Thoughts on the Runoff
On my return to the US, I was contemplating the data that I would want/need to conduct a systematic assessment of the upcoming runoff between Viktor Yanukovych and Yuliya Tymoshenko. Most valuable would be solid public opinion data that perhaps oversampled supporters of losing candidates, and those who did not vote, to assess their likely behavior in round 2. But, I do not have access to these data. However, in the grand, centuries-old tradition of blogging, I will not let the absence of these data deter my analysis and commentary.
I have compiled, as a starting point, election results from the Central Electoral Commission at the national and regional levels. While I plan to acquire and assess precinct-level data, that is a more time-consuming task. In the short-run, I am focusing on data aggregated at the oblast level. A caveat before I continue: the analysis of data at this level of aggregation runs afoul of the ecological inference problem. Indeed, the behavior I am interested in assessing is at the individual-level, but the data are at a higher level of aggregation. A second caveat, always applied to projections of second-round contests, is that the composition of voters may change. In fact, changing the composition of voters is to the campaigns' advantages, as I will argue below. With these caveats in mind, here is some preliminary analysis.
While all of the ballots have not been included in the tally, the results are unlikely to change significantly (although they will change slightly). According to the data from the CEC, Yanukovych claimed 35.32%, Tymoshenko 25.05%, Tihipko 13.06%, Yatseniuk 6.96%, Yushchenko 5.45%, and the remaining 13 candidates and against all picking up the rest of the valid ballots. All of the losers are players, to one degree or another, in round 2, but Tihipko is in a particularly strong position to negotiate terms with both camps.
Using the Kyiv International Institute of Sociology's polling from a few weeks before the first round, I performed a little thought experiment. The KIIS poll asked likely voters for all candidates who they would support in a round 2 matchup between Yanukovych and Tymoshenko. For many minor candidates the number of respondents indicating support was small, rendering their answers on this question problematic (for example, if two people indicated support for candidate Protivsikh, and one supported Yanukovych in hypothetical round 2 and the other Tymoshenko, it could be problematic to project a 50-50 split in the Protivsikh vote). In addition, the poll was taken well before the round 1 moratorium and positions may change. Nevertheless, it is an interesting starting point to consider.
I used the preferences of respondents to estimate how votes could break in round 2. Yuliya begins with a 2.5 million vote gap. Support from losers narrows the gap by 500,000 (that is, Yuliya picks up a higher proportion of votes from losers than Yanukovych), but the estimates still leave roughly 4.2 million voters (around 17%) from round 1 unallocated. Of particular note, based on the KIIS survey results, Yanukovych takes more Tihipko supporters than Tymoshenko. But, nearly 40% of Tihipko's supporters did not express a preference.
In addition, I compiled data from the presidential election, as well as the last two parliamentary contests, and used ArcGIS to generate various maps displaying the results. You may click on the maps for larger versions.
This map displays Yanukovych's performance across the country, with oblasts color-coded by .15 increments. The choice of increments can exert a strong effect on the way we interpret maps, and I may generate some different versions in later posts. There are few surprises here, with Yanukovych performing well in the east, especially Donetsk and Luhansk. Perhaps his relatively weaker performance in Dnipropetrovsk is a surprise, at first glance, but this is an issue that I will return to momentarily.
Tymoshenko's results are also unsurprising. She performed well in the west, especially Volyn and Vinnystia. Her strongest regional performances fall below Yanukovych, a point which could prove to be advantageous in round 2.
The third-place finisher, Sergey Tihipko, showed several areas of strength. Most notably, he placed well in Dnipropetrovsk and Odesa, as well as in Kyiv, Kharkiv, Zaporizka, and Sevastopol. Please note that his performance is displayed in .05 increments. Two more maps, for Arseniy Yatseniuk and Viktor Yushchenko:
Like Tihipko, Yatseniuk's performance is displayed in smaller increments (Yushchenko is in .15 increments due to his performance in the west). Yatseniuk had some strength in Chernivtsy and Ivano-Frankivsk, and Yushchenko did well in L'viv, Ternopil, and Ivano-Frankivsk.
Now, some thoughts for both campaigns.
I have compiled, as a starting point, election results from the Central Electoral Commission at the national and regional levels. While I plan to acquire and assess precinct-level data, that is a more time-consuming task. In the short-run, I am focusing on data aggregated at the oblast level. A caveat before I continue: the analysis of data at this level of aggregation runs afoul of the ecological inference problem. Indeed, the behavior I am interested in assessing is at the individual-level, but the data are at a higher level of aggregation. A second caveat, always applied to projections of second-round contests, is that the composition of voters may change. In fact, changing the composition of voters is to the campaigns' advantages, as I will argue below. With these caveats in mind, here is some preliminary analysis.
While all of the ballots have not been included in the tally, the results are unlikely to change significantly (although they will change slightly). According to the data from the CEC, Yanukovych claimed 35.32%, Tymoshenko 25.05%, Tihipko 13.06%, Yatseniuk 6.96%, Yushchenko 5.45%, and the remaining 13 candidates and against all picking up the rest of the valid ballots. All of the losers are players, to one degree or another, in round 2, but Tihipko is in a particularly strong position to negotiate terms with both camps.
Using the Kyiv International Institute of Sociology's polling from a few weeks before the first round, I performed a little thought experiment. The KIIS poll asked likely voters for all candidates who they would support in a round 2 matchup between Yanukovych and Tymoshenko. For many minor candidates the number of respondents indicating support was small, rendering their answers on this question problematic (for example, if two people indicated support for candidate Protivsikh, and one supported Yanukovych in hypothetical round 2 and the other Tymoshenko, it could be problematic to project a 50-50 split in the Protivsikh vote). In addition, the poll was taken well before the round 1 moratorium and positions may change. Nevertheless, it is an interesting starting point to consider.
I used the preferences of respondents to estimate how votes could break in round 2. Yuliya begins with a 2.5 million vote gap. Support from losers narrows the gap by 500,000 (that is, Yuliya picks up a higher proportion of votes from losers than Yanukovych), but the estimates still leave roughly 4.2 million voters (around 17%) from round 1 unallocated. Of particular note, based on the KIIS survey results, Yanukovych takes more Tihipko supporters than Tymoshenko. But, nearly 40% of Tihipko's supporters did not express a preference.
In addition, I compiled data from the presidential election, as well as the last two parliamentary contests, and used ArcGIS to generate various maps displaying the results. You may click on the maps for larger versions.
This map displays Yanukovych's performance across the country, with oblasts color-coded by .15 increments. The choice of increments can exert a strong effect on the way we interpret maps, and I may generate some different versions in later posts. There are few surprises here, with Yanukovych performing well in the east, especially Donetsk and Luhansk. Perhaps his relatively weaker performance in Dnipropetrovsk is a surprise, at first glance, but this is an issue that I will return to momentarily.
Tymoshenko's results are also unsurprising. She performed well in the west, especially Volyn and Vinnystia. Her strongest regional performances fall below Yanukovych, a point which could prove to be advantageous in round 2.
The third-place finisher, Sergey Tihipko, showed several areas of strength. Most notably, he placed well in Dnipropetrovsk and Odesa, as well as in Kyiv, Kharkiv, Zaporizka, and Sevastopol. Please note that his performance is displayed in .05 increments. Two more maps, for Arseniy Yatseniuk and Viktor Yushchenko:
Like Tihipko, Yatseniuk's performance is displayed in smaller increments (Yushchenko is in .15 increments due to his performance in the west). Yatseniuk had some strength in Chernivtsy and Ivano-Frankivsk, and Yushchenko did well in L'viv, Ternopil, and Ivano-Frankivsk.
Now, some thoughts for both campaigns.
- Mobilize core voters. While this is an obvious recommendation, it is particularly important because round 2 may be close. Not only do both campaigns need to energize voters who showed up to the polls, but they also need to engage voters who did not. Tymoshenko might face a greater challenge regarding the latter. Western oblasts had relatively high turnout. Yanukovych might be able to make gains increasing turnout in Odesa, Crimea, Kirovohrad, and Mykolaivsk (and perhaps in Kherson and Zakarpattya also). Tymoshenko may be able to benefit from higher turnout in Chernivets - a region that disproportionately supported Yatseniuk and had second-to-lowest regional turnout. But, the population of Chernivets is low compared to regions where Yanukovych may benefit from mobilization.
- De-mobilize potential voters for the opposition. Some of this strategy depends on how the campaigns perceive voter preferences among supporters of Yatseniuk and Yushchenko. If they are perceived as having anti-Yuliya attitudes, but even stronger anti-Yanukovych attitudes, then the Tymoshenko campaign benefits from some conciliatory gestures. Yanukovych benefits from reminding them how much they dislike Yuliya. If these voters are generally perceived as anti-Yuliya (especially Yushchenko voters given the personal animosity between the two), then Yanukovych benefits from attempting to mobilize them, and Yuliya benefits from de-mobilizing them.
- Get Tihipko Voters. Tihipko has many options to consider. It is unclear how loyal his supporters may be, so his endorsement may or may not bring these voters along. But, his supporters were strong in especially important areas. The following maps show how support for Tymoshenko and Yanukovych changed between 2006 and 2010, as well as 2007 and 2010. Another caveat: I am comparing parliamentary election results with presidential ones. The decision-making algorithm for voters, as well as the choice set, probably differs. But, the personalistic nature of parties and campaigns renders this comparison not wholly unreasonable.
These maps show Yanukovych's changes in fortune between the the parliamentary elections and the first round of the presidential contest. Note especially Yanukovych losses in Dnipropetrovsk and Zaporizka.
Similar to the Yanukovych maps, the Tymoshenko maps reveal some losses in areas of Tihipko gain (again, Dnipropetrovsk looms large).
Expect to see strong efforts to gain Dnipropetrovsk from both campaigns. In addition, the Kyiv region, Kharkiv, Kherson, Odesa, and Zaporizka could be strongly contested. - Picking up votes. For Yanukovych, Yuliya's potential to pick up votes from competitors in the West is a potential problem. In Yuliya's best region (Volyn), she beat Yanukovych by 44 points. In Yanukovych's best region, he beat Tymoshenko by 72 points. This map shows the gap between the candidates. Regions leaning to Yanukovych are reflected by positive values.
Yanukovych's top competitors were not as strong as Yuliya's in their respective regions of strength. But, while Yanukovych has less to gain in terms of the percentage of voters that he might acquire from competitors, his areas of strength have substantially larger populations.
