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

The reporting on the Turkish election results is treating the AKP’s fall to less than half the seats as a really big surprise. There was even a column in a Turkish newspaper that suggested the high threshold may have been a reason for the non-majority. This post attempts to set the record straight: these claims are bogus.

Was the AKP losing its majority a surprise?

Pre-election polls suggested the AKP would earn around 40% of the vote, which proved spot on. (The link there is an article from May 28, which I cited in my pre-election post.)

The Turkish electoral system is sufficiently proportional (although only modestly so) that a leading party with around 40% of the votes would be highly unlikely to get a majority of seats–UNLESS there were sufficient numbers of wasted votes for parties falling below the threshold. That threshold is set at 10%, and applied nationwide, in spite of all seats being allocated in 85 multi-member districts. In the past, the AKP had won parliamentary majorities on less than half the votes (and as low as 34%), but only due to a large below-threshold vote for minor parties and independents (many of them Kurds who this time coalesced behind the HDP).

Once the HDP cleared the threshold, there was essentially no way for the AKP to get a majority. In the same pre-election polls, the HDP was in the range of 8.5-11.5%. Thus it was a knife edge whether it would clear, but hardly a big surprise. And if the HDP had a good chance of clearing the threshold, the AKP could not get a majority unless its vote share was considerably higher than anticipated.

Did the threshold actually hurt the AKP this time?

Regarding that threshold, there was a piece in the Monkey Cage blog that was otherwise a very fine overview of “How the Kurds upended Turkish politics“, but in which one point makes no sense:

As editor of the English-language Hurriyet Daily News Murat Yetkin shrewdly pointed out, the AKP may have been a victim of its own dependence on the unfair 10 percent threshold rule. If the threshold had been lowered to 5 or 7 percent, argued Yetkin, the AKP still would have been prevented from adopting Erdogan’s presidential system, but its parliamentary majority would have been salvaged.

That point actually does not seem so shrewd to me. Yetkin himself further says:

Because of the complicated calculation system that the 10 percent threshold brings with it, the AK Parti has lost its parliamentary majority.

The system is indeed somewhat complex, but there is nothing complicated about how thresholds affect the outcome. If you attain the stipulated minimum vote share nationwide, you qualify for seats in any of the 85 districts in which you have sufficient votes.

How could it be that a LOWER threshold would INCREASE the seat share of the largest party?

he only thing I can imagine Yetkin may have meant–although he does not spell this out–is that the HDP vote itself would have been lower if the threshold had been lower, and implicitly, the AKP vote would have been higher. In other words, HDP, according to such an argument, may have benefited from strategic voting by voters who wanted to ensure it got 10%. Given that it got almost 13%, which was–as I noted above–more than pre-election polls predicted, while the AKP result was in line with those polls, such an argument (which, again, Yetkin did not actually make) seems a stretch.

A “normal” result, given the electoral system and HDP being over the threshold

The bottom line is that the result of this election, while quite a watershed for Turkish politics, has given Turkey a fairly “normal” result for an electoral system of its average magnitude (around 6.5). The threshold, in the end, did not bar any significant political force from representation, unlike in previous elections. A largest party winning around 47% of the seats on about 41% of the votes is nothing out of the ordinary. Whether Turkish-Kurdish politics can be as “normalized” as the election result itself is another matter. Stay tuned to the government-formation process and aftermath for clues.

10 thoughts on “Turkey 2015: The AKP’s non-majority is not a surprise, and higher thresholds can only increase the largest party’s seat share

  1. I have to say I find that commentary from The Monkey Cage absolutely mystifying.

    Regarding normality, though this is Turkey’s first hung parliament since 2002, before that election Turkey had last had a majority before the 1991 election.

    • Mystifying indeed, especially from a blog that exists for the stated purpose of bringing political science insights to a wider readership.

      I had hoped it was clear that by my remark on normality, I meant for a moderate-magnitude districted PR system (ignoring the threshold, which had little impact this time), not that the outcome is normal for Turkish politics.

  2. One area of elections research that I wish would be examined more would be notional conversions of proportional system to single member district systems, and vice versa. With allowance for the fact that the election and voting strategies would change, and therefore the votes allocated to each party, how would the same party system under proportional representation functioned using single member districts?

    In the case of the last Turkish election, obviously you get an AKP majority using single member districts, but both the Kurds and the CHP have regional strongholds in opposite ends of the country, and would have each returned a solid block of deputies. The party that would have been screwed would have been MHP, who would have struggled to win any seats as has been the experience of UKIP and FN in France.

    With single member districts, you probably would have seen a Kurdish regional party coalesce much earlier. Also under proportional representation there will probably be an AKP minority government. I’m not sure how much this really differs from an AKP majority government. I’m not sure if there are any practical differences between a single party minority government, and a single party majority government with rebellious MPs.

    Rerun the last UK election under proportional representation, and the likeliest result is a Conservative government dependent on UKIP, as opposed to a Conservative government dependent on euroskeptic Tory MPs.

    • Good points. The closest example to this type of work that I can think of is Adam Ziegfeld, 2013, “Are Higher-Magnitude Electoral Districts Always Better for Small Parties?” Electoral Studies 32 (1): 63-77.

      It simulates Indian elections under some form of PR, and Israeli under FPTP.

  3. Em 1996, a friend of mine made a simulation about who could be the portuguese elections results with FPTP.

    1975 – MPs in reality – MPs in FPTP

    PS (socialist) – 116 – 160
    PPD (centre-right) – 81 – 76
    PCP (communist) – 30 – 11
    CDS (christian-democrat) – 16 – 0
    MDP/CDE (“satellite” of PCP) – 5 – 0
    UDP (maoist) – 1 – 0

    1991 – MPs in reality – MPs in FPTP

    PPD/PSD – 135 – 216
    PS – 72 – 2
    PCP – 17 – 8
    CDS – 5 – 0
    PSN (pensioners’ interests) – 1 – 0

    The small differences between the sum of the real results and the results with FPTP is because he probably did not count with the emigration circles

    http://aleph20.letras.up.pt/F/LKDNXDNGK4X7T3ABPV3IGVBYYNLJAFP8HIBAD9ILH9AK4GATRM-14660?func=find-b&amp=&amp=&amp=&request=000115667&find_code=SYS&local_base=FLP01&pds_handle=GUEST

  4. Pingback: 2015b: Greece and Turkey back to the polls | Fruits and Votes

  5. Pingback: Turkey 2015b: Erdogan’s gambit pays off (apparently) | 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 )

Google+ photo

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

Connecting to %s