs to be manifestly written down in the coalition agreement’. Partial adoption is an intermediary category and means that ‘a limited version’ of the proposed action is added to the agreement (e.g. a tax cut of only half the size that was originally promised) (Schermann and Ennser-Jedenastik Reference Schermann and Ennser-Jedenastik2014a, 796). The authors found non-adoption and partial adoption to be the dominant outcomes of negotiations in coalition agreements.
Eichorst (Reference Eichorst2014) stresses that parties can utilize coalition agreements as a source of advertisement and boast to their members and supporters about the higher probability their pledges have of fulfilment. It works for them as it shows their bargaining skills in the coalition formation phase (Bolleyer Reference Bolleyer2007; Eichorst Reference Eichorst2014). The full adoption of the pledge best advertises the party’s coalition agreement function. Partial adoption is the second-best result but can also be interpreted as a party’s concession from its wishes when party preferences diverge. A party (or parties) must make some concessions to ensure at least something of its pledge will emerge in reality. The compromise shows voters that the party works on fulfilling its pledges but is not strong enough to guarantee exactly what it promised. Voters can be somehow disappointed, but certainly less than in the case of non-adoption.
In the following section, I present three hypotheses that point to the possibility that, under some circumstances, the compromise can be the most expected result of coalition bargaining. I look at two levels of variables: the pledge level and the issue level. I do not theorize about the difference between non-adoption and adoption but focus on the difference between full and partial adoption. I expect that the variables define the salience of the parties’ pledges and, simultaneously, the possibility of conflict between parties. If these two options coincide, full adoption is not likely because the coalition partners are in opposition, but the importance of the pledges for the parties determines special emphasis of the parties on their pledges, so that they strive at least for compromise.
Factors Leading to a Compromise in Adoption of Pledges
Pledge Level: Consensual Pledges
It seems intuitive that the individual party’s chance of adopting its pledges, and adopting them fully or partially, is affected by the coalition partners’ consensus with promised actions or results. Congruency in coalition is measured by the consensus of party pledges among coalition partners. A pledge is consensual if there is a similar pledge in coalition party manifestos. Consensual pledges were repeatedly found to increase the probability of pledge fulfilment (Thomson Reference Thomson2001; Kostadinova Reference Kostadinova2013; Naurin et al. Reference Naurin, Royed and Thomson2019; Schermann and Ennser-Jedenastik Reference Schermann and Ennser-Jedenastik2014b; Škvrňák Reference Škvrňák2015; Thomson et al. Reference Thomson, Royed, Naurin, Ferguson, Kostadinova, Moury, Pétry and Praprotnik2017). Based on arguments about pledge fulfilment, Schermann and Ennser-Jedenastik (Reference Schermann and Ennser-Jedenastik2014a) showed that consensual pledges also increase the probability of pledge adoption into the coalition agreement. The variable defining consensus had the greatest explanatory power in their models.
The pledge is usually seen to be consensually related to the second party’s pledge if the fulfilment of the second party pledge would automatically result in the partial or complete fulfilment of the first party pledge (Costello and Thomson Reference Costello and Thomson2008, 255). Thus, because the consensual pledges do not necessarily promise the same, there is some space for conflict about the final version of the pledge in the agreement. Consensus makes pledges salient: first, the fact that a similar action was promised by all coalition parties mirrors some public demand for the action or result. Second, if the pledge is nearly the same for more parties, it makes action in its direction very probable, and so it is to be expected that some version of it will be adopted. Consensus defines a kind of conflict. As Eichorst (Reference Eichorst2014) theorizes, the conflict decreases the value from advertising. Partial adoption is, therefore, a probable result:
H1: As the consensus about the pledge among coalition partners increases, the pledge is more likely to be partially adopted.
Issue Level: Party- and Voter Issue Salience
The salience of pledges can be defined by different means. In manifestos, parties formulate pledges on different issues. According to issue salience theory, they compete for votes by increasing the emphasis on some issues. The more emphasized issues reflect the saliency parties give to the policies on these issues. Party mandate responsiveness should therefore also be influenced by party issue emphasis. Basically, parties should push the policies on their salient issues in coalition agreements because the benefits from advertising an individually salient issue are great (Green and Hobolt Reference Green and Hobolt2008; Dolezal et al. Reference Dolezal, Ennser-Jedenastik, Müller and Winkler2014). The pledges on issues emphasized by the party should be crucial for the party because its electoral campaign was centred on its salient issues. Voters should be sensitive to pledges on these issues when evaluating party performance. During the coalition negotiations, the party should try to adopt its salient pledges, and of course, in their full versions.
However, the costs must also be counted. It is relatively cheap to push pledges in individually salient issues if it is non-salient for other coalition partners. However, when the issue is important for more coalition partners, conflict about the policies desired for the issue is probable, and parties must compromise. In such circumstances, the benefit from advertising is smaller, because the costs of pushing one’s own policy against the partners in a generally salient issue is small. As Eichorst (Reference Eichorst2014, 101) states: ‘On relatively more salient dimensions, the potential electoral benefits from advertising points of consensus are marginal when compared to those of advertising benefits on relatively less salient dimensions.’ Parties should thus advertise their position in their individually salient issues, which are not generally salient. I expect that the increase in issue salience for all coalition parties decreases the probability of fully adopting the pledge but increases the probability of partial adoption.
Parties are not necessarily able to manage the public debate and to emphasize their individually salient issues successfully when an external shock can draw attention to some previously not emphasized issue (Milita et al. Reference Milita, Ryan and Simas2014). This kind of importance, not identifiable in manifestos, reflects voter issue salience at the polls. Parties are motivated to adopt pledges in the voter salient issue to show their responsiveness to voters. However, the value from full pledge adoption on a generally salient issue is decreased by the possibility of conflicts with coalition partners. I, therefore, expect that the increase in voter issue emphasis in manifestos will decrease the probability of partial adoption, but increase full adoption.
H2: As the party issue salience decreases, the pledge is more likely to be partially adopted.
H3: As the voter issue salience increases, the pledge is more likely to be partially adopted
There is also a number of party characteristics that can have an impact on pledge adoption. Some parties can simply get more policy gains because they are bigger, or because they are formateurs of the coalition (Thomson et al. Reference Thomson, Royed, Naurin, Ferguson, Kostadinova, Moury, Pétry and Praprotnik2017). Finally, it can be also useful to think about the differences between party types. New parties and parties with inexperienced leaders are more likely to miss important information in the negotiation process (Moury Reference Moury2013, 4; De Winter and Dumont Reference De Winter, Dumont, Strøm, Müller and Bergman2008, 134; Deschouwer Reference Deschouwer and Deschouwer2008, 5). The experienced coalition partners can dominate the coalition bargaining and force new partners to compromise. Finally, the way a party is organized can also be important. So-called entrepreneurial parties establish small party organization with the leader’s strong position, with weak party members and party elites. This kind of party does not care about party manifestos (Louwerse Reference Louwerse2011). Inexperienced and entrepreneur parties also entered coalitions in the Czech Republic. I do not formulate a specific hypothesis connected to these differences, but they must be taken into consideration when interpreting the results. I, therefore, include party dummies in the analysis.
Case Selection, Variables and Data
This article examines the data on parties from three Czech coalition governments formed after the elections in 2006, 2010 and 2013. These three governments included two centre-right and one centre-left coalition. The government of Mirek Topolánek started in 2007 and consisted of the centre-right prime ministerial Civic Democratic Party (ODS) and two smaller parties, the Christian Democrats (KDU-ČSL) and the Green Party (SZ), which were in parliament for its first time. In 2010, the centre-right coalition of Petr Nečas was formed, including ODS as the prime ministerial party, and two new parties, the conservative Tradition, Responsibility, Prosperity 2009 (TOP 09), which was composed of splitters from KDU-ČSL, and the populist Public Affairs (VV). The last is the centre-left coalition of Bohuslav Sobotka. It consisted of Social Democrats (ČSSD), the new populist Yes Movement (ANO) and KDU-ČSL. These governments got the confidence of the Chamber of Deputies as the first governments after the elections. When Topolánek’s and Nečas’s governments ended prematurely, their successors did not create new coalition agreements.
I work with the manifestos from seven parties in three governments. Three of them, ODS, ČSSD, and KDU-ČSL were experienced mainstream parties. Four of them were new, i.e. they entered Parliament for the first time, and in the same period they also entered into a coalition government: this includes SZ in Topolánek’s government, TOP 09 and VV in Nečas’s government, and ANO in Sobotka’s government. It must be also mentioned that the parties differed not only in their experience, but also in party size. Finally, two parties serve as examples of entrepreneur parties. Table 1 summarizes the main characteristics of the coalition parties.
Table 1. Investigated coalition parties and governments.
Note: The total number of legislative seats is 200.
The dataset covers 1394 narrow coalition party pledges (for discussion of broad and narrow pledges, see Royed et al. Reference Royed, Naurin, Thomson, Naurin, Royed and Thomson2019).Footnote 1 As the sources of the data are electoral manifestos and three coalition agreements, all manifestos were coded manually. The reliability tests for extracting the pledges and main variables were done with the help of coders who are Czech native speakers, on the samples of pledges extracted from manifestos.
Royed’s (Reference Royed1996, 79) definition of a pledge as the ‘commitment to carry out some action or produce some outcome’ was used to initially identify the pledges in the manifestos, with precision given by the instructions of the AUTNES approach (Dolezal et al. Reference Dolezal, Ennser-Jedenastik, Müller and Winkler2014). Pledges were identified as every grammatical sentence consisting of a subject (‘we’, ‘the party’), a noun phrase (the subject of the pledges, the noun and some extensions – adjectives, descriptions of the subject) and a verbal phrase (a verb linked to the object and the subject). If there were more areas or more of the action’s objects in one sentence, the sentence was considered as formulating multiple pledges. For example, the sentence ‘we will increase salaries for teachers and non-pedagogical employees’ would be identified as two pledges (increasing salaries for teachers and increasing salaries for non-pedagogical employees). Possible interconnectedness of more pledges was also taken into consideration – two or more actions were identified as one pledge if one action was conditioned on a second action.Footnote 2 From the initial dataset of all, i.e. broad and narrow pledges, only the narrow pledges were picked up. The narrow pledge must be specified in sufficient detail for the researcher to evaluate whether the action or result was realized. Fulfilling the pledge must occur before the end of the parliamentary term. The criteria used to judge the pledge fulfilment must, in principle, be provided by the writers of manifestos and should be objectively measurable. If value judgements are required to assess the fulfilment of the pledge, the pledge is identified as broad. Examples of narrow pledges are: creation of new laws, rules, strategy, audit or revision in a specified area (for example, ministry), new obligation, etc. Examples of broad pledges are: improvement, making something fairer, financially viable, easier, etc without proper description (we will simplify tax controls, we will motivate teachers to teamwork, we will introduce fair taxes, etc).Footnote 3
Variable Measurement
The dependent variable is the adoption of the pledge into the coalition agreement. It is defined as a categorical variable having three values: non-adoption (0), partial adoption (1), and full adoption of the pledge (2). The non-adoption of the pledge means that no action or result regarding the pledge was mentioned in the coalition agreement. The exact pledge manifesting in the coalition agreement means full adoption, and a limited version of the pledge in the coalition agreement means partial adoption. For example, a pledge promising ‘we will decrease value added tax (VAT) by 10%’ can be seen as fully adopted if the pledge in the coalition agreement is the same, or with only a negligible change, not deviating from the original meaning. Partial adoption would be ‘we will decrease VAT by 8%’ or ‘we will decrease VAT by 10% for some groups of services’.Footnote 4
The independent variable for H1 is the degree of consensual agreement of the pledges among coalition parties, meaning that the pledge of one party is considered to be consensually related to the second party’s pledge if the fulfilment of the second party pledge would automatically result in the partial or complete fulfilment of the former party pledge (Costello and Thomson Reference Costello and Thomson2008). Because there were always three coalition parties, the variable gets values 0, 1 or 2. To identify consensual pledges, I also considered also the broad pledges as consensually related to focused narrow pledges.
To construct an issue-level variable for H2 and H3, every narrow pledge was categorized into one of 13 issue areas (economy, welfare state, budget, education, security, army, foreign affairs, Europe, infrastructure, society, environment, reform of institutions, immigration) according to the coding scheme of Dolezal et al. (Reference Dolezal, Ennser-Jedenastik, Müller and Winkler2014), with the addition of several topics prominent in Czech politics.Footnote 5 The explaining variable for H2, i.e. party issue salience is based on the percentage of narrow pledges dedicated to the issue. If a party dedicated 20% of pledges to the economy, every one of this party’s pledges devoted to the economy was coded with the value 20.
The explaining variable for H3, voter-issue salience was measured on the basis of two post-election surveys (Sociologický ústav Akademie věd ČR 2006, 2010, 2013). I used answers to a question about the first and the second most important issues in the election. The answersFootnote 6 of coalition party voters were attributed to one of the 13 policy issues described above and weighted in order to find the difference between the first and second most important issue (first answer was given double the weight of second answers). Finally, I created dummies for each coalition party and compared the effects of the parties in different groups.
Analysis
Before moving on to statistical analysis, Table 2 presents the distribution of pledges for individual parties. It is clearly seen that the trends in writing manifestos have changed over time, as well as the trends in adopting pledges. Parties first wrote very lengthy manifestos containing the highest number of narrow pledges. Coalition agreements were very short, so there was a small possibility of adopting pledges, and full adoption was very rare. This trend has changed. The number of pledges has decreased in manifestos, but the length of coalition agreements has increased. The writing style has changed to whole sentences that were more often adopted in their original form. Party manifestos in the 2013 election were very short, and the coalition agreement was longer than them all. Many sentences from the manifestos were easily copied and pasted into the coalition agreement. This completes the picture of the highest rate in adopting pledges in the Sobotka government, and also the high percentage of fully adopted pledges.
Table 2. Adoption of party pledges into coalition agreement (2006, 2010, 2013).
Table 3 presents the results of the multinomial logit regression model with the narrow pledge as a unit of analysis and its adoption into coalition agreement as the dependent variable. The dependent variable has three values. To compare the compromise (partial adoption) to the best result (full adoption), I take value 2 (full adoption) as the base. Multinomial regression simultaneously allows us to see the effect of the variables on the chance of a pledge to be non-adopted and partially adopted. Relative risk ratios above 1 mean a higher chance of the option taking place over a reference category. To interpret the effects of independent variables, I use graphs of marginal effects to illustrate the size of the effects by the predicted probability of different adoption categories.
Table 3. Multinomial logit model.
Standard errors in brackets. *<0.1; **<0.05; *** <0.01. Note: the log likelihood of null model for Model 1 is –1121.8.
Regarding H1, I expected that the increase in the number of consensual pledges increases the probability of partial pledge adoption compared with full adoption. Eighteen percent of party pledges were consensual with one party, 5% with both coalition partners, and the percentages did not differ much between parties. Thirty-five percent of pledges consensual with one party, and 47% consensual with two parties were partially adopted, contrary to 19%, respectively 32% of fully adopted pledges. The relative risk ratio in model 1 means that the odds of partial adoption increases (compared with full adoption) 1.28 times if one coalition partner has a consensual pledge, and 1.09 times if both coalition partners have a consensual pledge. H1 can be supported. Actually, there was an exception from the depicted trend. ČSSD in Sobotka’s government was able to fully adopt 50% of pledges that were consensual to one party, contrary to one third of partially adopted pledges in this category (with the equal percentage of fully and partially adopted pledges consensual for two parties). However, ČSSD was an exception in full adoption of its pledges.
The second and third hypotheses expect the effect of party and voter issue salience on pledge adoption. First, I expected that, while controlling for voter issue salience, the probability of partial adoption would decrease with the increase in party issue salience, contrary to full adoption. The value of the relative risk ratio shows that the chance of partial adoption is 2% smaller when party issue salience increases by 1%. The probability of partial adoption decreases from 27% for the party non-salient issues to 12% for the most salient issues, but contrary to my expectation, so does the probability of full adoption. H2 must, therefore, be rejected. Moreover, when party issue salience exceeds the value of around 16%, which identifies issues that were very salient in all parties’ manifestos, the differences in partial and full adoption probabilities disappear. Actually, even when contradictory to the expectation of issue salience theory, these issues (economy, welfare state, reform of institutions, or infrastructure) were among the most important for all parties.
For H3, I expected that the possibility of conflict given by increasing voter issue salience makes partial pledge adoption more probable than full adoption. However, a 1% increase in voter issue salience decreases the chance of partial adoption by 1.8% in relation to full adoption. Partial and full adoption probability increases (Graph 3 in Figure 1), but a look at the difference between partial and full adoption shows opposite directions to those expected. The probability of full adoption increases, while the probability of partial adoption slightly decreases. So, H3 must also be rejected. The same trend as in the case of party issue salience appears when voter issue salience exceeds the value of 25 (including the most voter salient, which, however, covers also party salient issues, such as: reform of institutions, economy, welfare state, and budget), and the probabilities of both, partial and full adoption, overlap. Contrary to my expectations, parties are able to get the best results in issues that are important for them all, whether defined by party manifestos, or by voters.
Figure 1. Graphs of marginal effects for independent variables.
A possible explanation for the non-support of H2 and H3 is the fact that party and voter salient issues overlap in the case of Czech coalition parties. In such cases, the conflict does not necessarily end in compromise, but in selective full and partial adoption. Certainly, not all pledges in one issue need to be of the same salience for the promising party. Some pledges in the issue are more important and are presented in pre-electoral debates and meetings, but that doesn’t mean parties take all their pledges in the issue with the same seriousness. The saliency of individual pledges is, however, difficult to determine.
Finally, party dummies should show the possible compromise forced on some parties. The biggest chance for non-adoption followed by partial adoption was for all parties except ČSSD in Sobotka’s government. Looking at the probability of partial versus full adoption, there are differences between individual governments. First, the parties in Topolánek’s government clearly differ from the rest. The probability of pledge adoption is very small, and the differences between partial and full adoption are the smallest (3% in case of SZ, 7% for ODS, and 10% in KDU-ČSL case). The particularity of Topolánek’s government can be explained by the different manifesto structure at that time, and the way pledges were adopted into a very short coalition agreement. The cases of Nečas’ and Sobotka’s governments are much more similar. First, the prime ministerial parties in both governments had higher probabilities of full pledge adoption. The difference was almost imperceptible in the case of ODS in Nečas’ government (one percentage point difference), but very noticeable in Sobotka’s ČSSD (30 percentage point difference between full and partial adoption). The other coalition parties had a higher probability of partial adoption, with a similar difference of around 14 percentage points between partial and full adoption, including new parties (VV, TOP 09, ANO), but also one traditional party (KDU-ČSL in Sobotka’s government). There are two possible explanations for this similarity. First, KDU-ČSL was outside the parliament for one electoral period (2010–2013). After that, it succeeded again and entered the coalition, but in view of this fact, it also had some characteristics of new parties – during parliamentary exile, it had changed some of its leading officials, including the leader. This possibly led to a decrease in party bargaining skills. Second, similar trends for the smaller coalition partners stress the fact that not only unfamiliarity and inexperience but also the smaller party size led the party to compromises in pledge adoption.
Conclusion
Coalition agreement is the first advertising board of coalition parties, this is the reason why parties should strive to fully adopt their pledges into the coalition agreement. However, because no party plays the coalition game alone, compromise in adopting pledges is a likely result in some cases. In this article, I test the expectation that the possibility of conflict about similar pledges or the same salient issues increase the probability of compromise, i.e. partial adoption of party pledges. Testing this expectation on the party data from three coalition governments in the Czech Republic, I found that parties are more likely to adopt their pledges partially (in comparison with full adoption) if their consensus increases. Contrary to expectations, parties are able to force full adoption in highly salient topics with the same probability as partial adoption, and do not necessarily compromise there. I present the possible explanation that not all pledges in salient issues are of the same importance for parties. Finally, I found that new parties compromise more often than traditional parties, and that party size, or the difference between prime minister and the rest of the parties, also plays its role.
This article contributes to the pledge approach, with the study of the pre-fulfilment phase of party mandate fulfilment. Party negotiation on the coalition agreement is a very important, but underdeveloped phase of party coalition negotiation, and previous knowledge anticipated the same pledge adoption and pledge fulfilment mechanism. I show that consensus is a conflicting factor resulting in compromise in pledge adoption rather than the typically expected positive effect. Finally, my article also contributes a new country dataset from Central-Eastern Europe, which is neglected in the empirical studies on party mandate fulfilment. The Czech case also shows an interesting change in pledge formulation and adoption over time.
The article shows new directions in the research of pledge adoption. Besides consensual pledges and different party types, there are possibly other variables that can influence the adoption of pledges. In addition, partial adoption is not the only way to compromise. Adopting a pledge in its broad version can serve as an alternative measure. Within the trend of changing party systems and new kinds of parties entering into coalitions, similar analyses on countries in Central-Eastern Europe would test and broaden the findings coming from this article.
Acknowledgement
This article was supported by the project ‘Postdoctoral job positions at UHK’.
About the Author
Petra Vodová is a post-doctoral researcher at the University of Hradec Králové in the Czech Republic. She is interested in the study of coalitions, especially in the study of the adoption of party pledges, and portfolio allocation.
