Presidential race prediction: A math model (II)

In 2015, this column featured a math model predicting that President Muhammadu Buhari would win the election by 59% of the votes. He won by 53.96%. My prediction was off by 5.04%. Today I’m turning in another prediction which is not very different from the one of four years ago. The numbers upon which the […]

In 2015, this column featured a math model predicting that President Muhammadu Buhari would win the election by 59% of the votes. He won by 53.96%. My prediction was off by 5.04%.

Today I’m turning in another prediction which is not very different from the one of four years ago.

The numbers upon which the math model is based are logical enough, but the reasons or premises behind those are anything but logical. Because, you see, voting is like the decision you take when buying something. Our shopping is rarely driven by logic. How many times did you go to the market to buy a few things but returned home with a basketful of things you didn’t intend to buy? I will hazard a guess that it is all the time.

Same way with elections. Therefore, like marketers, politicians have learned to appeal to our emotions rather than our logical reasoning. And in Nigeria we have strong emotional pillars that elicit different outcomes depending on how or who is employing them. One them is religion.

Keeping this in mind, I submit that very little has changed since eight years ago. Factors that swayed Nigerians in 2011, to a reasonable extent, swayed them in 2015 and will nudge their hands again in 2019.

In other words, Buhari’s payment to ex Biafran service, road construction in the southeast, committed efforts on the second Niger Bridge and completion of railway to Warri will translate to only a few votes.

So, I’m going to create a math model to explains this – even though a simple one.

Do not mind its simplicity however.  “In principle, it matters not whether the math methods employed to carry out math proof are simple,” wrote William J. Adams, author of ‘Slippery Math in Public Affairs’, “or the ultimate in technical sophistication.”

In doing this, I do not guarantee reality, only validity.  That is, a promise of (a perfect) accuracy of the prediction of voter behaviour is neither explicitly nor tacitly claimed.  Based on my postulates (a crystallization of my assumptions) however, my theorems (conclusions) are valid.

Another person could bring forth a different set of postulates (the hypothesis) and reach conclusions that are different from mine – keeping in mind the RIRO (if realistic input, realistic output) and the NINO (nonsense in, nonsense out).

Now that I’ve turned in my excuse early, let’s move on to the model.  First we’ll start with the postulates; in logic or rhetoric, these would be called premises.  The theorems or conclusions followed the postulates. The theorems are drawn based on the postulates.

The Postulates

Postulate 1: Factors which influenced election victory have not changed since 2015.  Such factors are: voting based on ethnicity and religion as well as Southeast and Niger Delta’s inclination to vote PDP.

Postulate 2: Buhari will win the states he won in 2015 and PDP will win the states it won in 2015 except  Benue and Nasarawa states. In 2015 winning Benue by Buhari was an anomaly. Therefore, in 2019, Benue will, as statisticians say, regress to the mean. It will go back to its normal position – to the PDP. The opposite will happen in Nasarawa State. Buhari has never won that state. But will win it this time, because all the PDP big shots have joined the APC.

Postulate 3: Undecided voters (defined as those who didn’t vote in 2015) will vote in 2019. This model uses the voter register of 2015.

Postulate 4: The undecided voters will vote according to the states the PDP and the opposition parties won in 2015, except in Nasarawa and Benue.

Postulate 5: All eligible voters will collect their permanent voter’s card (PVC) before the deadline and will vote.

Postulate 6: Atiku will get 25% of votes from Northwest (because of Kwankwaso, Tambuwal, etc.) and from Niger State. Buhari will get 25% of votes from Southeast (because Gov Obiano campaigns for Buhari in Anambra and Ebonyi, Okorocha’s influence in Imo and former Gov Kalu’s popularity as APC senatorial candidate of Abia North).

Postulate P: Other factors that may play a role in the result of the election are not being considered.

Theorems

Theorem 1: Atiku will win all the states in the Southeast (7,665,859) and South-south (11,133,620) and Northcentral (6,610,827) except Niger State (2,014,317), Nasarawa State and Kwara State (1,142,267), giving him a total of 25,410,306.

Theorem 2: Gen. Buhari will win all the states of the Northwest (17,620,436), all the states of the Southwest (13,731,090), Niger (2,014,317), Nasarawa and Kwara (1,142,267) states from Northcentral and all the states of the Northeast (6,208,197) except Taraba (1,340,652) and Adamawa (1,559,012) where he’ll share the votes with Atiku; giving Buhari a total of 40, 716,307 votes.

Theorem 3: Gen Buhari will win the election with popular votes from 19 states or the equivalent of 40 million votes in percentage which is 59% of the votes.

Explanations

This looks neat enough and makes mathematical sense.  If you grant my postulates P1 through P6 and PP, you must also grant my theorems T1 through T3 as valid consequences of them.

T1 follows from P1, P3, P4 and P5.  T2 follows from P2, P3, P4, P5 and P6. T3 follows from T1 and T2.

Postulate P states that other factors are not being considered.  But what other factors?

For a critical consumption of this model, let’s plainly state the factors.

Boko Haram is not going to have as much sway on the voting patterns of the Northeastern states with a sizeable population of Christians.  Northern Christians who voted for Jonathan in 2015 from Northeast are likely to vote Atiku in 2019 – regardless of the Boko Haram phenomenon.  Thus, Boko Haram is not being considered for some of the Northeastern states.

Let me give one example.  Do you know that there were people in Chibok who voted Jonathan in 2015? This was against reasonable analysis then that Chibok wouldn’t vote PDP because of how Jonathan’s wife treated the Chibok Girls’ parents. Also, the fact that attacks by Boko Haram has increased in last weeks wouldn’t change the minds of those who have decided to vote Buhari.

Also, the fact that President Buhari consistently continues to appoint a northern Christian as Secretary to the Government of the Federation will not translate to any votes in his favour. So this factor too is not being considered.

Those who think the religious tool PDP used in 2011 and 2015 can no longer sway should think again. Of course Atiku is Muslim too. But many see him as more liberal than Buhari.

Other factors abound – but they’re also not being considered.  If you disagree with this model – I know many will – let me know what you think.