Thursday, March 27, 2014

Final Thoughts & Predictions about Turkey's Local Elections

A few final thoughts and questions about Turkey's elections, which will take place on Sunday. I organize it around some points that we should be pretty certain about, and some questions that still remain.


  • The AKP retains a large core of supporters who will vote for it (or its candidates) in the upcoming elections.

Decades of political science research into voter psychology shows that most voters have a long-term leaning towards one party or ideology. That is rooted in some combination of one's upbringing, of our individual psychology, and possibly even our genes.

This same research also shows that people are better at "motivated reasoning" than at objectively examining the evidence. We tend to fit new facts into pre-existing cognitive schemes. In other words, think about the effect of all the leaked corruption tapes of the past several weeks. For opponents of the Erdogan government, these tapes "prove" what they have known all along: that Erdogan is corrupt and engaged in some seriously dirty dealing. But for AKP supporters, it "proves" what they have known all along: that a nefarious plot of coup-mongers and outsiders is seeking to undermine Erdogan by using dirty tricks. Thus, it's not surprising that journalists have found plenty of (anecdotal) evidence of AKP supporters dismissing the corruption allegations against Erdogan. Our minds work hard to avoid cognitive dissonance, so supporters of Erdogan will tend to reject evidence that Erdogan is a corrupt leader. (To add a further twist, people vary in their preference for maintaining cognitive consistency over updating their beliefs, and my suggestion is that AKP supporters tend to fall into that camp that prefers consistency). Expect this to continue, unless and until the evidence of his corruption either becomes overwhelming or until trusted sources within the AKP validate the charges against him.

  • The AKP has an impressive "machine" for mobilizing its supporters, and it uses state resources to enhance these efforts--giving it a considerable advantage over any of the opposition parties.

As a political scientist, I can only admire the AKP's ability to get its supporters out to rallies, to the polls, etc. It is miles ahead of CHP in this regard. One caveat: this has been true for a while. Certainly, it was true in 2009 and 2011. So, the AKP's mobilization advantage is probably more of a constant than a variable in terms of explaining what will happen on 30 March, unless you believe that the AKP has significantly improved its ability to mobilize voters since then.

(The good news for CHP voters? It's impossible to fall off the floor, so the opposition parties can only get better at attracting and mobilizing voters!).

  • As a result, the AKP will receive by far the most votes, and control the largest share of mayorships and councils after the local elections.
The best available evidence suggests that AKP will get about 42-45% of the vote nationally. This includes polling data as well as economic voting forecasting models.

For the CHP, the target should be 30%. Round numbers don't have any true value, but there would be something symbolic for the party in hitting 30% to show that it can mount an effective challenge as it moves on to the presidential campaign this summer and next year's (presumptive) parliamentary elections. Of course, the national vote doesn't matter in any substantive terms. But it has a lot of symbolic value: a strong showing by the AKP will embolden Erdogan to run for president, while a poor showing may deter him and/or convince AKP members to push for another candidate (and, possibly, another prime minister).

In terms of the three biggest cities, CHP will win Izmir. AKP will likely win Istanbul (though maybe not fairly...more on that). Ankara will be the major battleground, which AKP may well hold on to due to the opposition's failure to coordinate (more on that, too).

  • The economy was roaring in 2011. It is still growing now. That will benefit AKP, as it has in every election since 2002.

Millions of trees have died to allow political scientists to document the strong relationship between the national economy and the incumbent's vote share. We call this "economic voting." Mostly, it occurs among that portion of the electorate that is not so closely attached to one party. When the economy is growing, they decide to support the incumbents. When things go badly, they decide it's time to give the opposition a chance.

Long story short: AKP first came to power in part due to economic voting as Turkey was coming out of an economic crisis in 2002. Then it won re-election in every election since due a strong economy. And the economy is still growing at a healthy rate of 3% or so at present.

It may seem unfair, but there are many voters who will not worry too much about corruption or abuses of power as long as long as it does not harm them directly. But that's how it goes. And, remember, this is only among those voters who can even be persuaded that Erdogan has done something wrong (revisit the first point about AKP supporters' motivated reasoning).

Turkey may well be headed towards an economic crisis in the next 2-3 years. But it won't happen before the local elections. So AKP will benefit from positive economic conditions.

  • The AKP government will engage in electoral fraud, particularly in Istanbul

The stories are mostly anecdotal, but there are a lot of them: voters falsely registered at people's addresses, phantom buildings full of registered voters, oversupplies of ballots, etc. More to the point, the Erdogan government has demonstrated a willingness to subvert the rule of law to increase its power and get its way. Unless you truly believe that Erdogan considers elections to be too sacred to corrupt, then you would have to expect him to continue the same behavior.

In terms of where, electoral fraud is most likely to occur in contested municipalities that the AKP currently controls. The most obvious is Istanbul. In addition to being the largest city and a true electoral battleground, it also holds a lot of symbolic value for Erdogan. 

  • CHP and MHP, the two main opposition parties, will end up missing some crucial opportunities by failing to coordinate their candidates.

Prediction: the vote margin between AKP and CHP in both Istanbul and Ankara will be less than the MHP candidate's share of the vote. Meaning, of course, that a concerted effort by the CHP and MHP to unite behind a single mayoral candidate in either city could have produced an opposition victory.

Meanwhile, there are several western municipalities where a MHP incumbent may lose to an AKP challenger. For example: Manisa and Balikesir. If the two opposition parties had been smart and read their Cox, they would have traded candidacies: MHP would have supported CHP candidate in Ankara and CHP would have supported MHP candidates in Manisa and Balikesir. Some opposition supporters are apparently trying to organize this sort of coordination themselves, but it is unlikely to work at this late hour.

  • What will turnout be?

Despite having mandatory voting laws, turnout in recent parliamentary elections has been around 80-85%. In the last election, 1 in 6 eligible voters abstained. Normally, turnout results from socioeconomic factors: the poor, less educated, and younger voters (basically, the more marginal members of society) tend to vote less. However, people also don't turn out to vote because they either don't care about the outcome, don't believe that their vote matters, or don't like any of the choices.

One can see how that latter group would matter in light of the Gezi Park protests. Last summer, this was a common claim: the protests had mobilized a group of young, educated, and urban individuals, who had grown up in the increasingly prosperous post-2002 Turkey. As the story goes, these individuals had become more frustrated with life under Erdogan in the years leading up to 2013, but saw no effective outlet for their dissatisfaction in the media or opposition parties. Thus, the Gezi Park protests mobilized large numbers of them into political action for the first time. Now, will these protestors turn out en masse to vote for opposition parties? If so, how much will that matter?

This is a tough question to answer, and it also affects pre-election polling efforts (which make assumptions about turnout rates among different groups of voters). If polling firms fail to predict a surge in protest votes, then they could end up overestimating AKP electoral support. But maybe this surge in votes won't really happen. It is unclear whether the protestors really had been that politically apathetic in 2011 and earlier.


  • Just how much will all the scandals and leaks affect voters?


It is surprisingly hard to estimate just how the scandals might damage the AKP's standings. Particularly when you have scandalous leaks being released just days before the election. In the end, the confusing leaks and counter-narratives may confuse voters, it may lead some to rally around Erdogan, or it may lead them to decide it's time for a change. In the end, one would expect that the scandals have had to hurt the AKP somewhat, but it is harder to say how much.

There are, of course, many other things we don't know. To paraphrase Don Rumsfeld, there are things we don't know we don't know. But these are, I think, some of the major questions going in to Sunday's elections.

Thursday, March 20, 2014

What Next for Turkey?

"Formal democratic institutions are widely viewed as the principal means of obtaining and exercising power. Incumbents violate these rules so often and to such an extent, however, that the regime fails to meet conventional minimum standards for democracy."

"Although elections are regularly held and are generally free of massive fraud, incumbents routinely abuse state resources, deny the opposition adequate media coverage, harass opposition candidates and their supporters, and in some cases manipulate electoral results. Journalists, opposition politicians, and other government critics may be spied on, threatened, harassed, or arrested."

All of that comes from an insight article from over ten years ago called "The Rise of Competitive Authoritarianism," (ungated) by political scientists Steven Levitsky and Lucan Way. They were using it to describe many post-communist regimes in places such as Russia, Serbia, and Ukraine. But doesn't it describe Turkey under Erdogan quite well too?

This article provides useful insight into where Turkey is right now, and what might happen in the coming months and years.

"Competitive authoritarianism": both words are important. It is, at its core, not a democratic regime. But nor is it a fully authoritarian regime either. To wit: "Although the electoral process may be characterized by large-scale abuses of state power, biased media coverage...and an overall lack of transparency, elections are regularly held, competitive (in that major opposition parties and candidates usually participate), and generally free of massive fraud."

Similar undermining of the legislative process, judicial independence, and media freedom characterize competitive authoritarianism. But, in each case, it is only undermined, not eliminated as a viable force.

And that gets to a key point about competitive authoritarianism: it is inherently unstable. The uneasy balance between democratic institutions and authoritarian practices reflects the inability of either the incumbent rulers to crush the opposition or of the opposition to pressure the incumbents into acceding to true democratic practice. This tension forces a dilemma upon the incumbents when the opposition challenge rises: crushing the opposition may be too costly (or too damaging internationally), but allowing a true opposition challenge could mean losing power.

This is where Erdogan finds himself right now. After the Gezi protests, the corruption investigations, and the ongoing leaked phone conversations, it is clear that a large opposition movement is mobilized against his government. And the resilience of the protests from June until now shows that they will not easily be deterred by a show of force.

So far, all of Erdogan's actions have been designed to stifle opposition and to project a sense of national unity behind his governance. This is, as I've suggested before, partially a result of his governing style and psychology of his supporters. This also, I think, reflects his vulnerability: if half of what is in those leaked tapes is true, he would face serious criminal repercussions were he to lose power. So he is, in a sense, "gambling for resurrection."

This is bad news for Turkey in the short term, and it certainly explains his efforts to undermine judicial independence (reassigning police and judges), to block media (Twitter!), and it suggests that the likelihood of the 30 March local elections being fair--at least, or especially in, Ankara and Istanbul--is low. He is going to fight hard to hang on to power.

I still find it hard to believe that this will be a winning strategy for much longer. Levitsky and Way conclude by noting that competitive authoritarian states that are closer to the West politically and geographically have been more likely to transition to democracy, while states that are further removed have been more likely to transition into full-blown authoritarianism. It seems to me, at least, that there are large enough interests within Turkey that (want to) see it as a Western-oriented democracy. And, unlike in some former Soviet republics, there is no powerful neighbor to prop up an authoritarian regime. So I would argue, as I have before, that Turkey's medium-to-long term path is still towards greater democratization. But it may be a terrible rough path to get there in the short term against a leader determined to hold power at all costs.

The (evil) genius of Geert Wilders?

Last night, Geert Wilders, head of the radical right Party for Freedom in the Netherlands, stirred up quite a commotion. At a rally in The Hague, Wilders asked party supporters during a speech whether they wanted more or less Moroccans in the city and in the Netherlands. The crowd responded by chanting "less! less! less!", to which Wilders replied "We will arrange that." (Video of it here, in Dutch, of course).

Naturally, many in the Netherlands are outraged by these comments, which seem to call for mass deportations of Moroccan immigrants. A poll conducted today suggests that close to two-thirds (in Dutch) of Dutch citizens do not approve of Wilders' comments. And while it's hardly an objective metric, a Facebook page calling for a criminal investigation into Wilders' comments (on the grounds of inciting ethnic hatred) gathered over 40,000 "likes" within a day (the poll suggests split opinion on the question). Even the Washington Post noticed.

So how could provoking such a popular backlash be a good move for Wilders?

All of this occurred yesterday as votes were being counted in local elections around the Netherlands. And here is the catch: Wilders' PVV only contested two of several hundred local elections taking place. And the PVV finished first in one of those, and it lost vote share compared to its performance in the last local elections in both. So the PVV was really a non-entity in the local elections, compared to the other major parties in the Netherlands. And, yet, he's the one that everybody is talking about the day after the elections.

As they say, there's no such thing as bad publicity. And this fits a pattern that Wilders, and other radical right parties around Western Europe, have seemingly used pretty well: say something provocative, wait for the mainstream backlash, and then claim the moral high ground of being the only ones willing to speak unpleasant truths (remember that Islam video Wilders did a few years ago?). 

In the Dutch proportional electoral system, the PVV can afford to focus on the minority of Dutch voters who were not bothered by his anti-Moroccan statements. According to the same survey cited above, 89% of PVV supporters approved of his statement. He is, of course, speaking primarily to them, as well as to like-minded voters that he might drag away from other parties. The trick that Wilders has pulled is that he has managed to make an event that his party had very little to do with very much about him.

Tuesday, March 11, 2014

Is This the Beginning of the End for Erdogan?

After the death of Berkin Elvan, a teenager who died of injuries suffered after being hit in the head by a police tear gas canister during last summer's Gezi Park protests (while going to the store to buy bread for his family, reportedly), new protests erupted in Istanbul and other major Turkish cities. That led me to post this comment on Twitter earlier:



This Time Is Different?
Why do I say that now, when nearly nine months of protests plus apparent revelations of massive corruption schemes have failed to dent Erdogan's hold on power? Two reasons, one more substantive than the other:

First, anger about the death of an innocent child can galvanize protestors--and draw neutrals into supporting the protests--in a way that more abstract and political concerns (such as environmental degradation, government abuses of power, or corruption) probably can't. People respond more strongly to human stories than to abstract arguments or facts. It's easier to dismiss a heavy-handed prime minister or government bribe-taking as "the price of progress" or "they all do it." But the death of a 15 year old boy? That's tougher to ignore.

Second, and perhaps more importantly, is an apparent shift in the media environment. Before and during the Gezi protests, a major concern was media self-censorship. Infamously, CNN Turk showed a program about penguins during the first night of major protests in Istanbul. Media coverage was far more sympathetic to the government. Among the various leaked recording in the past several weeks, one appears to show Prime Minister Erdogan instructing a media chief not to cover a speech by an opposition party leader.

Today, at least, the media coverage seemed to be different:



This change in media coverage, if it reflects more than a one-day trend, could be a significant turning point.

You Can Only Accept What You Receive
 John Zaller's famous (among academic political scientists, anyway) Receive-Accept-Sample model of public opinion helps us to understand why.

Most people receive their information about politics from the media. The media thus plays a crucial 'gate-keeper' role: people will receive the information they report and not receive the information that they do not report. In the years leading up to last summer, the fact that much of the media had been intimidated into providing friendly coverage of the Erdogan government meant that most Turks simply weren't hearing any negative news. Obviously, the more attention you pay to the news, the more likely you are to receive any information.

Upon receiving new information, an individual can either accept or resist this new information. In other words, they can either believe information to be true or reject it as false. How do we know when each will occur? One factor is party support. All else being equal, AKP supporters will be more likely to accept positive information about the Erdogan government and to reject negative information. A second factor is one's level of political sophistication. A highly sophisticated individual (i.e., one is who educated, pays close attention to politics, etc) is better able to discern which information s/he wants to accept and which s/he should reject, while a less sophisticated citizen may accept everything s/he hears as being true. Ultimately, one's opinion is something like the balance of information on the topic: the more positive thoughts you have about the issue or person, the more likely you are to give a positive evaluation if asked about it. And vice versa.

Put these two factors together, and the importance of media coverage emerges for one particular group: those who are moderately sophisticated about politics. In other words, a fairly large group of people. Why? Highly sophisticated people are the ones most likely to have strong political leanings, and the ones least likely to be persuaded by new information. This is why opponents of the Erdogan government have tended to be more educated; they are the ones who can continue to resist the drumbeat of positive messages in the media and hold critical opinions. Unsophisticated citizens pay little attention to politics, and so they are the last ones to form an opinion or to change it. The strongest action involves those people in the middle (in terms of sophistication). They pay enough attention to see and hear the media coverage, but their ability to resist a continuous flow of "one-sided" media coverage (i.e., consistently positive coverage of the AKP) is limited.

This means that an increasing flow of dissenting information from the media will have the greatest ability to changes such people's political attitudes. So if the media begin to report consistently on the negative claims against Erdogan, the corruption allegations, etc, and to give equal time to positive coverage of opposition parties, then the balance of pro- and anti-Erdogan ideas in their minds will begin to shift. If media coverage becomes more balanced, then mass support for the Erdogan government may begin to erode. And it will be these sorts of people, moderately sophisticated about politics (and perhaps, politically moderate as well) who quit supporting Erdogan.

Mind you, this process can take a while. Zaller's case study of this shift from a "one-sided" to "two-sided" information flow was the Vietnam War--American media went from being relentlessly positive about it until 1968, after which negative coverage of the war became increasingly common. But public support for the war certainly didn't erode overnight, and many Americans continued to support it even as the coverage turned more negative. So don't expect a rapid shift overnight!

Erdogan, Not the AKP
I was careful to frame my question in terms of Erdogan, and not the AKP. One thing that should be clear for now is that the AKP represents the interests of a considerable portion of Turkish society--conservative and often pious Sunni Muslims. That is not likely to change. 

What could change is Erdogan's standing with the electorate and with his own party. Continuing negative coverage of his alleged corruption and of his authoritarian tendencies could erode his popular support--and, by extension, his party's. That isn't likely to happen before the local elections at the end of March, to be sure. But, with presidential elections this summer, and parliamentary elections due before next summer, skittish AKP leaders may decide at some point that Erdogan has become a liability for his party.

But, a powerful though divisive Prime Minister being forced out by his own party despite winning three national elections handily--who would ever do such a thing? Well, it happened to Margaret Thatcher. Nonetheless, the Conservative Party went on to win its first post-Thatcher election with a new, more moderate leader. One could easily imagine a post-Erdogan AKP winning general elections in the summer of 2015 with Abdullah Gul or the like as the party leader. Nonetheless, it would surely benefit Turkey to move forward from the Erdogan era towards a less confrontational and divisive style of politics.

Thursday, March 6, 2014

Is the AKP Losing Electoral Support?

A new national poll is out from Metro Poll. Here is a chart showing a tracking poll over the past several years.

It's a bit messy, but the most recent poll appears in red in the final column. The results show that the governing AKP (Justice & Development Party) of Prime Minister Recep Tayyip Erdogan sits on 37%, the center-left CHP (Republican People's Party) on 27%, and the nationalist-right MHP (Nationalist Action Party) just shy of 15%. Looking back to a year ago (Ocak is January, Nisan is April), AKP support was not greatly different than now, but support for the CHP was 15-18%, and MHP was hovering around 10%. What was noticeable was the large bloc of undecided respondents in early 2013 ("kararsizim" is undecided, "cevap yok" is no answer): 23% and 25% respectively. The number of undecideds has dropped as support for the opposition parties--and especially CHP--has increased.

This fits the interpretation of electoral trends I have been pushing since last summer. My argument has been that AKP has a core group of supporters--conservative, religious Anatolian voters and connected business persons--who make up about 30-35% of the electorate. When support for AKP went as high as 50% (almost) in the 2011 general elections, this was due to other groups of "swing" voters, who supported AKP due to its strong (apparent) record of competence in economic management and governance--which contrasted sharply with Turkey's experience in the crisis-ridden 1990s.

My argument is that enough information had come to light by early 2013 to tip these voters (who had been weak supporters of AKP in the 2007-11 era) into the "undecided" category. By early 2013, they would have reason to feel ambivalent: approving of AKP's economic record, skeptical if any of the opposition parties was ready to govern, but worried about the "creeping authoritarianism" of Erdogan. So, they were not quite sure if they could continue to vote AKP, but as yet unconvinced to vote for an opposition party. The events of the past nine months since April 2013--the crackdowns on the Gezi Park protestors and the revelation of were enough to push these individuals from indecision to support for an opposition party.


What does all of this mean then? On a broader, conceptual point, this suggests a moderate (and evidence-based) interpretation of AKP support. In the past six months, there have been two competing perspectives. One group of observers ("hopeless optimists"?) has waited for the AKP to collapse under the weight of the protests and scandal revelations--ignoring the fact that there are a significant number of Turks who support the AKP strongly and who will engage in heavy doses of motivated reasoning before abandoning that support. The other group ("professional doomsayers"?) has perpetually argued that the AKP's support remains high even in the face of these scandals, and that the opposition parties pose little meaningful threat. A lot of this has been based on by "man on the street" journalism in AKP-loyalist cities or neighborhoods (examples here and here). The point is that neither of these claims is based on a look at systematic evidence.

The truth, on the basis of this evidence, is somewhere in between. AKP is likely to remain the largest  party in Turkey for the foreseeable future based on current condition, maintaining around 35% or more of the vote. But the opposition has an opening to strengthen its position at the upcoming local elections and to present itself as a viable alternative in future general elections. If, as Jesse Colombo argues, Turkey is on the verge of an economic crisis, AKP support could erode more significantly. However, it is less clear which party AKP loyalists would support instead. In the long run, the opposition shares an interest in constitutional reform designed to strengthen the rule of law, establish meaningful institutional checks on centralized power, and protect individual and minority rights. All of this would help to prevent electoral authoritarianism should AKP, or any other party, regain the sort of support that it had in the 2007-11 era. That should be the overriding goal for 2015 and beyond.