Predicting the 2019 presidential election

As the 2019 election draws close, we should expect an array of election entrepreneurs and predictors to become ubiquitous. A few days ago, the influential London-based Economist magazine was reported by several media organizations to have predicted a Buhari win in the February 2019 election. For instance, The Nation of December 4 2018 quoted The Economist’s, ‘World in 2019’ report to have said: “The president, Muhammadu Buhari, will win re-election in February, as a new opposition coalition may collapse before the vote.”

The paper however noted that the magazine’s foreign correspondent, Aman Rizvi, disagreed with the official position of his magazine and tipped the PDP to win. The paper quoted Mr Rizvi as saying:
“Mr. Buhari’s approval ratings have languished below 50 % for most of 2018. Ominously, he has been hit by a wave of defections to the PDP….So, who will win? Many Nigerians do not care. The back-and-forth floor-crossing has convinced them that the same people will be in charge either way. The PDP holds a slight edge, if only because expectations for Mr. Buhari were higher and his failure to meet them was more recent.”

I made several unsuccessful attempts to access or buy the full report online. I was however able to access a published version of the report concerning Nigeria and entitled ‘The World in 2019: Nigerian Apathy’ (https://www.economist.com/the-world-in/2018/12/04/nigerian-apathy). The article, which was published on December 4, 2018, has no by-line but corresponded to the position and quotes attributed to the magazine’s foreign correspondent, Aman Rizvi, by the Nation newspaper. My calls to the Economist’s office in London did not yield fruits and my email to Jonathan Rosenthal, whom I was told helped to prepare the Africa section of the report, was not replied before I sent out my piece to Daily Trust for publication (around 7.30PM).

So what does this prediction – or the business of predicting the outcome of elections in general – mean? It is possible to identify at least five categories of election predictors (in no particular order):
The first category is the professional polling companies like Gallup Poll or Rasmussen (or NOI poll in Nigeria) which will actually ask a sample of voters about their preferences and then extrapolate from the sample to predict the outcome of a particular election. While several professional polling companies try to be rigorous in their methodologies and can often be right in their predictions, even the best among them can also get it wrong sometimes – even when they conduct exit polls (polls taken immediately after voters have exited the polling stations or booths –as we have in Nigeria). A major reason why polls by professional polling companies fail is the so-called ‘Bradley Effect’. This is the observed discrepancy between voters’ expressed preferences and election outcomes in some United States elections, especially where a White candidate is running against a Black candidate and voters mask their true preferences either out of social desirability or for fear of being called racist. The ‘Bradley Effect’ is named after then Los Angeles Mayor Tom Bradley, an African American, who lost the 1982 California governorship race despite being ahead in all voter polls going into the election. We also saw the ‘Bradley Effect’ play out in Donald Trump’s defeat of Bill Hilary Clinton in the last US presidential race despite all the polls favouring Hilary Clinton to win. In essence, even professional polling companies, with rigorous methodologies, can get it wrong.

The second category of predictors of electoral outcomes, especially in Africa, are diviners of all sorts (who claim to have crystal balls or have a way of foretelling the future) and our ‘Men of God’ (Pastors and Imams) who claim divine revelations of tomorrow. Since several people in this category of predictors have conflicting predictions of the outcome of the same election (and some contradict themselves once cash or enticing projects are dangled before them), it is difficult for many people to take them serious – except as a form of entertainment to the political process.
The third category of predictors of election outcome are major Western institutions and think-tanks, which sometimes ‘accidentally’ predict outcome of elections in the course of mapping out likely trajectory of events in several countries around the world. For instance the Economist has ‘World in’ reports published every year to highlight likely drivers of events around the world during the year in question while the Brookings Institution has Foresight Africa which it publishes in January every year to highlight likely drivers of events in the continent during the year. These reports are usually written by people regarded as experts on those countries. A particular report may take several months to write and then reviewed and corrected numerous times. Each election ‘prediction’ will depend on several factors including when the report was submitted for publication (which could be a few months before it was published). For instance, in the run-up to the 2015 presidential election in Nigeria, I was asked by the Brookings Institution to do a report for its Foresight Africa on the challenges and trajectories of the 2015 presidential election. I began writing the report around September 2014, and we had at least five revisions of the draft.

At the time the final draft was completed the APC had not even nominated its presidential candidate and there was apprehension that the alliance was going to unravel over a choice of presidential candidate. At that time, the ruling PDP was ascendant, with the Transformation Ambassadors of Nigeria (TAN) regaling us with stories of millions of Nigerians across the country signing ‘verifiable signatures’ and begging Jonathan to contest. When the APC eventually chose Buhari and Osinbajo as its presidential candidate and running mate respectively, I quickly updated the report which I had already completed and then submitted for publication in November 2014. It was published in January 2015 (https://www.brookings.edu/wp-content/uploads/2016/07/nigeria-elections-adibe-2.pdf).
What could pass for predicting the outcome of the 2015 election was the following passage in the concluding part of the report: “Allegations of corruption against top PDP officials will be powerful ammunition in the hands of the APC, especially with the choice of Buhari, widely seen as not corrupt, as the party’s presidential candidate. The party is, however, a fragile one that seems united only in its quest to wrest the presidency from Jonathan or to have power “returned” to the north. Though the election is expected to be very competitive, the odds still favour President Jonathan.” Understandably PDP spin doctors took that passage out of context and several newspapers reported: “Brookings Institution Tips Jonathan to Win”. This is the context in which the prediction by the Economist and similar predictions by thank-tanks and financial analysts could be understood. If my report for Brookings was submitted in January 2015, my prediction of the outcome of the election would be different. In essence predictions are influenced by the context and opinion of the writer.

A fourth category of election predictors are media analysts and public intellectuals. These will often base their analyses on a number of observable variables and will usually predicate their predictions on a caveat of ‘all things being equal’ – or ‘if this trend continues’. But the truth is that things are rarely equal as the dynamics and drivers of electoral outcomes can change very quickly with intervening variables (‘fogs of war’) that can become game changers.
A fifth category of election predictors are the partisans. These are party members and ardent supporters of particular candidates. Partisans will often look at the rosy side of things and will selectively focus on stories that will validate their choices. They work from answer to the question, so to say.

Do election predictions really matter? Yes, they do. For parties that are predicted to win, (especially by the major international institutions such as The Economist, Brookings Institution and the leading political/financial risks analysts), such predictions could provide a boost and create a bandwagon effect by unwittingly nudging voters to a particular political preference. In essence, election predictions could create a self-fulfilling prophecy – which is why some candidates in Nigerian elections procure favourable prophecies from ‘Men of God’ or organize fake opinion polls.

For now both the PDP and the APC appear to have serious internal issues that are impeding the effective takeoff of their campaigns: while there are fractionalizations and dissensions in several APC states (Zamfara, Imo, Ogun, Ondo etc) following the party’s contentious primaries, most of the PDP governors (South-east, South-south and even Northern states like Benue and Gombe) appear lukewarm towards the Atiku/Obi ticket. It is of course possible that these are teething problems that most campaigns face. However while it is still too early in the day for any reliable prediction to be made, many will be watching out for which of the two main parties will quickly resolve its internal problems and energize its base and which will be consumed by those problems.
______________
Email: [email protected]
Twitter: @JideoforAdibe