Our forecast model is finally updated with the results of our August poll and the result is a fairly sizable swing in the chances of both candidates as well as the probability of a run-off. 

The update:

  • After the update, both Nana Addo and John Mahama saw small but meaningful dips in their chances of winning a majority of votes on December 7th. 

  • Nana Addo’s chances dropped from 64.8% to 60% while Mahama’s chances slipped from 31.4% to 26.7%. Nana Addo still remains the slight favorite to win but the sudden spike in the chance of a run-off suggests the race is far from settled at this point. 

  • To be clear, a 13% chance of a run-off is still small. Our model expects the likeliest scenario to play out on Election Day is that of a Nana Addo win, followed by a scenario in which John Mahama wins and then lastly a scenario in which neither candidate clinches 50% plus one vote. 

  • So why the sudden jump from 3.8% to 13.3% for the chance of a run off? In one word: UNCERTAINTY. Our model calibrates the chances of each candidate winning by looking at how settled the race appears to be. 

  • The more confident the model feels about a particular candidate winning, the lower the chances of a run-off. Conversely, if the model is not so sure about any one candidate having an easy shot at victory, then it thinks the chances of a run-off will be higher. 

  • In this latest update, it was the 20% of voters who are undecided as to their choice for president that introduced a fair bit of uncertainty in our forecast. Add to that the 8.6% of voters who say they are voting for a 3rd party or independent candidate and our model sees an election that’s far from certain. 

  • That said, this is just the impact from one poll and we are still 3 months away from Election Day. The model expects the share of undecided voters to drop as we get closer and closer to Election Day. If our next poll shows a similar share (or higher) of undecided voters, then expect the chances of a run-off to be even higher. 

  • If instead we see more of the undecided voters breaking for Mahama, then the model might treat the race as becoming more competitive and we can reasonably expect a shift in the forecast positively towards both Mahama’s chances and that of a run-off. 

  • If instead we see more of the undecided voters breaking for Nana Addo, then the model will see the race as becoming a bit more settled and we can thus expect a drop in the chances of a run-off and a bounce for Nana Addo. 

  • Ultimately though, uncertainty rules the day and once again we are heading into an election where the swing voter will decide who wins. 
%d bloggers like this: