Responsiveness when parties are “weak”: A candidate-based analysis of voter-party congruence in Europe

Abstract

Outside the US, the crucial question of how well politicians represent the preferences of voters is usually investigated at the party level. Reversing this perspective, we examine representation in Europe from the point of view of individual candidates running in national parliamentary elections. This is especially insightful in a period that seems characterized by a decline in parties’ representational capacities and an increasing personalization of politics. We analyze representation by considering the incongruence between candidates’ left–right positions and the average placement of their party voters. By combining candidate survey data with mass survey data on voters, we assess how ideological incongruence varies according to predictors measured at the levels of candidates, parties, and party systems. The results highlight a systematic association between a partisan style of representation and candidates’ proximity to voters, as well as the interactions between representational roles and factors such as the anti-establishment nature of parties and ideological polarization in the party system.

Keywords

candidates congruence Europe left–right political parties representation

Forward to the past?

How well do politicians represent their voters? In their seminal article “Constituency influence in Congress,” Miller and Stokes (1963) answered that it depended on three conditions. Voters should choose a representative who shares their ideas; they should be able to punish unresponsive representatives at the next election; and, last but not least, voters should be knowledgeable about the opinion of the candidates, a condition not considered an untenable assumption (Eulau and Karps, 1977). In a context of uninominal electoral systems and weak parties such as that of the US, Miller and Stokes thought that the quality of the representation nexus differed more among issue domains and in regard to the individual characteristics of the (elected and prospective) candidates than among the voters of a constituency. When empirical studies on representation began in Europe, the question of how representation works changed. It became how well parties, with their distinct policy platforms and almost stable ideological leanings, are able to reflect the points of view of their voters. This approach was named the “responsible party government” model, since in a context where PR electoral systems and multi-party competition are predominant, parties—not individual representatives—were the main representative agent (see, e.g. Converse and Pierce, 1986; Dalton, 1985; Di Virgilio and Segatti, 2016; Thomassen, 1994; Vandeleene et al., 2019). Consequently, in Europe the analytical focus became different from that of Miller and Stokes: European studies have mainly evaluated the extent to which party voters’ views correspond to the aggregated (averaged) ideological stances or issue opinions of their party representatives.

Miller and Stokes’ approach and the European one share a common conceptual premise. Representing a voter means acting in responsive way, i.e. taking into account her/his wills (Pitkin, 1967). The best proxy for it is the degree of congruence between the representatives and the represented. Understanding which factors can influence congruence between individual politicians and voters is the main research question addressed in this article. As in the two aforementioned approaches, our article conceives responsiveness as congruence. However, our analytical perspective takes the side of Miller and Stokes, since we examine representation from the point of view of individual politicians—here, candidates for legislative offices. Differently from the European approach, our goal is to assess which factors make candidates more congruent with the average opinion of their voters, not the other way around.¹

Going back to Miller and Stokes is theoretically rewarding because the triangular relationship between voters, representatives and parties at the core of the responsible party model of representation appears under strain in many European countries. According to this model, parties are the agents of voters and at the same time the principals of the candidates (De Winter et al., 2020). Although the evidence is mixed, the double role of parties seems to become weaker (Rahat and Kenig, 2018). Leader personalization, decline of the party on the ground as a primary socialization agency, candidate-centered campaign activities, and consequently growing importance of candidates’ traits as “information shortcuts”—all of these are processes that increase the prominence of the direct relationship between prospective representatives and their represented (Adam and Maier, 2010; Costa Lobo and Curtice, 2015; Cutler, 2002; McAllister, 2007; Shugart et al., 2005; Van Aelst et al., 2016; Van Biezen et al., 2012). The aim of this article is not to ascertain whether and how the role of parties as principals of their candidates is declining, but to look at the potential consequences of it by raising also a relative novel research question in Europe. What are the candidates’ characteristics that make them congruent with their voters, and how are they conditioned by certain party- and election-level characteristics?

Congruence is conceived as the ideological distance between candidates’ self-placement on the left–right spectrum and the self-placement of their voters. To explain variation in the degree of congruence, we take into account three sets of factors: along with properties of parties and election-specific determinants (i.e. the “usual suspects”), we examine predictors measured at the level of individual candidates. We expect all these factors to have direct effects on the level of congruence. At the same time, we claim that party- and election-specific characteristics condition the effects of candidates’ individual characteristics on the degree of ideological distance with their voters.

The article is organized as follows. The next section discusses the main factors that can influence the degree of ideological congruence between party members and voters. In particular, we consider individual attributes, party properties, and election-specific determinants. In the third section we present the comparative dataset that we constructed, which combines candidate survey data with mass survey data on voters, and illustrate variable operationalization. The fourth section reports the results of a set of multi-level models where candidate–voters distance on the left–right scale was regressed on covariates gauged at three different levels. The final section concludes by summarizing the main findings and pointing to some avenues for future research.

Explaining (in)congruence between party candidates and voters

In most European studies, voters’ ideological positions are compared with some indicator of the position of the voted party, considered as a unitary actor. A common strategy is to calculate the absolute distance (or the correlation) between the average self-location of party voters and the average self-location of party members (Barnes, 1977; Converse and Pierce, 1986; Costello et al., 2012; Dalton, 1985; Dalton et al., 2011; Irwin and Thomassen, 1975; Kitschelt et al., 1999; Miller et al., 1999; Rohrschneider and Whitefield, 2012; Schmitt and Thomassen, 1999; Walczak and Van der Brug, 2013). Fewer studies have analyzed congruence at the individual level, computing the distance between single party voters and the average self-location of party members (e.g., Belchior, 2013; Enyedi et al., 2020).²

Voters’ left–right scores are usually drawn from mass post-election surveys. The ideological positions of parties are often derived from elite surveys (where respondents are either candidates or elected officials), while they are sometimes obtained from expert surveys or party manifestos. Parties tend to be treated as unitary entities with a single “ideal point” gauged as the mean of the distribution of the surveyed elite members or of country experts’ judgments. In contrast, we explore congruence as a relation between individual candidates and the overall electorate of their party.

In what follows, we identify a set of theoretically relevant attributes of candidates that directly or may conditionally affect their ideological proximity—i.e. congruence—with their electorate. Subsequently, we review the main party-level and election-level determinants highlighted in the literature and outline a number of expectations.

Candidate-level factors

Certain attributes can make candidates particularly close to their voters, thereby reducing incongruence. One of them concerns candidates’ career patterns and past political experience, since more experienced politicians are presumably more accurate in identifying the preferences of their party voters. Insofar as politicians strategically choose to take a certain position in the policy space in order to acquire votes, candidates with longer experience should be more congruent with their voters. This is consistent with the argument that lower convergence between parties and greater voter-party congruence may arise because of rational behavior by candidates (Downs, 1957).

Secondly, the level of proximity between a candidate and her/his electorate may depend on the style of representation. In this regard, the well-known distinction among “trustees,” “delegates” and “partisans” helps understand how a candidate conceives her/his role as a representative (Andeweg and Thomassen, 2005; Converse and Pierce, 1986; Eulau et al., 1959; Giebler and Wessels, 2013). Trustees regard themselves as acting as free individuals, constrained solely by their own judgment. Delegates conceive themselves as advocates of their geographically defined constituency. Finally, partisans perceive their role as representatives of the voters of the party to which they belong. Put simply, those candidates who perceive themselves as partisans should be ideologically closer to the opinions of their party voters than those who perceive themselves as delegates or trustees.³

In examining the influence of representational styles on congruence, a caveat is necessary. We acknowledge that the causal direction could also be the reverse—that is, a close matching with the ideological preferences of the party electorate might induce a candidate to define herself/himself as a partisan. At a minimum, the relationship is bi-directional, dynamic and mutually reinforcing. Following a standard approach in studies of representation, we assume that the style of representation weighs more heavily, as representative roles can be viewed as strategies that politicians choose to achieve their own goals more than something determined by voters’ preferences (Önnudóttir, 2014; Strøm, 2012). Indeed, the adoption of a given representational style can be better accounted for by a candidate’s socialization and career path, and by the incentives created by the electoral rules under which a candidate operates (Önnudóttir and Von Schoultz, 2020).

Party-level factors

Some individual candidates can be closer to their electorate than other candidates simply because they belong to a party with certain characteristics. Here we take into consideration the size, age and internal cohesion of parties, as well as parties’ position with regard to the party system. All else being equal, it can be argued that voter-party congruence is lower for large parties. While small parties can more easily identify the ideological profile of their voters, large parties often propound a wide range of policy positions in order to obtain votes from broad portions of the electorate (Mattila and Raunio, 2006; McEvoy, 2012). We expect also that candidates from parties with a longer electoral history are more congruent with their voters because they probably have better information on their electoral base. Those who run for office in a brand-new party are unlikely to know precisely where their voters are because they do not rely on a well-established and recognizable electoral base. In addition, congruence with voters is probably higher for internally cohesive parties. A party with ideologically heterogeneous members will probably attract voters holding a broad range of policy preferences, which will reduce the congruence between a candidate and the party electorate.⁴ Finally, where parties stand in relation to other parties may matter as well. On the one hand, candidates of anti-establishment parties may be subjectively inaccurate as to the left–right positions of their voters since they tend not to use this ideological yardstick. On the other hand, anti-elite parties, even when locating themselves on the left–right, may misunderstand their voters’ positions because the latter are often related to discontent and protest (Van der Brug and Fennema, 2007; Van Hauwaert and Van Kessel, 2018).

Party characteristics, however, do not only exert a direct effect on the level of ideological congruence of any candidate, irrespectively of who they are. In particular, a few party characteristics might moderate the effects of individual determinants of the level of ideological congruence with their voters. Even candidates with a partisan style of representation and those with a long career might be somehow less close to their voters in ideological terms if they belong to parties with certain characteristics: large size, newness, internal heterogeneity and non-mainstream nature.

Election-level factors

The distance between candidates and their voters may vary among candidates also because of the context in which representatives and the represented interact. Here we focus on two contextual variables routinely considered by academic research: the effects of the electoral system and the overall degree of ideological polarization.

As regards the former, systems with strong proportional effects are usually expected to produce levels of voter-party proximity higher than those of majoritarian (i.e. disproportional) systems (Blais and Bodet, 2006; Dalton, 1985; Ezrow, 2010; Golder and Lloyd, 2014; Powell, 2000; Wessels, 1999). As regards the latter, polarization in the party system tends to reduce the policy gap between candidates and voters because programmatic differentiation among parties increases the likelihood that voters will find a party that matches their views (Kim et al., 2010; Powell, 2011). Ideological polarization at election time should strengthen left–right identifications, increase their impact on voting decisions, and hence foster proximity between a candidate and the voters of her/his party.

Besides directly affecting candidate–voters congruence, the effects of electoral rules and ideological polarization can interact with predictors measured at the level of individual candidates. More precisely, the distance between a candidate and her/his party voters can be reduced by her/his political experience and partisan style of representation. However, this is unlikely to hold under certain system-level conditions such as high polarization and limited electoral disproportionality. Insofar as ideology works as a structuring force in national electoral contests (Weber and Franklin, 2018), increasing levels of polarization make candidates’ traits less relevant. Similarly, when the electoral system has strong proportional effects (Gallagher, 1991), individual attributes of candidates may be less relevant.

Summing up, in this study we test the following hypotheses:

H1: Candidates with individual traits such as a long political experience and partisan representational style tend to be closer to their party voters in ideological terms.

H2: Party-level characteristics such as party smallness, age, internal cohesion and mainstream nature make all candidates closer to their party voters in ideological terms.

H3: Election-level factors such as the degree of proportionality of electoral rules and left–right polarization make candidates closer to their party voters in ideological terms.

H4: The impact of those candidate-level traits that reduce ideological incongruence should be weaker in the case of large, new, internally heterogeneous and anti-establishment parties.

H5: The impact of those candidate-level traits that reduce ideological incongruence should be weaker when the electoral system has strong proportional effects and when left–right polarization is high.

Many of the factors mentioned in the former three hypotheses are well-known in the literature addressing the variation of ideological congruence from the point of view of the voters. Their effects have not been yet investigated from a candidate-based perspective. The latter two hypotheses constitute the most original contribution offered by this study.

Data and methods

We test the hypotheses outlined above using a comparative dataset that combines information about candidates and voters in 21 European democracies. On the basis of data availability, our data cover parliamentary elections held from 2005 to 2017, for a total of 39 country/election occurrences. Data about candidates were provided by the Comparative Candidates Survey, while data about voters were taken from a number of mass post-election surveys.⁵ To harmonize information obtained from different sources, we converted the variables of interest to common scales. In particular, left–right scores were rescaled to a range from 0 (extreme left) to 10 (extreme right).

In our dataset, individual candidates are the units of analysis. After dropping from the dataset those candidates who gave inconsistent answers to the survey questions aimed at assessing how they perceived their representation role, our data comprises 20,311 candidates, 280 parties and 39 elections (countries/years). In the analyses, the size of the dataset shrinks further—reaching around 14,000 observations—due to missing information on some socio-demographic covariates in certain national elections.

We operationalize the dependent variable (Candidate–Voters Incongruence) as the absolute distance between a candidate’s self-location and the average self-location of those who voted for her/his party in the general elections. The distance between a candidate and her/his voters is assessed along the left–right dimension. Figure 1 provides an overview of how our dependent variable varies across Europe. While the average candidate–voters distance in our selection of countries is relatively low (1.35 on a 0–10 scale), there is considerable variation around this value. On considering the country/year averages plotted in Figure 1, one notes that a distance of just one point is observed between a candidate and her/his party voters in Luxembourg (2009), Greece (2015) and Italy (2013), which present the lowest incongruence on the left–right scale. The greatest degree of incongruence (2.31 points) is found in the Montenegrin elections held in 2012—that is, the second time that the parliament was elected after the country had formally obtained independence in 2006.

Figure 1.

Average candidate–voters ideological incongruence in Europe, by election. Notes: Average Candidate–Voters Incongruence by election, with 95% confidence intervals. The vertical line represents the overall mean in the dataset.

As to the independent variables, at the level of individual candidates we gauge their level of political experience with a variable counting the number of times a candidate has run in a general election (Political Experience). To measure the candidates’ style of representation, we follow the standard approach based on the perception candidates have of their representational role (Önnudóttir, 2014). We then create three dummies: Partisan, Trustee and Delegate.⁶

Turning to party-level variables, the proxy for Party Size is the party’s share of votes, while Party Age is gauged counting the number of years between the first post-war election contested by the candidate’s party and the surveyed election. Party Heterogeneity, as the standard deviation of candidates’ left–right positions within each party, accounts for the degree at which parties lack internal cohesiveness. The dummy Anti-Establishment captures the distinction between mainstream (0) and anti-establishment (1) parties according to Abedi’s (2004) definition.⁷ As a robustness check, we build another indicator using data from the Chapel Hill expert survey (Bakker et al., 2015): Party Anti-EU, measuring the overall orientation of the candidate’s party toward European integration. Although the anti-establishment status of parties and their degree of Euroskepticism are not the same, we contend that in Europe a pro-EU attitude is a crucial aspect of belonging to the mainstream of a country’s political system.

As regards election-level predictors, the Disproportionality variable accounts for the disproportionality of electoral systems according to Gallagher’s (1991) formula.⁸ The overall degree of ideological polarization in a given election (Polarization) is operationalized using Van der Eijk et al.’s (2005) index.⁹

A set of variables control for other factors potentially relevant to explaining the degree of ideological incongruence between candidates and voters. At the individual level, three covariates account for basic socio-demographic traits of would-be representatives: a gender dummy (Women), age in years (Age), and an ordinal variable with seven categories operationalizing education (Education). Since in many countries male politicians are still more likely than women to hold central positions in their party organization (Cotta and Best, 2008; Norris, 2005), female candidates are perhaps farther away from their party electorate than their male colleagues. Moreover, proximity to party voters should be greater in the case of older candidates (as they are supposedly more experienced) and better educated ones (as they are more sophisticated and hence more aware of their voters’ opinions).

At the party-level, we control for the ideological position of a candidate’s party (Party LR), which is obtained as the average left–right self-location of all the candidates belonging to the party. Comparative research—which however has analyzed the topic mainly from a party-level or voter-level perspective—has revealed an empirical regularity whereby proximity to voters is greater for right-wing parties and lower for left-wing parties (Belchior, 2013; Converse and Pierce, 1986). Another characteristic of parties that we take into account is their ideological extremeness. Centrist parties tend to maximize votes in the short term, while ideologically extreme parties often pursue long-term strategies. Hence, congruence should be greater in case of centrist parties and lower for far left and far right parties (Przeworski and Sprague, 1986). We therefore include Party Extremeness, which measures in absolute terms a party’s distance from the center (5) of the left–right spectrum.

Finally, at the election-level we control for the democratic record of the country where the election took place. Incongruence between party candidates and voters would be higher in countries that had become democratic more recently, and more limited where democratic elections had a long-standing history. We hence incorporate Age of Democracy, which counts the number of years from the first democratic election to the surveyed election.¹⁰

Analysis and results

The dataset that we use is organized on three analytic levels. The most basic units of analysis—individual candidates—are nested in parties, and these are nested in elections (i.e. countries/years). A three-level random effects ANOVA reveals that roughly 2% of the overall variance in the candidate–voters incongruence is explained by the election level, and about 14% of the variance is explained by the party level. The individual level accounts for the remaining variance. To model the hierarchical structure of the data, we run a number of multi-level linear regressions with random intercepts on parties and elections. More precisely, we run three sets of regressions: models combining additively the independent variables at the candidate-level, party-level and election-level (Table 1); models where candidate-level covariates are interacted with party-level covariates (Table 2); and models where candidate-level covariates are interacted with election-level covariates (Table 3). We present the models’ results in the same order and comment on the most relevant among them.¹¹ Note that in each table a positive sign for coefficients means higher distance (i.e. greater incongruence), while a negative sign indicates lower distance (lower incongruence).

Additive models

Table 1 reports results from analyses that additively incorporate candidate-level, party-level and election-level determinants. Starting from individual-level predictors, we observe that the style of representation is systematically associated with reductions in candidates’ incongruence with the party voters (H1). As hypothesized, the positive coefficients on Delegate imply that delegates are more incongruent with their party voters than partisans (the omitted baseline category). At the same time, trustees are not systematically found to be more incongruent than partisans: although always positive, the coefficient on Trustee never reaches standard levels of significance. Contrary to our expectations, a candidate’s long political experience does not reduce incongruence, since those who have run as a candidate in a general election many times in the past are not closer to the voters of their own party. With regard to the socio-demographic control variables, almost all the estimated coefficients go in the expected direction. Female candidates tend to be more incongruent with their party voters than male candidates. Age affects incongruence following a quadratic function, as the youngest candidates and the oldest candidates are more incongruent with their party voters than the middle-aged ones. Moreover, better educated candidates are less distant from their party voters.

Table 1.

Determinants of candidate–voters ideological incongruence in Europe. Candidate-level, party-level and election-level predictors.

	(1)	(2)	(3)	(4)
	Level: candidate	Level: candidate + party	Level: candidate + party	Level: candidate + party + election
Political Experience	−0.009	−0.010	−0.010	−0.010
	(0.008)	(0.008)	(0.008)	(0.008)
Style: Trustee	0.010	0.012	0.012	0.012
	(0.021)	(0.021)	(0.022)	(0.021)
Style: Delegate	0.054**	0.053**	0.051**	0.055**
	(0.024)	(0.024)	(0.024)	(0.024)
Women	0.040*	0.040*	0.038*	0.039*
	(0.023)	(0.023)	(0.023)	(0.023)
Age	−0.017***	−0.017***	−0.016***	−0.018***
	(0.005)	(0.005)	(0.005)	(0.005)
Age X Age	0.000***	0.000***	0.000***	0.000***
	(0.000)	(0.000)	(0.000)	(0.000)
Education	−0.031***	−0.032***	−0.031***	−0.032***
	(0.010)	(0.010)	(0.010)	(0.010)
Party Size		0.009***	0.006***	0.009***
		(0.003)	(0.002)	(0.003)
Party Age		−0.003**	−0.001	−0.003**
		(0.002)	(0.001)	(0.002)
Anti-Establishment		0.121**	−0.018	0.134**
		(0.062)	(0.042)	(0.058)
Party Heterogeneity			0.732***
			(0.125)
Party LR		−0.051***	−0.048***	−0.048***
		(0.011)	(0.010)	(0.012)
Party Extremeness			0.081***
			(0.021)
Polarization				−0.087**
				(0.034)
Disproportionality				−0.011
				(0.008)
Age of Democracy				0.000
				(0.001)
Constant	1.904***	2.085***	0.832***	2.420***
	(0.118)	(0.160)	(0.222)	(0.201)
Variance component:
Election	0.012***	0.017***	0.005***	0.009***
	(0.007)	(0.005)	(0.002)	(0.003)
Party	0.141***	0.100***	0.045***	0.099***
	(0.017)	(0.013)	(0.005)	(0.013)
Residual	0.920*	0.921	0.922	0.921
	(0.023)	(0.023)	(0.023)	(0.023)

N	14535	14535	14535	14535
Log-likelihood	−20251.33	−20230.22	−20172.55	−20225.95
AIC	40524.66	40490.44	40379.10	40487.91
BIC	40608.09	40604.20	40508.04	40624.43

Notes: Multi-level linear regressions with random intercepts on election and party. Robust standard errors in parentheses. Omitted category for Style of representation: Partisan. Statistical significance: * p < 0.1, ** p < 0.05, *** p < 0.01.

Turning to party-level covariates, our data support the hypothesis that the smallness, the age, the internal cohesion and the mainstream nature of parties bring all candidates near to their party voters in ideological terms (H2). All else being equal, candidates are more distant from their party electorate if they belong to a larger party, while they are closer to their voters if they are members of a party with a longer electoral history (with the exception of Model 3). Moreover, candidates are farther away from the voters of their party if the latter is more internally heterogeneous. In addition, candidates of anti-establishment parties are more incongruent with the party voters than candidates of mainstream parties (again with the exception of Model 3).¹² Arguably, anti-elite parties may well be closer to their voters on policy dimensions other than the left–right. As expected, Table 1 also shows that candidates belonging to more right-wing parties are less distant from their party voters, while candidates belonging to more leftist parties are farther away from their party electorate. Furthermore, incongruence is more pronounced for candidates whose party is more ideologically extreme.

In regard to election-level predictors, our results only partially support H3. As the negative and statistically significant coefficients on Polarization indicate, higher differentiation among parties on the left–right scale reduces the distance between candidates and their electorate, which resonates with the literature on ideological congruence. At the same time, however, the disproportionality of the electoral system has no impact on ideological proximity gauged at the level of individual candidates. Finally, the election-level control variable that we introduced does not affect incongruence: candidates running for office in older democracies are not closer to their party voters.

Interactive models

We now explore a few cross-level interactions between factors at different levels of analysis. To begin with, we assess whether the effects of individual determinants—a long record of political experience and representative styles—on ideological incongruence/congruence, are moderated by a few key party properties: size, age, internal heterogeneity and non-mainstream nature. Our argument is that the effects of the individual attributes that increase the congruence (i.e. decrease the distance) with their voters are altered by specific party characteristics. In particular, hypothesis H4 states that ideological proximity increases for candidates with a partisan style of representation and with a long record of political experience, but this does not happen if they belong to a party that is large, brand-new, internally heterogeneous or anti-establishment. To test this hypothesis, in each model we include a cross-level interaction term. Table 2 reports the most relevant regression results, and Figures 2 and 3 display the significant interaction effects.

Table 2.

Determinants of candidate–voters ideological incongruence in Europe. Interactions between candidate-level and party-level predictors.

	(5)	(6)	(7)	(8)
	Interaction: N. Elections X Party Age	Interaction: N. Elections X Anti-Establ.	Interaction: Style X Party Size	Interaction: Style X Anti-Establ.
Political Experience	0.021	−0.019**	−0.011	−0.010
	(0.019)	(0.008)	(0.008)	(0.008)
Style: Trustee	0.012	0.012	−0.046**	0.013
	(0.021)	(0.021)	(0.020)	(0.021)
Style: Delegate	0.055**	0.055**	0.033	0.064***
	(0.024)	(0.024)	(0.030)	(0.022)
Party Size	0.009***	0.009***	0.008***	0.009***
	(0.003)	(0.003)	(0.003)	(0.003)
Party Age	−0.003**	−0.004**	−0.004**	−0.003**
	(0.001)	(0.002)	(0.002)	(0.002)
Anti-Establishment	0.133**	0.088	0.132**	0.150**
	(0.058)	(0.059)	(0.057)	(0.062)
Political Experience X Party Age	−0.001*
	(0.000)
Political Experience X Anti-Establishment		0.052**
		(0.025)
Trustee X Party Size			0.004**
			(0.001)
Delegate X Party Size			0.001
			(0.002)
Trustee X Anti-Establishment				−0.008
				(0.053)
Delegate X Anti-Establishment				−0.045
				(0.064)
Polarization	−0.084**	−0.086**	−0.085**	−0.086**
	(0.034)	(0.034)	(0.034)	(0.034)
Disproportionality	−0.010	−0.010	−0.011	−0.01
	(0.008)	(0.008)	(0.008)	(0.008)
Women	0.040*	0.039*	0.039*	0.039*
	(0.023)	(0.023)	(0.023)	(0.023)
Age	−0.018***	−0.018***	−0.018***	−0.018***
	(0.004)	(0.004)	(0.005)	(0.005)
Age X Age	0.000***	0.000***	0.000***	0.000***
	(0.000)	(0.000)	(0.000)	(0.000)
Education	−0.031***	−0.032***	−0.032***	−0.032***
	(0.010)	(0.010)	(0.010)	(0.010)
Party LR	−0.048***	−0.048***	−0.048***	−0.048***
	(0.012)	(0.012)	(0.012)	(0.012)
Age of Democracy	0.000	0.000	0.000	0.000
	(0.001)	(0.001)	(0.001)	(0.001)
Constant	2.372***	2.421***	2.448***	2.415***
	(0.204)	(0.199)	(0.199)	(0.199)
Variance component:
Election	0.008***	0.008***	0.009***	0.009***
	(0.003)	(0.003)	(0.003)	(0.003)
Party	0.099***	0.101***	0.100***	0.099***
	(0.013)	(0.014)	(0.013)	(0.013)
Residual	0.921	0.920*	0.921*	0.921
	(0.023)	(0.023)	(0.023)	(0.023)

N	14535	14535	14535	14535
Log-likelihood	−20224.07	−20222.39	−20223.44	−20225.61
AIC	40486.14	40482.77	40486.875	40491.211
BIC	40630.25	40626.87	40638.562	40642.898

Figure 2.

Impact of candidates’ past experience on candidate–voters incongruence, by party age and type. Notes: Marginal effects estimated from Model 5 (upper panel) and Model 6 (lower panel). In each graph, the black dots indicate the impact of Political Experience on incongruence at a given value of the party-level variable represented on the horizontal axis. Whiskers represent 95% confidence intervals. The superimposed histogram displays the distribution of candidates according to the party-level variable.

Figure 3.

Impact of candidates’ representational styles on candidate–voters incongruence, by party size and type. Notes: Marginal effects estimated from Model 7 (upper panel) and Model 8 (lower panel). In each graph, the black dots indicate the impact of being a delegate-type (as opposed to a partisan-type) of candidate on incongruence at a given value of the party-level variable represented on the horizontal axis. Whiskers represent 95% confidence intervals. The superimposed histogram displays the distribution of candidates according to the party-level variable.

The analysis offers some empirical support for H4. Specifically, candidates’ electoral history interacts with two party properties (age and mainstream nature), while representational styles interact with party type and (albeit to a limited extent) with party size. The two graphs in Figure 2 display the expected marginal impact of candidates’ record of political experience on the degree of incongruence with their party electorate, as estimated for different values of party age and party type. In particular, the upper panel reveals that a candidate’s longer political experience does reduce incongruence with her/his voters when her/his party is older than 50—a condition that applies to about 50% of the observations in our sample. Similarly, candidates’ experience decreases incongruence in the case of mainstream parties, but not in the case of anti-establishment parties (lower panel). This implies that a candidate seems better able to identify accurately the ideological stances of her/his party voters only if she/he is member of a non-anti-establishment party or of a relatively old party. The impact of political experience on incongruence is instead undiscernible from zero for any percentage of votes that a party won and for any level of party cohesiveness (see Appendix 7).

The marginal effects plotted in Figure 3 lend some support to the expectation that the impact of representational styles on incongruence is conditional on party-level traits. From visual inspection of the upper graph it appears that being a delegate rather than a partisan-type of candidate increases incongruence only for small and medium-sized parties (i.e. between 10% and 25% of votes), which are about 30% of the observations. In other words, while styles of representation do not matter for large parties (as hypothesized), they do not seem to matter either for very small parties. As to party type, the lower graph supports our expectation: compared to partisans, delegates are more distant from their party electorate if they are members of a mainstream party, but not if they belong to an anti-establishment party. At the same time, results concerning the interaction of styles of representation with party age and heterogeneity do not clearly support our expectations (see Appendix 7).¹³

Turning to the possible interactions between individual-level and election-level variables, Table 3 reports analyses showing the extent to which the effects of candidates’ styles of representation on incongruence are moderated by the degree of polarization in a given election and the disproportionality of the electoral rules. We conjectured that those individual attributes that reduce ideological incongruence between a candidate and the party voters may have a weaker effect when left–right divergence is sizable and when the electoral system has strong proportional effects (H5). The marginal effects of the two interactions are illustrated graphically in Figure 4.¹⁴

Table 3.

Determinants of candidate–voters ideological incongruence in Europe. Interactions between candidate-level and election-level predictors.

	(9)	(10)
	Interaction: Style X Polarization	Interaction: Style X Disproportionality
Political Experience	−0.010	−0.010
	(0.008)	(0.008)
Style: Trustee	−0.042	−0.023
	(0.067)	(0.030)
Style: Delegate	0.089	0.043
	(0.118)	(0.036)
Party Size	0.009***	0.009***
	(0.003)	(0.003)
Party Age	−0.003**	−0.003**
	(0.001)	(0.002)
Anti-Establishment	0.136**	0.133**
	(0.058)	(0.058)
Polarization	−0.089**	−0.085**
	(0.038)	(0.035)
Disproportionality	−0.010	−0.014*
	(0.008)	(0.007)
Trustee X Polarization	0.018
	(0.023)
Delegate X Polarization	−0.011
	(0.036)
Trustee X Disproportionality		0.007 (0.005)
Delegate X Disproportionality		0.003 (0.006)
Women	0.039*	0.039*
	(0.023)	(0.023)
Age	−0.018***	−0.018***
	(0.005)	(0.005)
Age X Age	0.000***	0.000***
	(0.000)	(0.000)
Education	−0.032***	−0.032***
	(0.010)	(0.010)
Party LR	−0.048***	−0.048***
	(0.012)	(0.012)
Age of Democracy	0.000	0.000
	(0.001)	(0.001)
Constant	2.423***	2.435***
	(0.211)	(0.200)
Variance component:
Election	0.008***	0.009***
	(0.003)	(0.003)
Party	0.099***	0.099***
	(0.013)	(0.013)
Residual	0.921	0.921*
	(0.023)	(0.023)

N	14535	14535
Log-likelihood	−20225.38	−20224.72
AIC	40490.75	40489.43
BIC	40642.44	40641.12

Figure 4.

Impact of candidates’ representational styles on candidate–voters incongruence, by levels of ideological polarization. Notes: Marginal effects estimated from Model 9. The black dots indicate the impact of being a delegate-type (as opposed to a partisan-type) of candidate on incongruence at a given value of the ideological polarization. Whiskers represent 95% confidence intervals. The superimposed histogram displays the distribution of candidates according to polarization.

Figure 4 seems to lend partial support to the idea that the impact of a partisan style of representation in reducing incongruence is stronger when polarization is low, and it is weaker when polarization is high. Close inspection of the graph reveals that styles of representation matter not for the entire range of the polarization index, but “only” for polarization scores between 2.5 and 3.5, an interval that represents as much as 60% of the observations in our sample. In this large subsample of the data, the impact of being delegate rather than partisan increases incongruence, and this effect has the hypothesized negative slope. Put differently, if we exclude elections with an extremely low polarization, we observe what we expected: as left–right polarization grows, the impact of a key individual attribute such as the style of representation becomes weaker, approaching zero above a certain threshold. At the same time, the difference between trustees and partisans (shown in Appendix 8) is never significant.

The findings appear less clear when examining the interaction between styles of representation and the disproportionality of the electoral system (Model 10). The impact of representational styles becomes weaker when disproportionality is lower. However, the estimated effects are never significant, the only exception being those national elections with disproportionality scores between 4 and 9. In these cases, delegate-type candidates are more incongruent with their voters than their colleagues adopting a partisan style. To be noted is that these elections represent less than 20% of the observations—that is, a very small subsample of our data (see Appendix 8).

Closing remarks

Democratic representation, understood as preference congruence between voters and members of the political elites, has seldom been investigated from the perspective of individual politicians. When such investigation has been undertaken, the focus has been almost entirely on the US Congress, and the analysis has typically involved the relation between elected representatives and their geographical constituency in a context in which parties on the ground were not structured the way they were in Europe. Here, parties prominent in the representational linkage nurtured a more party-centered tradition of studies, by discouraging research on congruence from adopting the point of view of individual representatives. This article helps fill this gap by exploring the degree of ideological congruence between candidates running for a legislative seat and their party electorate. In particular, it has evaluated how ideological proximity between a candidate and the voters of her/his party is shaped by factors measured at three levels of analysis: the candidate level, the party level, and the election level. Used for this purpose have been data on candidates and voters in 39 national elections held across Europe in the last 15 years.

Given the goal of this article and the nature of the data used, we are not in a position to assess whether European parties have become weaker or stronger if compared to the past. However, we can argue that individual candidates’ attributes do affect ideological congruence with voters. Our analyses reveal the prominent role of representational styles. Those who perceive their role according to a strictly partisan logic are closer to their party voters than those who think that representatives should behave as delegates of their geographical constituency. Many of our findings regarding party-level and election-level determinants are consistent with extant congruence scholarship adopting a voter- or party-perspective. Our analyses highlight that congruence is greater for smaller, older, and mainstream parties. These types of parties are particularly able to bring together voters and candidates with similar policy preferences. Although our data do not enable us to assess whether party cohesiveness is declining, we also found that candidates and voters are farther apart when parties are internally divided. Moreover, congruence is enhanced by the overall ideological polarization in the party system, while electoral disproportionality fails to show any impact.

The most interesting results reported by this article concern the interactions between individual sources of congruence and factors measured at the level of parties and at higher levels. Notably, candidates’ role orientations matter especially for certain types of parties. A closer match with voters is observed in the case of partisans (vis-à-vis delegates) not for any type of party, but only for non-anti-establishment parties and for small and medium-sized parties. We also discovered that a politician’s past electoral experience, which does not seem to matter per se, improves congruence with voters when she/he runs for office in a mainstream and relatively old party. We also found that the internal ideological heterogeneity of parties, however, does not clearly moderate the effects of individual characteristics. The findings are more mixed as regards the relationships between candidates’ traits and election-level characteristics. On the one hand, the disproportionality of electoral rules does not alter the impact of candidates’ attributes. On the other hand, we found some evidence that the impact of a partisan style of representation in reducing incongruence is stronger when ideological polarization is lower, and it is weaker when polarization is high.

Finally, our findings bring us to an issue that has been only marginally addressed in this article, namely the key role of the left–right axis in affecting the quality of representation in terms of ideological congruence. If candidates are ideologically closer to their voters when party divergence on the left–right is large, one could expect that also the representatives, not only for the represented, benefit from party distinctiveness. In other words, in the presence of polarization the categories of left and right seem to work as powerful heuristics not only for voters, but for politicians as well. As a consequence, to the extent that political elites seek congruence with voters, they may have incentives to differentiate party programs along the general left–right. This is a question that undoubtedly warrants further empirical analysis.

Supplemental material

Supplemental Material, sj-pdf-1-ppq-10.1177_1354068820968091 - Responsiveness when parties are “weak”: A candidate-based analysis of voter-party congruence in Europe

Supplemental Material, sj-pdf-1-ppq-10.1177_1354068820968091 for Responsiveness when parties are “weak”: A candidate-based analysis of voter-party congruence in Europe by Andrea Pedrazzani and Paolo Segatti in Party Politics

Footnotes

Acknowledgements

Earlier versions of this article were presented at the 12th General Conference of the European Consortium for Political Research (Hamburg, 22-25 August 2018), at the 32nd Congress of the Italian Political Science Association (Turin, 6-8 September 2018), at the 47th Joint Sessions of Workshops of the European Consortium for Political Research (Mons, 8-12 April 2019), and at the 26th International Conference of Europeanists (Madrid, 20-22 June 2019). The authors thank all participants for discussing the article. The authors also thank two anonymous referees for their helpful comments and suggestions.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Andrea Pedrazzani

Paolo Segatti

Supplemental material

Supplemental material for this article is available online.

Notes

References

Abedi

(2004) Anti-Political Establishment Parties: A Comparative Analysis. London: Routledge.

Adam

Maier

(2010) Personalization of politics: a critical review and agenda for research. Annals of the International Communication Association 34(1): 213–257.

Andeweg

(2011) Approaching perfect policy congruence measurement, development, and relevance for political representation. In: Rosema

Denters

Aarts

(eds) How Democracy Works. Political Representation and Policy Congruence in Modern Societies. Amsterdam: Pallas Publications–Amsterdam University Press, pp. 39–52.

Andeweg

Thomassen

(2005) Modes of political representation: toward a new typology. Legislative Studies Quarterly 30(4): 507–528.

Bakker

Edwards

Hooghe

. et al. (2015) 2014 Chapel Hill Expert Survey, Version 2015.1. Available at: www.chesdata.eu (accessed 24 February 2020).

Barnes

(1977) Representation in Italy: Institutionalized Tradition and Electoral Choice. Chicago, IL: University of Chicago Press.

Belchior

(2013) Explaining left–right party congruence across European party systems: a test of micro-, meso-, and macro-level models. Comparative Political Studies 46(3): 352–386.

Blais

Bodet

(2006) Does proportional representation foster closer congruence between citizens and policymakers? Comparative Political Studies 39(10): 1243–1262.

Clausen

Holmberg

DeHaven-Smith

(1983) Contextual factors in the accuracy of leader perceptions of constituents’ views. Journal of Politics 45(2): 449–472.

10.

Converse

Pierce

(1986) Political Representation in France. Cambridge, MA: Harvard University Press.

11.

Costa Lobo

Curtice

(eds) (2015) Personality Politics? The Role of Leader Evaluations in Democratic Elections. Oxford: Oxford University Press.

12.

Costello

Thomassen

Rosema

(2012) European parliament elections and political representation: policy congruence between voters and parties. West European Politics 35(6): 1226–1248.

13.

Cotta

Best

(eds) (2008) Democratic Representation in Europe. Diversity, Change and Convergence. Oxford: Oxford University Press.

14.

Cutler

(2002) The simplest shortcut of all: sociodemographic characteristics and electoral choice. Journal of Politics 64(2): 466–490.

15.

Dalton

(1985) Political parties and political representation: party supporters and party elites in nine nations. Comparative Political Studies 18(3): 267–299.

16.

Dalton

Farrell

McAllister

(2011) Political Parties and Democratic Linkage: How Parties Organize Democracy. Oxford: Oxford University Press.

17.

De Winter

Karlsen

Schmitt

(2020) Introduction. Candidates, Parties and Voters in Representative Democracies. In: De Winter

Karlsen

Schmitt

(eds) Parliamentary Candidates between Voters and Parties. A Comparative Perspective. London: Routledge.

18.

Di Virgilio

Segatti

(eds) (2016) La rappresentanza Politica in Italia. Bologna: Il Mulino.

19.

Dolný

Baboš

(2015) Voter-representative congruence in Europe: A loss of institutional influence? West European Politics 38(6): 1274–1304.

20.

Downs

(1957) An Economic Theory of Democracy. New York, NY: Harper.

21.

Enyedi

Pedrazzani

Segatti

(2020) Policy Representation in Europe. In: De Winter

Karlsen

Schmitt

(eds) Parliamentary Candidates Between Voters and Parties. A Comparative Perspective. London: Routledge, pp. 164–195.

22.

Eulau

Karps

(1977) The puzzle of representation: specifying components of responsiveness. Legislative Studies Quarterly 2(3): 233–254.

23.

Eulau

Wahlke

Buchanan

. et al. (1959) The role of the representative: some empirical observations on the theory of Edmund burke. American Political Science Review 53(3): 742–756.

24.

Ezrow

(2010) Linking Citizens and Parties: How Electoral Systems Matter for Political Representation. Oxford: Oxford University Press.

25.

Gallagher

(1991) Proportionality, disproportionality and electoral systems. Electoral Studies 10(1): 33–51.

26.

Giebler

Wessels

(2013) Whom to Represent? The Role of Individual Characteristics as Well as Institutional and Organizational Incentives. In: ECPR General Conference, Bordeaux, France, 4–7 September 2013.

27.

Golder

Lloyd

(2014) Re-evaluating the relationship between electoral rules and ideological congruence. European Journal of Political Research 53(1): 200–212.

28.

Golder

Stramski

(2010) Ideological congruence and electoral institutions. American Journal of Political Science 54(1): 90–106.

29.

Irwin

Thomassen

(1975) Issue-consensus in a multi-party system: voters and leaders in the Netherlands. Acta Politica 10(4): 389–420.

30.

Kim

Powell

Fording

(2010) Electoral systems, party systems, and ideological representation: an analysis of distortion in western democracies. Comparative Politics 42(2): 167–185.

31.

Kitschelt

Mansfeldova

Markowski

. et al. (1999) Post-Communist Party Systems: Competition, Representation and Inter-Party Cooperation. New York, NY: Cambridge University Press.

32.

Lupu

Selios

Warner

(2017) A new measure of congruence: the earth mover’s distance. Political Analysis 25(1): 95–113.

33.

Mattila

Raunio

(2006) Cautious voters—supportive parties: opinion congruence between voters and parties on the EU dimension. European Union Politics 7(4): 427–449.

34.

McAllister

(2007) The personalization of politics. In: Dalton

Klingemann

H-D

(eds) The Oxford Handbook of Political Behavior. Oxford: Oxford University Press, pp. 571–588.

35.

McEvoy

(2012) Unequal representation in the EU: a multilevel analysis of voter–party congruence in EP elections. Representation 48(1): 83–99.

36.

Miller

Stokes

(1963) Constituency influence in congress. American Political Science Review 57(1): 165–177.

37.

Miller

Pierce

Thomassen

. et al. (1999) Policy Representation in Western Democracies. Oxford: Oxford University Press.

38.

Norris

(2005) Recruitment. In: Katz

Crotty

(eds) The Handbook of Party Politics. London: Sage, pp. 89–108.

39.

Önnudóttir

(2014) Policy congruence and style of representation: party voters and political parties. West European Politics 37(3): 538–563.

40.

Önnudóttir

Von Schoultz

(2020) Candidates’ representational roles. In: De Winter

Karlsen

Schmitt

(eds) Parliamentary Candidates Between Voters and Parties. A Comparative Perspective. London: Routledge, pp. 120–141.

41.

Pitkin

(1967) The Concept of Representation. Berkeley, CA: University of California Press.

42.

Powell

(2000) Elections as Instruments of Democracy. New Haven: Yale University Press.

43.

Powell

(2011) Party polarization and the ideological congruence of governments. In: Dalton

(ed) Citizens, Context and Choice. Oxford: Oxford University Press, pp. 197–213.

44.

Przeworski

Sprague

(1986) Paper Stones: A History of Electoral Socialism. Chicago, IL: University of Chicago Press.

45.

Rahat

Kenig

(2018) From Party Politics to Personalized Politics? Party Change and Political Personalization in Democracies. Oxford: Oxford University Press.

46.

Rohrschneider

Whitefield

(2012) The Strain of Representation: How Parties Represent Diverse Voters in Western and Eastern Europe. Oxford: Oxford University Press.

47.

Schmitt

Thomassen

(eds) (1999) Political Representation and Legitimacy in the European Union. Oxford: Oxford University Press.

48.

Shugart

(2001) Electoral “efficiency” and the move to mixed-member systems. Electoral Studies 20(2): 173–193.

49.

Shugart

Valdini

Suominen

(2005) Looking for locals: voter information demands and personal vote-earning attributes of legislators under proportional representation. American Journal of Political Science 49(2): 437–449.

50.

Strøm

(2012) Roles as strategies: towards a logic of legislative behaviour. In: Blomgren

Rozenberg

(eds) Parliamentary Roles in Modern Legislatures. Abingdon: Routledge, pp. 85–100.

51.

Thomassen

(1994) Empirical research into political representation. In: Jennings

Mann

(eds) Elections at Home and Abroad. Ann Arbor: University of Michigan Press, pp. 237–264.

52.

Van Aelst

Sheafer

Hubé

. et al. (2016) Personalization. In: De Vreese

Esser

Hopmann

(eds) Where’s the Good News? Comparing Political Journalism in 16 Countries. London: Routledge, pp. 112–130.

53.

Van Biezen

Mair

Poguntke

(2012) Going, going,…gone? The decline of party membership in contemporary Europe. European Journal of Political Research 51(1): 24–56.

54.

Van der Brug

Fennema

(2007) Causes of voting for the radical right. International Journal of Public Opinion Research 19(4): 474–487.

55.

Van der Eijk

Schmitt

Binder

(2005) Left-right orientations and party choice. In: Thomassen

(ed) The European Voter: A Comparative Study of Modern Democracies, Oxford: Oxford University Press, pp. 167–191.

56.

Van Hauwaert

Van Kessel

(2018) Beyond protest and discontent: a cross-national analysis of the effect of populist attitudes and issue positions on populist party support. European Journal of Political Research 57(1): 68–92.

57.

Vandeleene

De Winter

Baudewyns

(eds) (2019) Candidates, Parties and Voters in the Belgian Partitocracy. Cham: Palgrave MacMillan.

58.

Walczak

Van der Brug

(2013) Representation in the European parliament: factors affecting the attitude congruence of voters and candidates in the EP elections. European Union Politics 14(1): 3–22.

59.

Warwick

(2016) The ideological congruence illusion: the impact of valence. Legislative Studies Quarterly 41(2): 445–469.

60.

Weber

Franklin

(2018) A behavioral theory of electoral structure. Political Behavior 40(4): 831–856.

61.

Wessels

(1999) System characteristics matter: empirical evidence from ten representation studies. In: Miller

Pierce

Thomassen

. et al. (eds) Policy Representation in Western Democracies. Oxford: Oxford University Press, pp. 137–161.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.02 MB