The Runner-Up Effect

Santosh Anagol, Thomas Fujiwara
2016 Journal of Political Economy  
Exploiting regression discontinuity designs in Brazilian, Indian, and Canadian first-past-the-post elections, we document that second-place candidates are substantially more likely than close third-place candidates to run in, and win, subsequent elections. Since both candidates lost the election and had similar electoral performance, this is the effect of being labeled the runner-up. Selection into candidacy is unlikely to explain the effect on winning subsequent elections, and we find no
more » ... nd we find no effect of finishing in third place versus fourth place. We develop a simple model of strategic coordination by voters that rationalizes the results and provides further predictions that are supported by the data. All use subject to University of Chicago Press Terms and Conditions ( that the runner-up effect (on winning) is stronger in cases in which the second-and third-place candidates received a large number of votes, compared to the winner. Fourth, we also find larger effects on winning future elections immediately after the 1975-77 "emergency" in India; this is a period in which it is plausible that the number of voters acting strategically was higher relative to other periods. Finally, the model is particularly suited for analyzing cases in which the second-and third-place candidates are similar in voters' perception, and consistent with this, we find that our effects are stronger when second-and third-place candidates are from parties with similar platforms. Another (nonexclusive) possible mechanism behind the runner-up effect is that at least one political player (e.g., voters, candidates, parties, the media) evaluates candidates on the basis of their rankings, even though rankings provide no additional information beyond the underlying vote shares. In other words, they engage in a "rank heuristic." 5 For example, parties might use a heuristic in which, at least in some cases, they provide more support for runners-up. Another possibility is that candidates perceive that, even holding electoral performance constant, they were closer to winning after coming in second instead of third place and are therefore more motivated to invest in future campaigns. We discuss the possibility of such heuristics driving the effect. In particular, we note that it is not clear how an explanation based on heuristics could explain all the additional predictions from the coordination model that find support in the data. We also present tests of a heuristic-based explanation in relation to the media. One possibility is that the media give more coverage to runnersup than to third-place candidates. We do not find evidence for this in the Canadian context, where comprehensive newspaper archive searches for candidate names are possible. Furthermore, we do not find that the runner-up effect is larger in Brazilian and Indian regions with greater media presence. We reiterate that heuristic behavior and strategic coordination are not mutually exclusive explanations. They are likely mutually reinforcing (e.g., if runners-up are more likely to be coordinated on, they should be motivated by rank). We also emphasize that while our results are consistent with coordination by voters, it is possible that this coordination (that shifts vote shares) also occurs at a different level. Candidates, parties, and/or other "elites" may coordinate their support and rely on election rankings 5 An individual who observes ranks but not vote shares should (rationally) infer that any runner-up received substantially more votes than a third-place candidate. However, and more relevantly for our purposes, an agent observing only ranks implies that some other agent (e.g., the media) acted following (or imposing) a rank heuristic by deciding to supply only the coarser information. For examples of rank heuristics in other contexts, see Pope (2009) and Barankay (2012). runner-up effect journal of political economy
doi:10.1086/686746 fatcat:vdvt3fush5aspkwzubgjx3a2ea