A friend looked over my recent posts and said, basically, that this analysis of things that had already happened was all find and dandy, but might it be possible to predict the results of the second round? On Tuesday, I speculated that a coalition of the Labor Party (Darbo partija), Social Democrats, and Order and Justice (Tvarka ir teisingumas) would be at most 11 seats away from a majority, meaning that the three parties have 11 second round elections against other parties that they have to win. But what are the chances of getting those 11 seats? Similarly, considering that the conservatives (TS-LKD) are the most represented party in the second round, how likely is it that the Prime Minister Andrius Kubilius’s words—that his party will be the largest faction in the new Seimas—will come true?
I’m not an elections prognosticator—there’s already one alum of my university who does that rather well—but, I am a bit of a hobbyist. So I decided to take the results of the first round and develop a predicting methodology for the second round. For the TL;DR crowd, here are the results:
- The conservatives will pick up 11 seats.
- The Social Democrats will pick up 11 seats.
- The Labor Party will pick up nine seats.
- Order and Justice will pick up four seats.
- The Polish Action (Lenkų akcija) will pick up two seats.
- The liberals (LRLS) will pick up one seat.
- The Path of Courage (Drąsos kelias) will pick up one seat.
That’s 37 of 68 seats. The rest my methodology considers too close to call. But it does indicate that the Social Democrats, Labor, and Order and Justice still have some work to do to get to a majority. What was the methodology? Both of the parties in each run-off were awarded bonus points based on certain criteria. If a party scored more than three bonus points, it was solid for that party. If a party scored more than one point, it was a slight favorite. If both parties picked up less than one point, it was considered a toss-up.
Here’s the rather arbitrary way the points were distributed:
- +1 for an incumbent (or if the candidate is a member of the incumbent party). +.5 if the incumbent party is not represented at all. (This second part was a bit tricky, since parties are so volatile in Lithuania)
- +1 if the candidate won the first round by more than 20 points. +.5 for a win of more than ten points. +.25 for a win of more than five points. The inverse was true for the first round runner-up.
- +1 if the candidate’s party also was the party that received the most votes in the constituency in proportional voting.
- +1 if the candidate’s party received votes at a percentage two standard deviations (or more) above the party’s national mean in the constituency. +.25 if it was more than one standard deviation above the mean. The inverse was applied if the party underperformed its national average within the constituency.
- +1 if the candidate’s party was the leading proportional party and received 28% or more of the vote in proportional voting (one standard deviation above the mean of all top parties in all constituencies of ~24%).
The rationale was:
- Incumbency would be rewarded, even in an election about change. There would even be a trace effect if a new candidate represented the incumbent party. But both sides benefited a bit if it was a clean slate.
- Beating up on the second place candidate in the first round bodes well for the second round.
- Representing a popular party in the constituency helps.
- If your party is especially locally popular, it helps.
- If your party demolished the others in proportional voting, it helps.
Interestingly, only two candidates received 4.5 bonus points (against .25 for their opponents), and both represent the Polish Action. Somehow I’m a little skeptical that these are locks, since I suspect the non-Polish population in both constituencies will rally and vote for the Lithuanian candidate. The question is if there will be enough Lithuanians voting to offset the Polish vote. In the other constituencies, the matrices get too complicated for me to consider.
So how does it play out spatially? Glad you asked (as always, click to enlarge):
It’s no surprise that there’s some correlation with the maps from the previous post; that correlation is built into the means by which I hand out bonus points. The toss-ups (mostly) have the showdown indicated on the map. The party that got more votes in the first round is the first mentioned. I’ll also include a few explanations of abbreviations: TT = Order and Justice. LCS = other liberals. Indep = Independent candidate. VZ = Peasants & Greens. Also, the three “decided” constituencies feature candidates who broke the 50% threshold in the first round and face no run-off.
In the Vilnius inset, the northwestern toss-up is Fabijoniškių constituency. It’s TS-LKD vs. Darbo p. The eastern constituency, Naujosios Vilnios, is Lenkų a. vs. Darbo p. The large one in the southwest, Lazdynų constituency, is TS-LKD vs. LRLS.
All three Kaunas toss-ups are TS-LKD vs. Drąsos k. In Klaipėda, the northern toss-up, Danės constituency, is TS-LKD vs. LCS. The southern constituency is actually Pajūrio constituency, which makes up much of the coast. It’s TS-LKD vs. Darbo p.
A few small consituencies are obscured in the main map. Alytaus constituency is a Darbo p. vs. SocDemai toss-up. The northern constituency of Panevėžys, Nevėžio, is TS-LKD vs. Darbo p., and the southern constituency, Vakarinės, is Indep. vs. Darbo p. Marijampolės tiny constituency is mostly blocked by the label for the surrounding Suvalkijos constituency. It’s a slight SocDemai favorite.
So there are some predictions. We’ll see in just over a week how ridiculous they were.