Turkish election, 2015

Turkey is holding general elections on 7 June. It is an unusually important election, coming as it is after the first direct presidential election. President Recep Tayyip Erdogan is seeking parliamentary support to amend the constitution to make the presidency far more powerful–currently it is mostly ceremonial. For that he needs either his Justice and Development Party (AKP) to win two thirds of the seats, or to have willing coalition partners that would get a constitutional amendment bill to two thirds.

This time around, a potentially significant party for the Kurds is running. In past elections, Kurdish candidates have generally run as independents. According to Henri J. Barkey, in Foreign Policy, the AKP is polling in the low forty-percent range. The main party of the Kurdish minority, People’s Democratic Party (HDP), is polling anywhere from 8.5% to 11.5%. Which end of that range the HDP ends up in is critical, because Turkey uses a 10% nationwide threshold.

In 2007, I questioned whether the Turkish system even should be considered democratic, because of its very high threshold, combined with district-level allocation of all seats. That is, there is a system of 550 seats divided among 80 or so constituencies, for a mean district magnitude of around 7. However, even if a party wins sufficient votes to to get seats in one or many of these districts, it actually will get no seats at all if it has not won 10% nationwide. I have previously offered some egregious examples of district-level results generated by this combination of districting and nationwide threshold. Independents, however, can and often do win seats–many of these have been informally affiliated with Kurdish parties that couldn’t win 10% nationwide but could elect some “non-party” candidates in specific districts where their supporters are concentrated.

In 2011, I did some calculations meant to estimate how many parties should be expected based on this system. Over the 1995-2007 period, the averages are about what my estimates implied they should be. I concluded that maybe the system was bifurcating into two parallel systems–one of parties passing the 10% threshold, and another of non-party MPs representing Kurds and other interests.

The 2015 election will be a major test of both the representativeness of the electoral system and the strategy of the HDP. If the party clears 10%, it can hold the balance of power and force Erdogan to make concessions to the Kurdish minority in order to get his presidential powers.* If HDP fails, the AKP will probably get a manufactured two-thirds majority and be able to govern and amend the constitution at will. And, in that event, the Kurds will be worse off for not even having a bloc of independent legislators present in the parliament. For their sake, and the sake of Turkish democracy (such as it is, I hope the HDP and Kurdish leaders know what they are doing. As Barkey, the worst thing would be for the HDP to just miss the threshold, and AKP to get two thirds.


* Some weeks ago, I recall seeing an article in the Economist that suggested Erdogan was intending to strike just such a deal. The tenor of Barkey’s Foreign Policy piece is quite different. I don’t know the Turkish situation well enough to know which is likely.

59 thoughts on “Turkish election, 2015

  1. Turkey must be a rare case of a polity that used to have two elective and policy-shaping [*] “estates” (Assembly and Senate), and then abolished one, then later went back to having two — but a different two (Assembly and President). Are there [m]any others?
    [* Some “estates” – not the very best term, true, but more accurate than “chambers” as it can include the Head of State – are elective but not policy-shaping (eg, the Austrian and Irish Presidents). Others are policy-shaping but not elective (eg, the House of Lords, the Canadian Senate, the Jordanian and Thai monarchs, and formerly the Greek President). But usually elective correlates to policy-shaping. By the latter I mean “leaves fingerprints on laws and other policy,” as opposed to legal powers. Eg, Queen Elizabeth II has vast de jure power but during her time has given Royal Assent to death warrants and also to Acts abolishing the death penalty: ie, 99% of the British statute book is not influenced by her personal views].


  2. If the HDP fails to win seats, the AK Party would be short of a manufactured two-thirds majority. If the HDP wins seats, it would compromise with the AK Party on a new Constitution. It is possible that the AK is getting more and more popular with each election, but wins fewer seats. Hopefully Turkey will embrace semi-Presidentialism.


  3. If the AKP needs the support of the HDP to pass legislation, I would imagine that pretty high up on the HDP wishlist would be electoral reform of one sort or another. Would the AKP give the HDP a drop in the threshold in exchange for a new constitution?


    • Electoral Reform for a new constitution might be too high of a bar for the AK Party, the AK party will most likely win more than 60% of the seats in Parliament and thus can have a referendum on a new Constitution even if the HDP wins seats or not. It looks like the AK party has become the natural governing party of Turkey and thus it is becoming a dominant party system.


  4. Is HDP running as a national list, like DEHAP in 2002, taking the all-or-nothing risk of the national 10% threshold?
    Or are they running as independents, as in 2007 and 2011, taking the spread out risks of the SNTV-like coordination problems in constituencies where they aim at more than one seat?


      • Well, strictly speaking it won’t be a national list, but a number of district lists, probably not in many districts outside Kurdish-majority areas.


      • There are no national lists in Turkey. There are 80-some districts. A party must run in some minimum number of them. I do not remember details, but I think I reviewed them in one of my posts around the time of the 2007 election (and I can’t say for sure if there have been changes or not).

        Obviously, the HDP won’t be able to collect any votes towards clearing the threshold in any districts where it does not run. I assume they are running in nearly all of them, therefore. If anyone has this information, I’d like to know.

        It is an interesting point about the presidential election, but one needs to be careful drawing inferences about legislative list votes from votes for presidential candidates–especially with non-concurrent elections.

        In any case, the HDP candidate won 9.8% of the presidential vote. Is there any reason to believe that the candidate would be less popular than the party nationwide? I have no knowledge to draw on here, other than the observation that there were only three candidates running for president, in an actual nationwide contest. And that makes me question whether the lists could do better across many districts and with additional parties, and perhaps some independent Kurdish candidates, running.


      • Makes sense; there are probably small numbers of Kurds (or other HDP sympathisers) living all around Turkey, especially in the cities, and even though they may not amount to any seats in their respective districts, they could help with reaching 10%.


  5. I’m pretty sure amendments can also be proposed by 3/5-majority in parliament and passed by referendum. I believe that’s how direct elections for president were instituted in the first place – which to some extent raises the question why Erdogan didn’t just push for full presidentialism in the first place.

    My reading of the polls, however (looking at the wikipedia page), is that even if the HDP Kurdish party doesn’t make it, AKP may just squeak past for a majority (nowhere near 2/3), as CHP and MHP together poll around 45%, better than the AKP’s low 40’s. If I recall correctly, the D’Hondt calculation is made at the district level, so AKP could still get a majority of seats despite not getting the majority of votes going to threshold-surpassing parties. If these polls are correct, and if the HDP does in fact pass the threshold, it looks like a hung parliament in which Erdogan will be forced to form a coalition or even – and this is the really interesting possibility – be forced to go into cohabitation with a prime minister supported by MHP, CHP and HDP.


    • Just read the FP article. It confirms that 3/5 majority is a path to constitutional amendment.

      But it remains unclear to me what Erdogan is pushing for, exactly. The article says a supermajority “will enable him to lay the groundwork for a new political system, which would shift power from the prime minister to a French-style executive president.” Isn’t that what he has already achieved? He’s a directly-elected president in a premier-presidential system, with the power to dissolve parliament and an electoral system that enables his party to control a legislative majority. It also says “Erdogan is a towering figure of Turkish politics. He dominates his party, which in principle is being run by his handpicked prime minister, Ahmet Davutoglu. ” Sounds like France to me…
      (my guess is that Erdogan is really looking for full presidentialism, with a cabinet responsible only to him and perhaps some decree powers.)


      • Turkey better be careful if it embraces semi-presidentialism, look at what happen to Sri Lanka and now it is going back toward parliamentarianism. What is wrong with parliamentarianism as almost all parliamentary democracies Prime Ministers can serve an unlimited amount of time whereas most Presidents are term limited with some exceptions.


      • Journalists tend not to be very good on this, and AKP people may not be helping understanding. I suspect when journalists and AKP-ites say “French style presidency” they mean a president who dominates both his own party and the legislature. Of course, we know that this is so in France only under certain conditions; presidential dominance is a contingent outcome. I assume Erdogan has in mind measures to make it the norm in Turkey, but I do not know what those might be. That is, the outcome he wants is not in dispute, but the intended reforms to bring it about are not clearly stated (at least in English language media).


    • In 2007, the advantage ratio (%seats/%votes) for the AKP was around 1.20, after removing the below-threshold votes from the denominator. If that degree of advantage (from, yes D’Hondt, and the moderate magnitude) were to occur again, the AKP would need around 55% of the above-threshold vote to get two thirds of the seats.

      If the HDP is close to, but below, the threshold, thereby wasting around 10% of the total valid vote, and there are still other parties that win some smaller vote shares, this is possible if the AKP polls in the high forties (as a percentage of all valid votes).

      In 2011, however, I calculate the threshold-adjusted advantage ratio as a lot lower (closer to 1.05). I am not sure what explains this.

      In other words, it would be far from a sure thing, but it is not out of the question, either.

      And as a couple of posts have noted, there are avenues towards constitutional amendment for which 60% of the seats is sufficient.


    • How do the CHP and the MHP feel about the peace process with the PKK, the continuation (and conclusion) of which must be the number 1 priority of the HDP? From what I understand the MHP at least is an ultra-nationalist grouping, so is not that positive about it. It might therefore be difficult to from such a broad anti-AKP coalition.


  6. How could Turkey improve it’s electoral system? If had no electoral threshold, it seem as the district magnitude is fairly moderate for a moderate multiparty system. What caused the excessive fragmentation of the party system before 2002 despite having a 10% threshold? Did Turkey have a regional volatile multiparty system before 2002? Are the party lists open or closed in Turkey?

    Also don’t overseas voters also vote in this election for the first time? How will that change the election result?


  7. Is there a precise figure where a threshold begins to reduce, rather than increase proportionality? When you look at good government agencies like the Venice Commission or even international courts like the European Court of Human Rights*, the absence of an accepted standard for the threshold has been problematic.

    The Turkish Constitutional Court previously invalidated a 25% provincial threshold while upholding the 10% national threshold.

    It seems to me that a mathematically informed brief by political scientists on electoral thresholds would carry considerable weight with the ECHR. It may even be that the party fragmentation that the threshold is designed to combat may be less threatening to democracy than the emergence of hegemonic parties with artificial majorities under high thresholds.

    *The ECHR is not an EU court and has jurisdiction over Turkey because the country belongs to the Council of Europe.


    • It would be interesting to see exactly how such a challenge would work. As I read it, the European Convention on Human Rights gives the right to ‘fair elections’. Now, I don’t know how enthusiastic the ECHR would be to extend this from the current interpretation of ‘no rigging of elections, universal suffrage, no unreasonable restrictions on standing for election’ to involving themselves in electoral system design.

      If ‘parties must be treated fairly in seat distribution’ was considered part of the Convention, I could see (hypothetically) Nigel Farage (or Nick Clegg, or Natalie Bennett) claiming that it’s unfair that UKIP won 12.6% of the vote but 0.2% of the seats, and demanding that the UK adopt proportional representation.

      A guarantee of proportionality might be considered an infringement on individual countries’ ability to choose electoral systems.


      • First Protocol, Article 3
        Right to free elections
        The High Contracting Parties undertake to hold free elections at reasonable intervals by secret ballot, under conditions which will ensure the free expression of the opinion of the people in the choice of the legislature.

        The ECHR has heard cases on thresholds. Presumably those cases were argued in terms of the second leg of the article rather than being dismissed for want of jurisdiction.


      • Alan, I’m not saying that the ECHR don’t have jurisdiction over electoral rules. I’m saying that they might not want to intervene due to concern about setting precedents regarding standards of proportionality.


      • Of course, for all those countries with mixed-member systems the threshold does not apply to all the seats. I suppose, however, that I was more generally listing countries with majoritarian systems than focusing on the threshold aspect.

        I would be surprised if the French government’s commitment to introducing an element of PR is much stronger than the Blair government’s.


    • “Is there a precise figure where a threshold begins to reduce, rather than increase proportionality?” Surely all threshold reduce proportionality.


      • “Surely all threshold reduce proportionality.” Mechanically, yes, the effect can only be downward–or no effect. One might say that a well functioning threshold is one that no one challenges without clearing. So, if the threshold effect is well understood by party leaders and voters, such that no one comes up short, then a high-threshold system can be very proportional. Look at the classic (1970s or so) West German system. There were essentially no votes cast for parties falling below 5%. Then look at Germany 2013. Same threshold, same basic rules. But very different levels of proportionality.

        Of course, the Turkish system would introduce disproportionality even without a threshold, simply because of variable magnitude, rather than nationwide PR.


      • The Venice Commission on the 2002 Turkish general election where the governing party first introduced the 10% national threshold:

        26. More than half the electorate was deprived of representation and those parties that were elected had a percentage of seats twice that of their percentage of votes, so that a proportional system became a majority one.

        The commission argues that an escape clause, such as winning one district, cures many of the problems of high thresholds as well as allowing regionalist parties. Absent some mathematical formula that predicts when a proportional system becomes a plurality system, an escape clause may be a better approach. They favour electoral thresholds of 3%-5%.

        The commission is considering developing a ‘common European rule’ although of course it is only a consultative body and it recognises the difficulties with one size fits all rules.

        If, as the Conservative manifesto threatens, the UK attempts to withdraw from the jurisdiction of the ECHR, it’s hard to see the CoE allowing the UK to retain membership. The last country to withdraw from the jurisdiction of the ECHR was Greece in the time of the colonels. However, the CoE would then have no members that use FPTP and would presumably be more open to a common European rule on proportionality.


      • Alan, France would still be in the CoE, thus creating resistance to a Europe-wide proportionality standard. Also, one would imagine that a common proportionality standard would mean Erdogan would leave the CoE, given that staying in it would destroy his chances of re-election. Mixed-member majoritarian systems might also come under question. In an era where deepening of European regulation is becoming increasingly unpopular, I don’t think Europe-wide standards for electoral systems would be popular or accepted.


      • Not only France – you’d still also have Hungary, Greece and Italy. Either way I find it difficult to believe the ECHR will decide something like that is a matter of human rights.


      • Somehow I managed to forget to mention Lithuania, Ukraine, Russia, Georgia, Azerbaijan and, less importantly, Andorra, San Marino and Monaco (Liechtenstein has an 8% threshold, although there have been recent proposals to lower it). Regardless of one’s opinion on the importance of proportionality, Turkey and some of these other countries currently have much greater problems with regard to democracy.


      • The Venice Commission mentions a threshold no greater than 3-5% and/or a lifeline clause for parties that win a seat in their own right.

        Russia meets that standard.

        Ukraine meets that standard.

        Lithuania meets that standard for individual parties but not coalitions.

        Hungary meets that standard for individual parties but not coalitions.

        Georgia meets that standard in national elections.

        France does not although the French government remains committed to introducing an element of proportionality in the national assembly.

        Azerbaijan does not.

        I have not researched the microstate electoral thresholds because they are mentioned by the commission as special cases although I will note that Liechtenstein was extensively criticised by various bodies, including both the Council of Europe and the Venice Commission as well as Human Rights Watch, for the recent constitutional amendments that vested wide and unaccountable powers in the prince.

        I concede the lengthy, majestic and glorious electoral tradition of Azerbaijan may be against me.


      • JD’s comment should probably go here

        Of course, for all those countries with mixed-member systems the threshold does not apply to all the seats. I suppose, however, that I was more generally listing countries with majoritarian systems than focusing on the threshold aspect.

        I would be surprised if the French government’s commitment to introducing an element of PR is much stronger than the Blair government’s.


  8. Does anyone know if Turkey is one of those countries that tends to get results out very quickly, or is it one where the results are not known for several days after the election? Related: are there exit polls, and if so, do they have a decent track record?


      • That would be an advantage ratio of 1.15 for the largest party, which would be lower than in past elections (around 1.2). The strategy and success of those political forces that are around the threshold makes all the difference!


  9. Pingback: Mexican elections 2015 | Fruits and Votes

  10. The coalition could be with the MHP rather than the HDP. In fact, that would be more likely, perhaps.

    I am unsure whether the government-formation procedure makes a minority government feasible.


  11. From what I understand of Turkish politics, the AKP is furthest from the MHP ideologically and on practical political issues. For example, seeking MHP support severely contrains the government’s options on the Kurdish issue, more even than a deal with the HDP itself does. If they need a coalition, one with the CHP would probably be the best option, followed by the HDP. But given how divided the opposition, seeking votes issue by issue as a minority government is feasible and probably what they will wound up doing.

    In terms of results, the key result is that not only was the HDP strategy successful -they easily cleared the threshold- but their percentage rise matches the percentage fall for the AKP. This supports my theory that alot of Kurds voted for the more Islamic parties in the past to stick it to the more Kemalist parties.


    • However, the AKP ended up winning about the same vote share as the pre-election polls suggested (according to the article linked in the post). The HDP, on the other hand, polled higher than the range the article cites (which was 8.5-11.5).

      So, unless the polls were just wrong, it does not seem as if a swing from the AKP to the HDP explains these parties’ final results fully.

      The MHP also came in somewhat higher than the polls.


      • The polls could have been wrong, or there could have been a voter flow from the AKP to the HDP, and then a last minute flow from the CHP to the AKP to prevent a hung parliament or “rainbow coalition” and reduce the possibility of a Kurdish party being influential (sort of like the last minute voter flows to the Tories when it looked like the Scottish Nationalists would hold the balance of power).


        • Odd that every comment here refers to the HDP as the Kurdish party. The HDP was founded in 2012 as the political wing of the Peoples’ Democratic Congress, a union of numerous left-wing movements that had previously fielded candidates as independents, the Turkish version of SYRIZA and Podemos. Meanwhile, the Kurdish BDP had succeeded the Democratic Society Party (DTP) in 2008, following the closure of the latter party for its alleged connections with the Kurdistan Workers’ Party (PKK). At the 2014 municipal elections, HDP ran parallel to BDP, with the BDP running in Turkey’s Kurdish-dominated southeast while the HDP competed in the rest of the country. The BDP was co-chaired by Selahattin Demirtaş, and after the two parties were re-organised in a joint structure, on 28 April 2014, the entire parliamentary caucus of BDP joined HDP, whereas BDP was assigned exclusively to representatives on the local administration level. So on 22 June 2014 Selahattin Demirtaş became co-leader of the HDP. However, the other co-leader is the marxist Figen Yüksekdağ. She was taken into custody in 2009 due to her political activity. She founded the Socialist Party of the Oppressed (ESP) shortly after in 2010. It was a participant in the Peoples’ Democratic Congress. She resigned as leader in 2014 to join the HDP (with which the ESP merged later the same year), and on 22 June 2014 was elected the co-leader of the HDP. The ESP was a Marxist-Leninist political party that defined itself as “a militant revolutionary socialist party fighting for a workers’-labourers’ federative republic in Turkey and Northern Kurdistan.” Some sources hold the view that ESP was a legal front for the banned Marxist–Leninist Communist Party.

          Turkey had previously been blessed with a plethora of centrist and leftist parties falling below the 10% threshold. In 2002 five major parties got more than 5% but less than 10%. In 2007 the Democrat Party got 5.42%,and the Youth Party 3.04%. But this time the only surviving splinter was the left-wing Patriotic Party (formerly Workers’ Party) which got only 0.35%. The HDP was the left unity party.

          So the HDP is a fairly far left party with strong support from the Kurdish and Armenian minorities. At one point Figen Yüksekdağ ruled out negotiating with the centre-left secular CHP since they were ‘closed to dialogue,’ but they are the only plausible ally for the HDP. That leaves the king-maker as the right-wing nationalist MHP.

          I think. But I have not updated my research.


        • Most commentary that I saw referred to the HDP as “pro-Kurdish” or “Kurdish-aligned”, or similar wording. It is indeed not a specifically Kurdish party. In fact, one article I ran across said it was more like Syriza than any other party in the wider region.


      • I think that after the 2002 election, the Kurds recognised that a purely ethnic party just wouldn’t work; not only would it lack broad electoral support, but it would attract more unwanted attention from the authorities. They tried running as independents, but for real political power, they needed a party. The far-left was an unfilled niche, and an alliance with the Kurds and the far-left would be mutually beneficial.

        I don’t know if the Kurds, as an ethnic group, are particularly left-wing, but it certainly seems like the best option for Kurds to hold political power.


      • The Kurds consistently poll as the most pro-American group in the Muslim world. I think only the Kosovars polled higher. It would be very ironic if they allied with the far left in Turkey, but I suppose no weirder than most of what’s passing for alliances in the Middle East at present.
        Caveat that it’s usually Kurds in Iraq who get polled on attitudes to US foreign policy… maybe Kurds in Turkey have different views.


      • I think pro-US sentiment amongst Iraqi Kurds might be overstated due to gratitude at the removal of Saddam Hussein, but the Kurds do seem somewhat more pro-US than other Middle Eastern ethnic groups (which isn’t saying much).

        As for the coalition with the far-left, it really comes down to a process of elimination. There was no question of the AKP joining up with the Kurds, and the Kemalist parties did nothing for the Kurds in government, which only left the far-left.


      • The HDP certainly got some votes from non-Kurds and seats in non-Kurdish majority areas, especially around the coast, but the map of HDP support matches that of the Kurdish population almost exactly.


  12. HDP has already ruled out a coalition with the AKP, apparently also ruling out support for an AKP minority government: “”We will not form any coalition [government], whether inside or outside, with the AK party,” – aa.com.tr/en/news/533203–turkey-hdp-leader-rules-out-coalition-bid-with-ak-party


  13. From Blomberg News, the MHP has also ruled out a coalition, and for good measure has reaffirmed its rejection of the peace process with the Kurds (presumably ruling out an anti-AKP rainbow coalition). That leaves the CHP as a colaition partner, or new elections. The CHP has expresed interest, but the Bloomberg article points up “a serious issue of incompatibility”,



      • But then again, In 2002 the 10% threshold gave the AKP a false majority, leaving the CHP robbed of its potential coalition partners. Do they still remember? If so, could we see a rainbow coalition to lower the threshold?


  14. Pingback: Luxembourg term limit referendum | Fruits and Votes

  15. Pingback: Some further thoughts on the Turkish election, 2015 | Fruits and Votes

  16. Pingback: Turkey 2015: The AKP’s non-majority is not a surprise, and higher thresholds can only increase the largest party’s seat share | Fruits and Votes

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.