Why threaten to drop out of a presidential election you are likely to win?

(Joshua Tucker: Continuing our series of Election Reports, we are pleased to welcome the following pre-election report on the upcoming Oct. 27 Georgian elections from Bret Barrowman, a PhD candidate in political science at George Washington University who is conducting dissertation research in Tbilisi, Georgia. You can follow him on Twitter @bbarrowm.)
*****
With current President Mikhail Saakasvhili constitutionally ineligible for a third term,  Georgian voters will replace him as president Oct. 27  for the first time since the Rose Revolution almost a decade ago. The candidate of Saakashvili’s United National Movement (UNM), Davit Bakradze, trails Giorgi Margvelashvili, who is backed by Prime Minister Bidzina Ivanishvili’s Georgian Dream (GD) coalition, by a large margin in the most recent poll.
Despite this popularity gap, the presidential election cycle has been characterized by uncertainty stemming from three factors: the two-round voting system, public opinion polls with high non-response rates, and Ivanishvili’s intention to withdraw from politics if his preferred candidate is elected. Furthermore, Margvelashvili, at Ivanishvili’s suggestion, recently announced his intention to withdraw from the race if he does not win outright in the first round.
The leading presidential candidate’s threat to abandon the race presents an interesting puzzle — why would Margvelashvili give up when all available information indicates that he is the favorite even if he doesn’t win outright? The context of Georgia’s highly personalized political system provides a couple of plausible explanations. First, Margvelashvili’s announcement to withdraw could be cheap talk intended to mobilize GD supporters for the first round. Alternatively, the announcement might not be strategic at all, but rather a reflection of confidence in the GD camp that Margvelashvili can pick up enough undecided voters to avoid a second round.
Background
Georgia’s president is directly elected using a two-round format. If no candidate receives a majority of votes in the first round, the two highest-polling candidates will compete in a runoff. Of 23 candidates contesting the election, there are three leading contenders. The latest poll commissioned by the National Democratic Institute shows Margvelashvili leading as the preferred candidate of 39 percent of respondents, less than the majority necessary to avoid a second-round contest. Bakradze and former Parliament speaker Nino Burjanadze followed with 18 percent and 7 percent, respectively. A combined 26 percent of respondents were undecided, did not respond or answered “no candidate,” making predictions of the election results difficult. Barring, however, some dramatic revelation such as last year’s prison abuse scandal that drove the nail in the coffin of UNM’s parliamentary majority, there is little reason to doubt that Margvelashvili is the favorite.
Like many post-Soviet states, Georgia’s political system remains highly clientelistic. Elections are won not by selling distinct policy platforms to voters but through campaigns centered on personalities and the mobilization of personal networks through a combination of coercion and the distribution of resources. Although incumbents do not appear to engage in outright electoral fraud, over the past two election cycles, both GD and UNM have allegedly attempted to induce voters with coercion, or through the distribution of private and club goods. Several NGOs have documented instances of both public- and private-sector employees fired for political opposition, high turnover in central and local public sector employees following the GD takeover of the parliamentary majority in the 2012 election, apparently arbitrary bonuses paid to civil servants, and outright violence or threats of violence against supporters of political opponents. At a broader level, politics continues to be characterized more by personal animosity than by competition between policy platforms.
This clientelistic nature of Georgian politics is particularly important for this election cycle because of an important upcoming change in formal political institutions. Upon the inauguration of the next president, a set of constitutional changes will take effect, transferring several important functions from the office of the president to that of the prime minister. Most importantly, the prime minister and the government will become responsible for all domestic and foreign policy; appointment of the defense and interior ministers, along with provincial governors, will transfer to the prime minister or government from the president; and the government and parliament will become responsible for the state budget. As a result, the office of the president will become largely symbolic. This shift in formal institutional power takes on additional importance, given Ivanishvili’s intention to resign his position if his preferred presidential candidate wins. As a result, the newly empowered premiership will be occupied by Ivanishvili’s hand-picked successor, whom he will announce a week after the presidential contest.
Analysis
Georgian Dream’s electoral strategy for the presidential race, including Margvelashvili’s threat to withdraw if he doesn’t win the first round outright, makes sense in Georgia’s clientelist political context. GD is a loose coalition of parties with little in common in terms of policy programs. During the 2012 parliamentary election cycle, GD served as a catch-all for popular disillusionment with the UNM. In this sense, Ivanishvili’s personality and resources served as a focal point for opposition to the ruling party. With the GD now in power, and UNM still relatively unpopular, the biggest challenge for Margvelashvili may be mobilizing his supporters for the first round, lacking any issue as galvanizing as UNM’s abuses of power.
With this in mind, the best outcome for GD is obviously to win the presidency outright in the first round. This outcome would signal a clear balance of power in favor of their network, potentially attracting elites on the margins who want to contest or retain offices in future local and parliamentary elections. However, due to the two-round voting system and the incomplete information provided by public opinion polls, GD may not be certain it can achieve its best outcome. If GD perceives that its supporters are complacent, or voters on the margins are ambivalent, it could be relatively cheap talk designed to encourage turnout among anti-UNM voters. Whether this sort of threat would actually have any benefit is uncertain, but in any case, it likely has little cost. Core GD voters are unlikely to jump ship or stay home, and Margvelashvili can, probably with little cost, change his mind later depending on the results of the first round. Margvelashvili’s announcement to withdraw if he does not win the first round outright could therefore be a strategy for dealing with this uncertainty in an environment in which mobilization matters more than policy differences.
Alternatively, the announcement may not be strategic at all, but rather a reflection of confidence that Margvelashvili will pick up enough undecided voters to avoid a runoff. Despite the high percentage of likely voters who were undecided, non-responsive, or expressed no preference (26 percent) in the September poll, Margvelashvili appears to hold a commanding lead, and only needs to pick up about another 10% of voters to avoid a runoff. Furthermore, there is good reason to believe a large portion of undecided voters will turn out for Margvelashvili. Bakradze, with the support of around 18 percent of likely voters, is already more popular than his party (12 percent of respondents indicated UNM as the party closest to them in the same poll).* Bakradze’s popularity, therefore, is probably the upper bound of any underlying UNM support in the larger population, so there does not appear to be a reservoir from which the UNM could mobilize votes. Meanwhile, around 50  percent of respondents in the NDI poll identified GD as the party closest to them, higher than Margvelashvili’s level of support. Burjanadze, finally, is a bit more of a dark horse in the sense she could serve as a focal point for undecided voters or supporters of other candidates who oppose both the GD and UNM. Like Bakradze, however, she outperforms her party, the United Georgia-Democratic Movement. Moreover, while her party is relatively well financed it is unlikely that she will be able to generate enough support to force a second round. In any case, Margvelashvili’s announcement may just be speculation in response to what he and Ivanishvili consider a highly improbable hypothetical situation.
Recent comments from Ivanishvili and Margvelashvili provide evidence for both of these interpretations. Margvelashvili has described his announcement as a demonstration of resolve. GD could therefore be trying to constrain its activists and voters to first round participation if Margvelashvili can credibly threaten to deprive them of an opportunity to participate in the second. Ivanishvili, on the other hand, has tended to dismiss a second round, and as a result, Margvelashvili’s withdrawal, as a remote hypothetical. Most likely, the announcement is some combination of the two — reasonable confidence in a first-round victory, but with a hedging strategy. In either case, there is obviously a difference between threatening to withdraw from a second round and actually withdrawing. If Margvelashvili fails to obtain a majority on Sunday and follows through on his pledge, it could reflect a different strategic logic entirely.
*H/T to Dr. Timothy Blauvelt of  Ilia State University in Tbilisi for providing this observation, among several other helpful suggestions. Any mistakes in the interpretation are mine.