Thursday, May 28, 2015

5 Predictions on Turkey's Upcoming Elections

With Turkey's parliamentary elections just 10 days away, here are five (not so) bold predictions:

1. AKP will win the election with about 42.85% of the vote.

Earlier, I presented a predictive model of the AKP vote share based on the state of the economy in the year leading up to the election. Two slightly different versions of this model produce predictions of 42.5% and 43.2% for AKP. Average those together, and one comes up with a prediction of 42.85% for AKP.

This is slightly higher than the poll-averaged prediction of 40.9% that Michael Daventry has produced. In relying solely on economic estimates, one possibility is that the economy is actually performing worse than recent projections have suggested. If so, then the appropriate prediction would be lower.

It's also important to note that my model offers no prediction of how the opposition parties will fare. I predict that AKP will lose about 7 percentage points relative to its 2011 vote share, but any predictions about which parties will gain those votes must be based on polling data. What the polling data suggest is CHP should get around 25-28% and MHP around 15-18%. Then there's the fourth party...

2. The big question remains whether HDP get 10% or not.

I've argued elsewhere that the biggest question of this election is whether HDP get 10% or not. Because of Turkey's unusually high 10% election threshold, HDP face an "all or nothing" situation in which they will win either 60 or more seats (if they get 10% of the vote) or no seats (if they fall short of 10%).

HDP have increased their electoral support by attracting an increased share of conservative Kurds and liberal/leftist Turkish voters. A number of polls show HDP above 10%, but it remains "too close for comfort."

My prediction is that HDP will receive over 10% of the vote in actuality. The question raised in point 5 below is whether that will be reflected in the official election results.

3. There is a real chance AKP could end up short of a majority...or with a super-majority.

The "all or nothing" dilemma surrounding HDP largely affects AKP's parliamentary prospects. Due to Turkey's electoral geography, AKP stand to gain the vast majority of those 60+ seats if HDP fall short of 10%.

In Turkey's parliament, 276 are needed for a majority. With 330 seats (60%), it is possible to trigger a national referendum on amending the constitution. Thus, the stakes are high.

I simulated a few possible election scenarios to determine possible party compositions after the election. There are online simulators that one can use as well. The model that seems the most plausible is the "HDP/MHP Surge" simulation, which assumes that those two parties will pick up most of AKP's lost support, depending on the electoral district, with CHP support remaining constant. That model predicts the following vote/seat share:

Vote % Seats
AKP 42.86 271
CHP 25.99 136
MHP 16.03 82
HDP 10.53 61
Others 4.58 0

Thus, this model predicts that AKP will end up with just shy of a parliamentary majority. But here is the flip side: if HDP fall short of 10%, then this model would put AKP close to 330 seats, having gained most of those 61 seats.

4. The breaks are more likely to go AKP's way.

I think it is more likely that AKP will overperform the vote and seat predictions offered here (and elsewhere). I suggest two reasons why.

First, AKP have all of the structural advantages: the power of incumbency, control of the public (and much of the private) media, the greater organizational capacity, and the 'trust' factor that the opposition parties cannot match. In the end, some wavering Turkish voters may decide to stick with AKP rather than vote for an untested opposition party. This is particularly true in the case of HDP. There are a lot of anecdotal stories of liberal, Western Turkish voters who plan to vote HDP. But will it actually happen in the event, or will many of them go back to their default vote (CHP)?

Second, I examined one further aspect of my seat share simulation above. I examined the district-level seat projections to examine two questions: which party is predicted to win the last seat in each district, and which party would win the 'next' seat if there were one (put different, which is the first party excluded)? In the chart below, the first column suggests the number of seats that each party is vulnerable to losing, while the 2nd column shows the number of seats that it could gain by doing a little better than predicted. This suggests that AKP may be more likely to overperform these predictions:

Last Seat
Next Seat
AKP 35
CHP 16
MHP 22
HDP 12

Note that AKP is the only party with a negative ration between last seat total and next seat total. It would stand to win the next seat in 47--over half--of the electoral districts. What this means is that a slight increase in AKP's vote share could lead to a significant swing in seats towards AKP. MHP seem particularly vulnerable: they are 'defending' the last seat in 22 districts. In 18 of these, AKP would win the next seat. Thus, a slight shift from MHP to AKP in these district could significantly increase AKP's seat share at the expense of MHP.

5. This will not be a fair election.

In one very real sense, it is already not. I've argued elsewhere that Turkey is best understood as a case of competitive authoritarianism, a situation in which formal democratic institutions are routinely undermined by the rulers in order to maintain their grip on power. We have seen plenty of evidence of this throughout this election campaign:

--Public media coverage of AKP/Erdogan rallies has greatly exceeded coverage of opposition parties
--President Erdogan has routinely violated constitutional mandates that he remain politically neutral
--The Supreme Electoral Council (YSK) has consistently refused to sanction AKP for these violations

In short, the opposition is competing on an uneven playing field that is tilted in every instance towards AKP.

In addition, there is a very high likelihood of election fraud on 7 June. The March 2014 elections were marred by a number of documented instances of fraud (a good summary can be found in the latter pages here). There were various tactics at work: destruction of valid ballots, miscounts of vote tallies, registration of nonexistent voters, the 'power outages' while votes were being counted, reports of vote buying, etc.

It would be naive to expect no election fraud in these upcoming elections. Based on the contours of this election, there are two plausible targets of election fraud:
--Fraud to keep HDP below 10%
--Targeted fraud in those 18 'vulnerable' MHP districts to swing them to AKP. (If you are interested, they are: Aksaray, Bayburt, Burdur, Canakkale, Cankiri, Duzce, Giresun, Gumushane, Istanbul 3, Kahramanmaras, Kars, Kilis, Kirikkale, Kirsehir, Manisa, Nevsehir, Sakarya, and Usak. Note that those are based on my projections. You can also use this simulator to examine district-level seat predictions).