There is a lot of excitement in the build up to Nigeria’s 2015 presidential election and justifiably so for a number of reasons; possibly for the first time since 1993, and certainly since 1999, the opposition parties have managed to form an alliance with some similitude of a national structure reaching each geopolitical zone of the country. Therefore, there is no shortage of commentary in the blogosphere, social media and traditional media as to what the outcome of the election would be. There is also some analysis out there using assumptions about historical correlations between incumbency (gubernatorial and presidential) and election outcomes (See here and here). A lot of these suggest we would have a close race, but most expect the incumbent president to win re-election.

History teaches us that one of the best predictors of the future is a long look at the past. We (in this case Eronmonsele) therefore thought it would be a fun exercise to dive in a little deeper into the numbers that are publicly available to see what new insights we could contribute to this debate. Our analysis supports the popular notion that in the most likely scenario, the incumbent, President Goodluck Ebele Jonathan’s (GEJ) will win with about 56% of the vote with the opposition getting 43%. Here’s how we came to that conclusion:

Methodology and Assumptions

Our analysis looks at the 1999, 2003, and 2011 presidential elections and extrapolates what the 2015 presidential election result could be based on a few adjustments to account for the current political climate e.g. the merger of ACN, CPC, and ANPP as well as GEJ’s perceived decline in popularity, which recent NOI polls (here) suggest are not as bad as either of our Facebook timelines would infer. We assume General Mohammadu Buhari (GMB) will win the APC nomination, going by his popular appeal in the Northern region and recent endorsements.

We model three scenarios, a base scenario (Scenario 1), which seems to be the likeliest based on our judgment of today’s political climate, an optimistic scenario for GEJ (Scenario 2) and an optimistic scenario for GMB (Scenario 3).  In each scenario, we make assumptions regarding likely voter turnout using 2011 turnout numbers and percentage of 2011 votes retained or gained by the two contestants based on our (considered) judgment and NOI polls described previously. For each geo-political region and in each scenario, we use a triangular distribution to model a range of outcomes – minimum, likeliest and maximum. We conduct a monte-carlo simulation, running our model 20,000 times (the model picks a random set of assumptions for each run) to estimate the probabilities of various outcomes. For simplicity (supported by historical presidential elections), we mostly consider each geo-political zone as homogeneous blocks, isolating states with unique characteristics. For instance, Ondo state is separated from the Southwest since it has not had an AD, ACN, or APC governor since 2003.

Voter Participation and Retention

Scenario 1: Base Case with Adjusted Voter Participation vs. 2011 (Conservative Case for GEJ)

In the base scenario, we assume that voter participation would be similar to 2011’s election in most parts of the country with an increase or decrease of plus or minus 5% when compared to 2011.  However, in the South-East and South-South (excluding Edo), the 2011 election results demonstrated a strong firewall for GEJ, which was partly due to the historic nature of his candidacy, being the first minority candidate from the region with a good shot at winning.  Voter turnouts surged to unprecedented levels of 84% in Imo and 86% in Bayelsa, against the national average of 52%. We expect a moderation in the enthusiasm and voter participation coming down by 10% from a 67% regional average to about 57% which is still significantly higher than the national average (regressing to the mean). Our projected voter participation percentage variances are shown in table 1 below.

Table 1- Voter Participation Projections for Scenarios 1 and 3
Table 1: Voter Participation Projections for Scenarios 1 and 3

We estimate votes earned by each candidate as a percentage of votes retained from the previous election, and new votes acquired based on votes lost by the other candidate and third party candidates. We assume that while there will be a lot of new voters coming in and old voter churn, the average voter numbers will remain similar, as exits are cancelled out by ingresses and the triangular distribution addresses the uncertainty.

We assume in this scenario, that GEJ takes home 92.5% of the votes he earned in 2011 in most PDP controlled states taking account of the influence of incumbent governors. These include Ondo & Ekiti, the Northeast excluding Adamawa (since this state recently had an APC governor, we are assuming the party is more organized here than its predecessor parties were in 2011), Northwest, North Central excluding Nasarawa, Kwara, and Niger (GMB had 64% in 2011 compared to his vote average in the North Central of 31%). We additionally assume he retains the aforementioned percentage in Edo and Niger states, and 85% in the Southwest excluding Ondo & Ekiti, in Adamawa, Nassarawa and Kwara, given the change of the political climate in those states and the likelihood of more APC resources on the ground. We expect GMB to retain 95% of the votes he earned in 2011 and 87.5% (+/- 5%) of the votes earned by ANPP and ACN as a result of the merger, despite the decamping of both Gov. Shekarau and Mallam Ribadu.  For the South-East and the South South we used the NOI monthly tracking polls to project that GEJ would likely win 93.8% of the total votes in the Southeast and 86.9% of the total votes in the South South.

Table 2 below shows a summary of the simulation’s outputs.

Table 2 - Scenario 1 Results
Table 2: Scenario 1 Results

Scenario 2: Base Case with Stable Voter Participation (Optimistic Case for GEJ)

In this scenario, we used all the assumptions in scenario 1.  However, instead of projecting a decrease in turnout in the South-East and South-South, we assume it stays the same as it was in 2011.

Table 3: Scenario 2 Results
Table 3: Scenario 2 Results

Table 3 shows a summary of the election results for this scenario.

Scenario 3: Optimistic Case for GMB

In this scenario, we make voter participation identical to the base case’s (scenario 1) with a 10% (+/- 5%) decline occurring in the South-East and South-South while the rest of the country is largely unchanged (0% +/- 5% change). We assume that with aggressive campaigning the APC is able to reclaim 30% of Jonathan’s voters in the APC states of the Southwest (GEJ retains 70%) and 12.5% in Ekiti and Ondo (GEJ retains 87.5%). We assume that GEJ retains 92.5% of his 2011 votes in the Northwest, Niger state and Northeast excluding Taraba and Adamawa (both states are regional outliers because GEJ recorded 61% and 56% in both of them respectively in 2011, despite his vote average in the Northeast of 31%). We expect GEJ to retain 85% of his 2011 votes in Taraba, Adamawa, Edo and North Central excluding Niger state.

Results

Scenario 1 Results: Base Case with Adjusted Voter Participation vs. 2011 (Conservative Case for GEJ)

 

In Scenario 1, GEJ wins by a handsome majority of 56% to GMB’s 43% (as shown in table 2 below), also winning in 24 states and obtaining the minimum requirement of 25% of the vote in 33 states as shown in Figure 2.

 

 

Scenario 2 Results: Base Case with Stable Voter Participation (Optimistic Case for GEJ)

In this scenario, GEJ wins by an even clearer majority of 57% to GMB’s 43% (as shown in table 2 below), also winning in 24 states and obtaining the minimum requirement of 25% of the vote in 33 states as shown in Figure 3.

From this, we can deduce that for every 10% decline in voter participation in the South-East and the South-South, the president could lose a percentage point of the popular vote.  Hence, it would be important for the president to shore up turnout in these regions in case the election turns into a nail biter.

Scenario 3: Optimistic Case for GMB

In the third scenario, we obtain a statistical tie, with GEJ and GMB each obtaining 49% of the vote.  Additionally GMB also has a 47% chance of winning the election in this scenario. Also, both GMB and GEJ obtain the minimum requirement of 25% of the vote in more than 24 states with an almost certain probability.

Table 4 Scenario 3 Results
Table 4: Scenario 3 Results

Conclusions

Our analysis demonstrates that if the election were held today GEJ would be the clear favorite to win with between 52% and 56%. For GMB have a decent shot at winning, the APC must do the following:

  • Get as many votes in the North-West and North-East as GMB did in 2011 with high voter participation
  • Receive a high percent of the votes Mallam Ribadu and Governor Shekarau got in 2011
  • Campaign aggressively in the South-West and the North-Central with the ideal of making the presidential race in both regions a statistical tie, or a win for GMB
  • Campaign hard in the South-East and South-South with the goal of winning between 25% and 30% in both regions
  • Deploy boots on the ground across the country, especially in the South-South and South-East to shrink the president’s firewall in both regions

While we had great fun playing with the numbers, we are aware of that our analysis is more directional than predictive. With better polling figures, we could potentially hone our assumptions and use better statistical tools such as Bayesian estimation to arrive at numbers that would seem more precise, but as with all models of the future, this analysis is merely prosthesis for the mind and contrary to the opinions of our ragging fans neither of us can see the future.

 

[1] MBA Candidate, Darden Graduate School of Business, University of Virginia, VA, USA; Email: nigerianelections@gmail.com (Corresponding Author)

[2] PhD Candidate, Pardee RAND Graduate School, RAND Corporation, 1776 Main Street, Santa Monica, CA, 90401, USA