Perceptual Accuracy of Elected Legislators

• In Event: Perceptions and Bias

Abstract

Policy congruence—the fact that policies match public preferences—is one of the cornerstones of democratic quality. It can essentially be produced in two ways: (1) voters elect the ‘right’ representatives (representatives holding preference that are similar to their own) and these then act following their own preferences; (2) elected representatives forming themselves a perception of the public’s preferences and then acting on those perceptions (Miller and Stokes 1963). Of course, elected politicians may deliberately opt not to directly follow the public’s will (see the trustee role that MPs may deliberately adopt). But in many cases, representation supposes some knowledge of what the public cares about and wants its representatives to do. Consequentially, this paper analyzes the latter mechanism. It examines individual politicians’ perceptions of public opinion with regard to policies and, more concretely, it focuses on the accuracy of these perceptions.
A small and older body of work specifically tackled elite’s perceptions of voters (Hedlund & Friesema, 1972; Clausen, 1977; Clausen, Holmberg, & deHaven-Smith, 1983; Granberg, 1987; Granberg & Holmberg, 2002; Holmberg, 1999). Notwithstanding a single study in the 2000s (Norris & Lovenduski, 2004), only recently the perceptual path of democratic linkage has been rediscovered in studies by Belchior (2014) in Portugal and by Broockman and Skovron (2015) in the US. The key finding is the reaffirmation of the fact that elites’ perceptions of public opinion are not very accurate. They are biased by "wishful thinking" (e.g. Holmberg, 1999). Uninformed about the actual opinions of voters, MPs rely on their own opinion about an issue and project it on voters. This work has also found, in line with Miller & Stokes’ (1963) initial work, that elites’ success in making sense of public opinion is unevenly distributed across issues. When issues are salient or ‘politically charged’ (Hedlund & Friesema, 1972) or when the public has homogeneous ideas (Clausen et al., 1983) MPs’ estimates are more accurate. Third, political context plays a role: elites’ estimations are more accurate in systems that promote direct legislator-voter links (Clausen et al., 1983). Fourth, and most importantly for our quest, the previous studies have not been successful in explaining the often substantial differences among representatives (Clausen et al., 1983; Holmberg, 2007). Still, some studies did find that individual differences in accurately estimating the public’s wishes are correlated with ideology (Belchior, 2014). In an impressive recent study, Broockman and Skovron (2015) show that, among US state legislature candidates, conservative candidates hold less accurate estimations compared to their more liberal counterparts. Yet, they do not offer any theoretical explanation for this empirical result.
This paper therefore addresses two research questions: (1) How accurate are politicians’ assessments of public opinion? (2) How can differences in perceptual accuracy across elected representatives, across parties and across policy issues. Based on existing research, we have several expectations. For example, legislators who perceive their seat to be more unsafe are expected to be more invested into learning about public opinion. Yet, the party or interest group affiliation might play a role as well as does the polarization of public opinion. In sum, it is plausible that there is a tension between factual information about what the general public and the party’s electorate wants on the one hand and the own preferences an MP holds on the other hand.
To address our research questions, we draw on unique evidence from elected representatives and the general population from four different political systems: Wallonian Belgium, Flanders Belgium, Switzerland, Netherlands and Germany. In these, we have surveyed 800 active legislators face-to-face. The core of the survey consists of several statements for which we ask legislators to disclose their personal opinion and estimate the distribution of public opinion. We pair these data with additional information. First, to compare legislators' estimations with the actual public opinion, we add information from an independently held population survey on the exact same statements we asked legislators. Next, we add background information about each individual legislator such as their political experience, their committee membership or how they have been elected to the data. The resulting data allow us to address one of the most fundamental puzzles about policy congruence.

Authors

• Luzia Helfer, Geneva University, Political Science and International Relations

• Stefaan Walgrave, University of Antwerp

• Nathalie Brack, Université libre de Bruxelles

• Jean-Benoit Pilet, Universite Libre de Bruxelles

• Frederic Varone