Wednesday, April 30, 2014

What happened in Ankara 2009 Elections?

Note: To read an extended summary of the two analyses of the local elections (this article and the next one) in Turkish please see here.

Several people have been analyzing different aspects of the preliminary results from Turkish local elections collected through the Yerel Seçim 2014 Sandık Takip Sistemi (the ballot box tracking system), an initiative of CHP, the main opposition party in the parliament.

Erik Meyersson, an economist at Stockholm School of Economics, posted a few articles analyzing this data. His first article reports preliminary evidence from Ankara and Istanbul Mayor elections indicating a positive relationship between invalid vote share and AKP-CHP vote margin at the ballot box level. The magnitude of the relationship was smaller when district- or voting station-specific factors were netted out, but it remained statistically significant. While definitely not conclusive, these results might be a sign of electoral manipulation whereby a disproportionately greater [lower] share of votes cast for CHP [AKP] were invalidated (either blatantly or through indirect methods).


One may wonder if the relationship between AKP-CHP vote share difference and the invalid ballot share was also there back in 2009. If the answer is no, then what we have in 2014 is certainly more likely to reflect voting manipulation, because it is harder to come up with a structural explanation of the documented relationship that would hold today but not 5 years ago.However, AKP has been in power for more than 10 years now, and there is no reason to doubt that incentives and capabilities to manipulate back in 2009 were dramatically lower than it was in the last elections. Then, one should expect a similarly significant relationship in the previous local elections. To see what the answer is, I repeat Erik's analysis of the 2014 elections using 2009 official results for Ankara.

First, I will focus on election results for district municipalities --as opposed to metropolitan area mayor. The figure below plots the unconditional relationship between AKP-CHP vote share difference (akpchpdiff) and the share of invalid votes (invshr) at the ballot box level. The results indicate a positive and significant association, just like in the last elections !



The slope coefficient (4.46) is somewhat smaller compared to what Erik has found (5.86). To see whether the relationship remains qualitatively intact after we account for district level unobserved factors (like socio-demographic characteristics) that might influence the vote shares, I add dummy variables for each district. Hence, the figure below simply shows the relationship between the residual variations in akpchpdiff and invshr respectively after netting out the association between cross-district variations in these variables.


Results are equally significant albeit the size of the estimated coefficient (2.46) is about half the unconditional coefficient. This indicates that indeed some district specific factors (including but not limited to incentives to steal votes) were partly (but not fully) responsible for AKP's electoral advantage vis a vis CHP as well as the share of invalid ballots. I should note that the magnitude of my estimate is strikingly close to the estimate (2.72) Erik reported in his corresponding specification (see here).

Finally, we can go one step further and control for voting station fixed effects. To be fair, that the relationship may completely disappear at this stage does not say much about the relevance of voting manipulation. On average there were roughly 14 ballot boxes in a given voting station, as opposed to about 415 ballot boxes per district. Thus, there may not be enough variation in our variables of interest within each voting station to obtain a sufficiently precise estimate even if there is a systematic relationship between vote shares and invalid ballot shares. Or it may very well be the case that the feasibility of and incentives for manipulation are quite unlikely to differ much across ballot boxes in a given voting station.

The figure below plots the same relationship, this time conditional on voting station fixed effects.


Not surprisingly the relationship is much lower in magnitude. However, it is positive and very precisely estimated. Once again, the coefficient (0.39) is very similar to what Erik found in his analysis (0.40). At this point, I should add that the 2014 data Erik has been using belongs to metropolitan mayor elections while the one I have used in this analysis come from district municipality elections. While, voting behavior in the two types of elections should be highly correlated, it is still interesting that the factors which are driving these results (whatever they are) seem to have persisted throughout last 5 years. To what extent this remarkable persistence between 2009 and 2014 elections can be attributed to persistence in potential electoral fraud is a big question that begs an answer. Perhaps more on that later on.

Troubling Regularity of the Irregularity: Evidence from 2014 Local Elections in Ankara in the light of 2009 Outcomes

Note: To read an extended summary of the two analyses of the local elections (this article and the previous one) in Turkish please see here.


As I am writing this, Turkey is still waiting for the Supreme Election Commitee (YSK) to announce the official results of March 30, 2014 local elections. Vote recount requests and other complaints are abound. There are serious allegations of election rigging and some supporting evidence that raises further suspicions regarding the fairness of the whole process. Several people have been analyzing different aspects of the preliminary election results collected through the Yerel Seçim 2014 Sandık Takip Sistemi (the ballot box tracking system), an initiative of CHP, the main opposition party in the parliament.

Erik Meyersson, an economist at Stockholm School of Economics, posted a few articles analyzing this data. His first article reports preliminary evidence from Ankara and Istanbul Mayor elections indicating a positive relationship between invalid vote share and AKP-CHP vote margin at the ballot box level. The magnitude of the relationship was smaller when district- or voting station-specific factors were netted out, but it remained statistically significant. While definitely not conclusive, these results might be a sign of electoral manipulation whereby a disproportionately greater [lower] share of votes cast for CHP [AKP] were invalidated (either blatantly or through indirect methods).


In a follow-up article Erik showed that these findings somewhat generalize to some other provinces like Adana, Izmir and Manisa. What is more interesting was that the relationship was significantly weaker in provinces where AKP tends to dominate the race by relatively larger margins and political competition was limited. This finding is crucial because it is consistent with a deliberate and strategic behavior. Allocating resources and concentrating manipulative efforts to provinces where they are more likely to matter make more sense (for manipulators), as the argument goes.


Finally, Erik's analysis shows that in pro-CHP districts where support (vote share) for CHP was above CHP's Ankara-median, the same relationship was strong and highly significant while in other districts (pro-AKP sample) it was extremely weak and statistically insignificant, further lending credibility to a voting manipulation story. Yet, this analysis relies solely on data from 2014 local elections. The idea that pro-CHP districts are more likely to be targeted by potential manipulation efforts is reasonable. However, he uses district level election outcomes from 2014 to identify the pro-CHP and pro-AKP districts. Selecting the regression sample based on how parties performed in the same election may be somewhat problematic, because then the sample selection process itself might be directly influenced by (instead of affecting) the actual intensity of ballot invalidation and hence the invalid vote share. This might in theory exacerbate a potential endogeneity problem.


More importantly, current electoral performance in a district may not be a good proxy for incentives to manipulate. For example, suppose that, in district A, the invalid vote shares were high due to reasons other than manipulation. Imagine that these reasons correlate positively with CHP's performance. Say, for example, that monitoring by CHP supporters was stronger and resulted in invalidation of incorrectly cast ballots that would otherwise favor AKP. As a result, this district was more likely to belong to the sample of pro-CHP districts. However, in this scenario, the fact that CHP did well in district A has nothing to do with higher incentives to manipulate.


In this article, I'll try to complement Erik's analysis of invalid vote shares and the outcomes of the 2014 elections by using the official results of the previous local elections in 2009 (for Metropolitan Mayor) as a proxy for the incentives for electoral manipulation. Data on 2009 elections for metropolitan mayor is retrieved from the online database of the Turkish Statistical Institute (TUIK). This approach (using information from previous elections) seems to be better suited for the question of interest. The underlying reason is that election performance in 2009 serves as an objective (and ex-ante observed) indicator of districts where marginal return on manipulation in 2014 would be higher. The hypothesis goes as follows: If there were attempts to selectively invalidate CHP votes and/or count incorrectly cast AKP votes as valid, then one would expect such attempts to be (1) stronger in districts where the main opposition parties (CHP or MHP) received the highest share of votes in 2009 and (2) weaker in districts where AKP was the first party. Moreover, among the districts where AKP was the first party, we would expect (3) more intense manipulation when the AKP's win margin was relatively low and (4) much less manipulation when it was high.


In this article, I focus on Ankara, where the suspicions about the election results have been the strongest. To investigate whether these four predictions receive support in the data, I regress the vote shares of AKP, CHP and MHP on the share of invalid votes (invshr) using five different samples. Unit of observation is a ballot box. The following figure plots the regression coefficients along with the 95% confidence intervals.The plot on the left-hand side reports the unconditional relationships, and the one on the right-hand side shows the results conditional on district-level fixed effects (click on figure to see a larger version).





In the first row, I regress each party's vote share on invshr using the entire sample. In the last elections, all of the 25 districts in Ankara voted for Metropolitan mayorship. In 2009, however, 7 out of these 25 districts were not allowed to vote in the Metropolitan area elections. Therefore, “All Districts” sample includes the 18 districts which voted both in 2009 and 2014.

The results on the first sample are in line with Erik's findings. In ballot boxes where invalid votes were more prevalent, AKP had a significant advantage while CHP's vote share was lower. There is a positive relationship for MHP, but it is much weaker and only significant at the 10 percent level. Estimated coefficients are smaller in magnitude, but still significant, when district fixed effects are accounted for.


The second row is striking because it suggests that in the 5 districts where either CHP or MHP was the leading party in 2009, the relationships are somewhat stronger and substantially more precise (except for MHP).


The results in the third row are consistent with the idea that in the 13 districts, where AKP obtained highest share of votes in the previous local elections, the relationships between invshr and party vote shares were on average smaller in magnitude and less significant. However, these results appear to disguise a systematic difference between districts that saw more intense competition and the ones where AKP had a landslide lead.


To see why, look at the fourth row. It shows the regression results for districts in which AKP had the lead, but only by a win margin that is less than 10%. These correspond to Etimesgut, Kalecik and Mamak. The last row, on the other hand, focuses on districts where win margin in 2009 was above 20% (Sincan, Altindag, Pursaklar and among others).


Of course, this analysis is just a preliminary glimpse at the data and way more work lie ahead. Nonetheless, the complementary evidence I provide here strengthens the case that, at least for Ankara, where the race between AKP and CHP was extremely close, a close and independent examination may change (or shall I say might have changed) the winner. Yet, as Erik has also pointed out in his articles, these seemingly systematic anomalies in the election results do not prove that voting manipulation of the kind we have been investigating actually took place. On the other hand, as more systematic evidence is uncovered (here is one recent example), it becomes increasingly harder to come up with alternative stories that would explain away these irregularities.


Mansur Yavaş, CHP's candidate for Ankara metropolitan area mayor, has filed an appeal to the Constitutional Court after the Higher Election Commitee (YSK) turned down his complaints. Let's not be naive here. Most probably, the disputes regarding the local elections in Ankara and some other provinces will not lead to a significant revision of originally reported outcomes. However, this does not mean that we should abandon the quest for fair elections and more transparency in politics. On the contrary, we shall put more pressure on the government. I see that Turkey is slowly moving toward that direction. It is very encouraging to witness all the data collection and monitoring efforts of citizenship initiatives (like Ankara'nin Oylari in Ankara and Oy ve Ötesi in Istanbul) and empirical analyses by scientists and individual volunteers. Groups on Twitter like @sandikbuglari are working to publicize new evidence and findings, trying to raise more awareness. I consider all these efforts as part of the revival (or perhaps the birth) of Turkish democratic civil society, a process that gained significant impetus with the Gezi Park resistance.