The effect of candidate electoral experience on ballot placement in list proportional representation systems

Abstract

This article investigates how candidate experience in previous elections, measured by the number of prior candidacies, affects ballot placement in list PR systems. The central argument is that selectorates find experienced candidates attractive because of proven loyalty and higher voter recognition. Moreover, prior candidacies strengthen candidates’ familiarity with intra-party selection processes. Returning candidates, however, also risk becoming victims of a seniority trap: after not succeeding to access realistic positions after a number of elections, additional candidacies become harmful as selectorates lose interest. Using data on 18,393 candidacies by 9,905 candidates of parties in Flanders (Belgium) (1987–2019), a set of logit and Cox models test the hypothesised curvilinear effect of prior candidacies on access to realistic positions. Both approaches lead to similar outcomes, confirming that experience positively affects ballot placement but with decreasing marginal utility. These findings have important implications for career strategies and contribute to knowledge on nomination processes and candidate turnover.

Keywords:

• Electoral experience
• Belgium
• candidate selection
• Cox regression
• list PR systems

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 We do not include the 1989 election for the Parliament of the Brussels Capital Region, as this regional election took place before the fourth state reform (1993), which turned Belgium into a full-fledged federal state with directly elected regional assemblies. All regional elections involving Flemish parties after this state reform are included in the analysis.

2 We unfortunately have no access to a similar longitudinal data set on the careers of francophone Belgian candidates.

3 Successor candidates form a separate list of candidates who will replace the effectively elected representatives in the event of resignation, death or illness. Similar to intra-party seat allocation amongst effective candidates, successors are ordered on the basis of ballot position and preference votes.

4 In the literature on campaign spending effects in list PR systems, several authors have excluded irrelevant challengers, low quality candidates or ‘list fillers’ from the analysis as these candidates would exaggerate the differences between incumbents and challengers in terms of spending (Green and Krasno, 1988; Samuels, 2001). These studies assume that list fillers only invest very moderately in the campaign, as their candidacy is merely symbolical.

5 Alternatively, in a second approach we consider all subsequent candidacies by a candidate that at some point switched parties as ‘party-switching’, to assess whether the label of party-switching affects nomination prospects throughout the remaining electoral career.

6 In some very exceptional cases, candidates are in fact sitting incumbent at the time of their first participation in Belgian parliamentary elections: the category of co-opted and provincial Senators, did do not need to run for election but were instead appointed by the language groups in the Senate and the provincial councils respectively; members of federal or regional government that were appointed by their parties without having run for election previously. This explains why the percentage of challengers amongst the candidates with zero prior candidacies (see Figure 1) does not equal 100%.

7 As regards party switching, the alternative operationalization of considering the switching as a ‘sticky’ characteristic for candidates did not yield any significant results either.

8 The flexibility of this semi-parametric approach is an advantage to check for the hypothesize curvilinear effect of prior candidacies. For the sake of robustness, however, we also show the results following three alternative modelling choices, of which all lead to similar results. First, we run a parametric model with a lognormal baseline distribution (Table J). The latter is appropriate given that it allows for hazard functions that are first increasing to a maximum and then decreasing. A plot of the baseline hazard is also included in the Online appendix (Figure A). Second, we display the results from an accelerated failure time (AFT) model (Table K). As we specify a polynomial term, we opt for an AFT model with normally distributed errors. Third, a piecewise constant hazards (PCH) model was estimated where each election year is considered a cut-off point to define the intervals (Table L).

Additional information

Notes on contributors

Gert-Jan Put

Gert-Jan Put is an FNRS postdoctoral researcher at Université de Namur. His research interests include candidate selection, intra-party competition and electoral systems. He has published in Political Behavior, Political Geography and Party Politics, amongst others.

Gertjan Muyters

Gertjan Muyters is a doctoral candidate at the KU Leuven Public Governance Institute. His research focuses on candidate turnover and political careers.

Bart Maddens

Bart Maddens is Professor of Political Science at the KU Leuven Public Governance Institute His research interests include political finance, elections and multi-level systems. His articles have appeared in journals such as West European Politics, Party Politics and Electoral Studies, amongst others.