Results of Parliamentary Elections 2016 — The Wisdom of the Crowd

Hans Gutbrod & William Dunbar | January 10, 2016

In recent years, election polls have received increasing attention in public debates in Georgia. But what do they really say about likely election outcomes? As various sides focus on particular numbers, one important message often does not get through: the eventual results are likely to be quite different, primarily because not everyone who participates in a survey will eventually go to vote.

There are other reasons, too, such as last-minute political developments, voters not making up their minds until late, or (in some cases) voters being nervous about revealing their real choice. These are problems for polling everywhere, but in Georgia the issue is polarized further by there being few polls, and few people who comment on polls.

What we need, then, is a more structured discussion on what polling numbers mean, and a survey of experts can help us achieve such a discussion. With that in mind, we asked 50+ people in late November 2015 what they thought the likely election results next fall were going to be. Most of those have taken a longer interest in Georgian politics, and if they are not certified as experts, they at least are the equivalent of engaged citizens. Pooling their estimates may be a sensible approach, as repeated studies have shown that aggregate estimates often are surprisingly accurate.

What, then, are the predicted results? In aggregate, the group sees the Georgian Dream ahead. Our group predicted 32% on average. The majority (70%) of those giving a guess saw Georgian Dream between 20–40% of the vote, still quite a broad corridor, suggesting the overall uncertainty on the outcome. There were extreme guesses on both sides, with one suggesting 78% for the Georgian Dream, and another 14% — these extreme estimates typically cancel each other out. (Note that with the majoritarian system, a strong presentation throughout the country can translate into many seats, as long as the opposition parties are split.)

The group saw the United National Movement at 27%, trailing GD by 5%. Most estimates saw the UNM between 17% and 37% of the total vote, highlighting that in the extremes, too, people see UNM trailing behind the Georgian Dream. Here there also were some extreme suggestions, with 58% and 12% as the respective extremes.

What about the other parties? As it currently stands, Irakli Alasania’s Free Democrats are at about 10%, and only 10% of those volunteering their guess thought that the Free Democrats would get more than 15% of the vote. The collective wisdom is that the Free Democrats will solidly get into Parliament.

Nino Burjanadze with her Democratic Movement and the Alliance of Patriots, both parties that want to chart a course friendlier to Georgia’s northern neighbour, could just about make it into Parliament. Nino Burjanadze is seen as receiving 7%, and Irma Inashvili just about 6% — but for both, getting in remains uncertain. If they found common ground, electoral arithmetic would suggest to them that some kind of coalition may be advantageous, as they would increase their chances of parliamentary representation. Similarly, the New Political Center/Pinecone still has some way to go, to assure themselves of making it into parliament. Forecasts put it at 3% — this still gives them a chance of catching up.

Now these are, of course, long-range forecasts, and much can still happen. Some forecasters highlighted that new parties were likely to get about 10% of the vote. We may also get last-minute surprises, as they previously impacted elections.

Obviously, no long range estimate is perfect, as we know from weather forecasts. The response to those limitations is not to do away with forecasts, as some suggested, but to make them better, and to improve the discussion on what they really mean. Getting a crowd to pool their estimates is one effective contribution, to such an improved discussion. Likely, this would also improve debate on election forecasts, as the broader public would understand that polling can provide a sensible corridor of expectation, but that its purpose is not to predict an exact point landing.

Our next round of collecting forecasts is planned for June 2016. Want to contribute? Send an email to this address, to participate and to receive results.

One clap, two clap, three clap, forty?

By clapping more or less, you can signal to us which stories really stand out.