Markets wrong in their elections optimism

I did my first election forecasting for Turkey during the 2007 general election, when I was working as a market economist at Citi’s Turkey headquarters.

There wasn’t actually much forecasting. The basic model I built with my boss could calculate the number of seats each party would win, based on the percentage of votes they were predicted to get. We simply took the average of all the available polls to feed the percentage of votes into the model.

In the end, we failed miserably, not because our model was wrong, but because two right-wing parties did not merge, each ending up with a couple of percentage points rather than the 10-15 percent that the polls were predicting for the new party. Prime Minister Recep Tayyip Erdoğan ended up advising Citi’s bigwigs, who were visiting him after the elections, to fire us - as it was obvious, based on our prediction of a coalition government, that we were not good at our jobs.

Similar election simulators are now on the web. Taking the average of the latest polls yields 42-44 percent for the ruling Justice and Development Party (AKP), 25-27 percent for the main opposition Republican People’s Party (CHP), 15-17 percent for the Nationalist Movement Party (MHP), and 9-11 percent for the Peoples’ Democratic Party (HDP).

You don’t need a model to realize that the Erdoğan’s fate is dependent on whether the HDP passes the 10 percent threshold. But feeding in the numbers shows that the AKP could end up just below or above the minimum number of seats (276) necessary to form a single-party government if the HDP does enter parliament. Interestingly, the AKP could still be below the 330 seats needed to take Erdoğan’s constitution to a referendum even if the HDP remains under 10 percent.

Will the HDP make it? It all depends on whether they can get enough votes outside their core constituency, especially from those who would vote for them so that they can prevent Erdoğan from becoming the supreme ruler of the realm. The problem is that their leader Selahattin Demirtaş’s promise to block Erdoğan is not credible. After the elections, his best interest would be to receive concessions in exchange for support for the new constitution.

Economists Finn Kydland and Edward Prescott, who won the Nobel Prize for this idea of “time inconsistency,” showed that rational people would not be deceived by such promises. However, talking to political analysts and polling companies, I have come to realize that many people will nonetheless vote for the HDP: These voters’ time horizon is very short, and they are willing to take a chance on the HDP because their priority is to block Erdoğan. The HDP should therefore be able to pass the threshold - unless cats get into power distribution units and cause blackouts again.

But I also think, again based on my conversations with the experts, that the AKP will probably find a way to get support for the constitution if it manages to stay in power - from the MHP if not HDP. Some commentators have pointed out that Prime Minister Ahmet Davutoğlu and Erdoğan’s criticisms of the MHP have been rather toned down, especially compared to their vicious attacks against the CHP and the HDP.




In that sense, the markets’ pricing of the “golden scenario” of a single-party AKP government, but without enough seats to take the constitution to a referendum, (which partly accounted for last week’s strong performance of Turkish assets), may be premature.