People walk in front of a giant poster of Turkish President Recet Tayyip Erdogan on April 6, 2017 in Istanbul, 10 days ahead of the referendum on whether to change the current parliamentary system into an executive presidency. / AFP PHOTO / BULENT KILIC (Photo credit should read BULENT KILIC/AFP/Getty Images)
People walk in front of a giant poster of Turkey's president Recep Tayyip Erdogan. Mr Erdogan is hoping to cement his powers at a referendum on a new constitution on Sunday © AFP

Turkey will vote on April 16 in a referendum on key constitutional amendments that aim primarily to replace the country’s parliamentary system with a strong executive presidency.

Even though the latest polls suggest strong momentum for a Yes outcome, with around 53-54 per cent of the vote, the almost evenly-distributed poll results suggest that tossing a coin might be as good an indicator.

That is a deviation from five out of six of the country’s historical referendums, which yielded clear yes/no results in 1961, 1982, 1988, 2007 and 2010. The only referendum with a really close-call took place in 1987, when 50.1 per cent of Turks voted in favour of the removal of political bans following the military-backed 1982 constitution.

It is, however, not an easy job for pollsters to make a healthy analysis under Turkey’s state-of-emergency climate, in which thousands of people, including quite a few well-respected journalists, have been arrested or lost jobs since a failed coup attempt in July 2016.

Response rates (ie the percentage of people who respond to the survey) are unusually low, as people are unwilling to reveal their true vote to “strangers” who give them a call (and who know respondents’ names and contact details) or meet them on the street for a face-to-face interview. In some instances, pollsters need to invite 2.5-3 times more people than usual for an interview to reach the targeted number of respondents.

The other difficulty arises when making a qualitative assessment to survey results. The proportion of undecided voters is believed to be around 10-15 per cent, a level that could be a real game changer. A pro-rata allocation of the undecided may not be the best approach due to “soft-issues”, as several people we met suggest that No-sayers tend to hide their views more often than Yes-sayers due to worries over the state of emergency. If this view holds true, even a dominant 55-60 per cent No outcome should not come as a surprise.

Putting the guessing-game aside, what matters to investors is the impact of the referendum on the investment climate, economy and markets. From this perspective, the focus should rather be on possible post-referendum scenarios and trying to make a dynamic analysis of possibilities.

A Yes vote will remove most of the uncertainties, at least in the short term. However, this outcome could just be a step towards further debates, as the full-scale presidential system will kick in only along with elections scheduled for 2019 — or maybe earlier.

Depending on timing and the margin of a Yes vote, a snap election may even be preferred by the market, as it eliminates two more years of uncertainty until the 2019 elections. Then there are the discussions over subjects such as freedom of speech, EU relations, reinstatement of the death penalty, increasing fiscal spending and the overall shape of the economy.

Despite expecting a short-term improvement in investment sentiment on a Yes vote, a permanent improvement will come only when and if political tensions fade away considerably.

A No vote will also yield a mixed picture, with further uncertainties and a possible early election on the one hand, and maintenance of a 94-year-old proven and working parliamentary system on the other.

However, a No vote probably has more potential to eventually lead to other unexpected scenarios, such as a speculated break-up of President Recep Tayyip Erdogan’s ruling Justice and Development (AKP) party or rise of a rival conservative political party which may significantly weaken support for the AKP.

Those scenarios may even include the possibility of an even stronger parliamentary representation for the AKP, if the Kurdish HDP and nationalist MHP fall below the 10 per cent threshold for parliamentary representation in a snap election. In the case of a two-party parliament, the AKP may be able to make even larger constitutional changes, as reaching a decisive two-thirds majority could be possible.

In conclusion, there is much doubt over the reliability of the poll results. It could be that even a surprising Brexit or Trump-like outcome is possible, but it does not make sense to draw strong conclusions in response to a given outcome.

Regardless of the outcome, the base case for the market is to see “more of the same”, ie to see occasional volatility because of constant political noise on different topics, many of which have been seen extensively since mid-2013.

With that, East Capital continues to focus on companies that are less affected by political uncertainties due to their business models, that have strong export capabilities and that typically pay high dividend yields, due to strong balance sheets.

Emre Akcakmak is a portfolio adviser at East Capital, an emerging and frontier markets asset manager

Get alerts on Turkish referendum when a new story is published

Copyright The Financial Times Limited 2019. All rights reserved.
Reuse this content (opens in new window)

Follow the topics in this article