Friday, October 9, 2015

Wartime Elections

In a forthcoming Electoral Studies article, Michael Thunberg, Nazar Boyko, and I assess the effects of conflict on the 2014 snap presidential and parliamentary elections in Ukraine. Our findings have relevance for the upcoming local elections to be held at the end of October as well as elections outside of Ukraine.

Civil and international crises impede state capacity and inhibit the conduct of free and fair elections. Yet, democratic (and semi-democratic) countries have held elections during wars and insurrections. The United States conducted elections during its Civil War on the territories the Union controlled, and many contemporary examples exist across the globe. Election administrators face the challenges of maximizing enfranchisement, security, and integrity when state sovereignty and the safety of participants are under threat. Our article analyzes how Ukraine adapted its election administration to manage the process in wartime conditions. We found that:
  • Precinct Electoral Commissions (PECs) located closer to the conflict zone were more likely to be relocated and report lower levels of citizen participation. The map below shows the location of PECs that were closed on election day: dark blue dots are PECs that were never established and red dots are PECs that were established but did not report election day results. 
  • While conflict effects were measurable in Donetsk and Luhansk, they did not extend outside Donbas,* suggesting that Ukrainian state authorities successfully contained conflict. PECs situated closer to active combat experienced substantially lower participation than those located closer to the borders of contiguous regions (Kharkiv, Dnipropetrovsk, and Zaporizhia). The effects of distance from the conflict zone on voter behavior disappeared outside of Donbas.
  • Partisan effects of wartime management especially impacted voters who had supported Viktor Yanukovych and the Party of Regions in past contests. It is unclear how much of the diminished participation in open PECs in Donbas was due to concerns about safety or the absence of parties/candidates that citizens supported.

Closed PECs in the 2014 Snap Parliamentary Election. Figure by Roman Sverdan, CIFRA Group
The upcoming local elections face similar difficulties. Although the so-called Donetsk People's Republic and Luhansk People's Republic have postponed their local elections, enfranchising voters and administering elections in Donbas is fraught with problems. Elections will not be administered on territories that Ukrainian security services do not control. Just as in October 2014, large numbers of voters located in Donbas will be unable to cast their ballots in October 2015.**

Outside of Donbas, the decentralization of administration for local elections creates additional impediments. The Central Electoral Commission is not collecting and disseminating data as it would in national elections; these responsibilities have been delegated to Territorial Electoral Commissions at the local level. Variation in local conditions and the quality of TEC personnel may generate different experiences for voters, candidates, and parties.

* The exception to this observation is Crimea, where elections could not be held.
** Enfranchising Internally Displaced Persons is an additional challenge. We did not analyze effects on these voters in the article as data are unavailable.
Replication Data for the Electoral Studies article.
Data and Stata .do files for the tables are available for download.

Thursday, July 16, 2015

Technical Parties, Partisan Election Management Bodies, and Election Outcomes

In the December 2015 edition of Electoral Studies (online now, gated), Nazar Boyko and I evaluate how efforts to "stack the deck" on electoral commissions affect voting outcomes in Ukraine. Institutional rules* and weak parties help major players to place affiliated individuals on precinct electoral commissions. We investigate whether or not the presence of "technical party"** commissioners benefits their patrons, and find evidence of statistically significant, albeit substantively small, "bonuses" in election returns.

While the ideal for election administration is neutral counting and compilation of ballots, the findings suggest that election administrators interpret rules (such as whether or not to validate disputed ballots) with partisan outcomes in mind. Various reforms could mitigate these effects, such as tightening eligibility requirements for parties/candidates to nominate commissioners, limiting replacements of commission staff, and professionalizing commissioners via improved training and enhanced state compensation.


Data used in the analysis (.dta)

* Notably, staffing procedures that balance partisanship and liberal party/candidate registration requirements facilitate stacking.

** See Brian Mefford's blog post about the 2015 by-elections in Chernihiv for a current example of extremes in the deployment of technical candidates. Mefford notes that up to 30 or so technical candidates are associated with major candidates, with 127 candidates registered for that race alone.

Monday, April 13, 2015

Candidate Characteristics and the Construction of Party Lists

I am presenting a second Midwest Political Science Association Conference paper (available here) with Michael Lynch that addresses how political parties manage party lists.* Specifically, it investigates how careful cultivation of down-list candidates (those situated in hopeless list positions) may benefit the party. We take advantage of a unique feature of Lithuania's party-list vote - the presence of candidate preference votes - to gauge how characteristics associated with personal vote-seeking activities affect spatial variation in preference vote acquisition. We find that preference vote performance is enhanced in regions where candidates have a local connection (e.g., nativity or residency), even among those who do not contest the constituency tier of the mixed-member system. Our findings suggest that parties can enhance their overall performance by diversifying party lists, even among unelectable list positions.

*The authors thank Monika Aleknaviciute and Maksym Palamarenko for research assistance. The research was supported, in part, by a grant from the National Science Foundation (SES – 0751662, PIs Ellis Krauss, Robert Pekkanen, and Matt Soberg Shugart).

More on State Capacity and Election Administration During Conflict

Later this week, Nazar Boyko and I are presenting a paper (available here) at the Midwest Political Science Association Conference that explores how institutional, temporal, spatial, partisan, and experiential factors are associated with administrative capacity to conduct elections. We analyze data from surveys of officials in District Election Commissions (DECs) and Precinct Election Commissions (PECs) conducted during the October 2014 parliamentary elections in Ukraine.*

In general, we found that election administrators expressed confidence in their level of preparedness and security, as well as the integrity of the election. Variation in evaluations of preparedness or security were largely attributable to the location of the DEC/PEC and training. Unsurprisingly, commissioners in Donetsk and Luhansk were more likely to express lower levels of confidence in preparedness and security than their peers elsewhere. Assessments of integrity were associated with location as well, but concerns about manipulation were not limited to Donbas. Participation in training (standard training and supplementary security training) mitigated many of these concerns. In sum, the research suggests that the conflict's effects were largely contained in Donbas, and that state interventions (especially to enhance commissioner readiness) enhanced perceptions of readiness and integrity.

*We thank Andriy Magera and Ukraine's Central Electoral Commission for access to personnel data and support in conducting personnel surveys; IFES for technical assistance in developing the surveys and permission to interview subjects at IFES sessions; Michael Thunberg for research assistance; and the participants in the GWU Post-Communist Politics/Social Science Workshop for valuable feedback on an earlier version of the research. The DEC survey received financial support from PACT/UNITER/USAID. The PEC surveys were supported by a National Science Foundation RAPID grant (SES-1462110).

Friday, January 23, 2015

State Capacity and Election Administration During Conflict

After the August 2014 announcement that Ukraine would hold early parliamentary elections, Nazar Boyko and I developed survey instruments to better understand how election administrators assessed their training and preparation, compensation, election processes, and potential outside influences.* The surveys are part of a larger project that investigates state capacity through election administration, using the unique circumstances of conflict and occupation to gauge how the state has responded to challenges.

The surveys were implemented from late September 2014-late November 2014, in collaboration with the Kyiv International Institute of Sociology. We recently made presentations at the Southern Political Science Association Annual Conference and at a George Washington University-sponsored workshop to introduce some of our preliminary observations from the survey data.

We have found evidence of institutional, temporal, spatial, and partisan effects on the responses of administrators, and are currently analyzing the data to better understand how district-level and precinct-level commissions functioned during the parliamentary election.

A few highlights of our findings include:
  • DEC respondents reported heightened pre-election concerns, lower levels of party-based financial support, and more awareness of technical parties than their PEC counterparts.
  • PEC respondents tended to show higher levels of concern/lower levels of confidence in the pre-election survey compared to the post-election survey. 
  • Respondents located in the Eastern regions of Donetsk and Luhansk showed elevated concerns about safety and security, and less confidence in their preparation. While this outcome is not surprising given conflict in the region, the effects did not seem to "spill over" into contiguous regions.** 
  • Post-election survey participants in the West were the most likely to report that falsification was common, followed by participants in the East. However, reports of falsification were limited.
  • Commissioners nominated by affiliates of the former ruling party were more likely to express concerns than commissioners nominated by former opposition parties. 
The paper describing our preliminary findings is available online. We welcome comments and suggestions.
*This research was funded by grants from PACT/UNITER/USAID and the National Science Foundation (SES-1461220). We also acknowledge essential collaboration from the Central Electoral Commission of Ukraine, Kyiv International Institute of Sociology, and IFES.

** We will refine the analysis to include other spatial measures, including distances between commissions. If these findings hold, however, it suggests that the conflict - and its consequences - have been compartmentalized by administrators. That is, the conflict was viewed as a localized phenomenon and fear of contagion was not present.

Monday, January 19, 2015

Local Machine Politics and Elections in Ukraine

At this year's Southern Political Science Association Conference, I presented a paper (co-authored with Fredrik Sjoberg) assessing local political machines and electoral competition in Ukraine.

While the business sector has legitimate policy interests and the right to lobby for particular outcomes, it has often been accused of using illegitimate means to secure its preferred policies. Extending studies that describe machine politics in Ukraine, our paper investigates the connection between "boss" candidates and allegations of fraud, using data from the 2012 parliamentary elections. We combined candidate biographies, election results, and crowdsourced election observation data, and found that higher levels of competition among boss candidates is associated with higher levels of vote buying allegations. The analysis did not produce significant findings for other types of voter manipulation.

While the research should not be used to excoriate all business sector candidates who compete in Ukraine, it provides evidence that business sector participation is associated with specific fraud tactics, especially those which provide a competitive advantage to bosses.