Tuesday, December 18, 2007

How Kibaki Might Win

You Missed This: Opinion Polls Are Raining On Kenyans

It certainly is true that we are suffering from opinion poll fatigue, but it would be remiss of us to avoid the simple analysis that is possible with these polls. Here is a perfectly logical scenario how Mwai Kibaki can honestly and truthfully win the election on December 27th.

According to the African Elections Database voter turnout for the presidential election in 1997 was 68% and in 2002 was 56% for an average of 62%. Let's assume that overall voter turnout in this year's presidential election will be about the same as the average -- 60-64%.

If you accept, as I expect you would if you know anything about Kenya, that ethnicity plays a larger-than-life role in Kenyan poilitics, then it would not be too big a leap to say that it is likely that the GEMA, Luo and Kamba turnout in this election will be higher than other ethnicities without a presidential candidate.

In my model I assume that turnout in disinterested districts will be 50% and in districts largely populated by one of these three groups turnout will be 75%. Here is how I have categorized the districts using voter data from the Electoral Commission of Kenya.

Province District Ethnicity Registered Voters Modeled Turnout
Nairobi Nairobi Other 1,275,021 50%
Coast Mombasa Other 370,286 50%
Coast Kwale Other 187,945 50%
Coast Kilifi Other 227,555 50%
Coast Malindi Other 127,777 50%
Coast Tana River Other 84,527 50%
Coast Lamu Other 47,539 50%
Coast Taita Other 132,690 50%
NEP Garissa Other 86,846 50%
NEP Ijara Other 16,874 50%
NEP Wajir Other 110,138 50%
NEP Mandera Other 101,808 50%
Eastern Moyale Other 33,478 50%
Eastern Marsabit Other 63,972 50%
Eastern Isiolo Other 57,573 50%
Eastern Meru Central GEMA 297,063 75%
Eastern Meru North GEMA 260,530 75%
Eastern Meru South GEMA 123,972 75%
Eastern Tharaka GEMA 47,261 75%
Eastern Embu GEMA 162,902 75%
Eastern Mbeere GEMA/Kamba 92,465 75%
Eastern Kitui Kamba 245,094 75%
Eastern Mwingi Kamba 145,155 75%
Eastern Machakos Kamba 488,828 75%
Eastern Makueni Kamba 356,400 75%
Central Nyandarua GEMA 264,811 75%
Central Nyeri GEMA 409,590 75%
Central Kirinyaga GEMA 279,441 75%
Central Murang'a GEMA 208,244 75%
Central Maragwa GEMA 224,409 75%
Central Kiambu GEMA 428,497 75%
Central Thika GEMA 371,323 75%
RVP Turkana Other 138,414 50%
RVP West Pokot Other 123,229 50%
RVP Samburu Other 68,816 50%
RVP Trans Nzoia Other 261,507 50%
RVP Uasin Gishu Other 319,615 50%
RVP Marakwet Other 68,654 50%
RVP Keiyo Other 81,819 50%
RVP Nandi North Other 141,535 50%
RVP Nandi South Other 131,204 50%
RVP Baringo Other 115,399 50%
RVP Koibatek Other 72,833 50%
RVP Laikipia GEMA 182,091 75%
RVP Nakuru GEMA 657,219 75%
RVP Narok Other 170,244 50%
RVP Transmara Other 73,310 50%
RVP Kajiado Other 201,812 50%
RVP Kericho Other 234,285 50%
RVP Bureti Other 114,149 50%
RVP Bomet Other 202,150 50%
Western Kakamega Other 289,349 50%
Western Butere/Mumias Other 221,751 50%
Western Lugari Other 112,509 50%
Western Vihiga Other 256,941 50%
Western Bungoma Other 357,493 50%
Western Mt. Elgon Other 57,578 50%
Western Busia Other 183,267 50%
Western Teso Other 85,859 50%
Nyanza Siaya Luo 220,546 75%
Nyanza Bondo Luo 118,383 75%
Nyanza Kisumu Luo 211,268 75%
Nyanza Nyando Luo 165,618 75%
Nyanza Rachuonyo Luo 140,107 75%
Nyanza Homa Bay Luo 135,140 75%
Nyanza Migori Luo 226,258 75%
Nyanza Suba Luo 74,845 75%
Nyanza Kuria Other 64,891 50%
Nyanza Kisii Central Other 241,236 50%
Nyanza Gucha Other 205,173 50%
Nyanza Kisii North Other 238,221 50%



14,294,732

Overall provincial and national turnout given these assumptions is:

Province Registered voters Turnout Votes Cast

A B A*B
RVP 3,358,285 56% 1,880,640
CEN 2,186,315 75% 1,639,736
EAS 2,374,693 73% 1,733,526
NYZ 2,041,686 66% 1,347,513
WES 1,564,747 50% 782,374
NAI 1,275,021 50% 637,511
CST 1,178,319 50% 589,160
NEP 315,666 50% 157,833
Total 14,294,732 61% 8,768,291


For voting patterns by province I used data from the Steadman Poll of December 8th, 2007 as published in the East African Standard. (Where data were missing e.g. for Kalonzo in many provinces I guesstimated a number.)

Province Kibaki Raila Kalonzo

C D E
RVP 36% 61% 1%
CEN 91% 8% 1%
EAS 46% 6% 44%
NYZ 19% 78% 1%
WES 21% 69% 4%
NAI 41% 47% 6%
CST 35% 53% 10%
NEP 27% 65% 3%
Prov > 25% 6 6 1


When I put all of that into a model here is the outcome.

Province Kibaki Outcome Raila Outcome Kalonzo Outcome

A*B*C A*B*D A*B*E
RVP 677,030 1,147,190 18,806
CEN 1,492,160 131,179 16,397
EAS 797,422 104,012 762,751
NYZ 256,027 1,051,060 13,475
WES 164,298 539,838 31,295
NAI 261,379 299,630 38,251
CST 206,206 312,255 58,916
NEP 42,615 102,591 4,735
Total 3,897,138 3,687,754 944,627
Final Outcome 44% 42% 11%

Interesting analysis.

-Silaha

3 Comments:

At Thursday, December 20, 2007 at 2:09:00 AM EST, Blogger Jackie said...

jaksWhat? are you for real? I work for the Kibaki campaign team although im an ODM inside out. i can tell you right now there is now way western will vote 21% to PNU. we were there . very hostile crowds even the PC had to cancel some 4 scheduled meeting coz of insecurity fears. The exact same thing happenned in Molo and Moyale. The temerature around the country is very anti Kibaki. all we hear is "raisi alikuwa wapi miaka tano? sasa anakuja kufanya nini hapa?" i can tell you for sure as a member of the PNU team (only for salary purposes) they are not winning. Raila will take the seat hands down.

 
At Tuesday, December 25, 2007 at 9:40:00 AM EST, Blogger Silaha said...

Jackie:

Thanks for your comment -- I think you are the first comment on this blog!

I do not purport to know what the numbers are going to be on the ground and in each of the provinces that is why I used Steadman's numbers for the provincial shares. If you have better "real" numbers give them to me and I'll plug them into the model.

In fact the only numbers that I made up were the 50%/75% turnout numbers for non-interested and interested ethnicities. My gut, about as non-scientific as you can get, now tells me that the actual turnout may be higher than that.

The point I am trying to bring across is that with the GEMA population at ~30% it is possible for MK to win even if he does not have overwhelming support among the non-GEMA.

-Silaha

 
At Thursday, December 27, 2007 at 4:45:00 AM EST, Blogger Unknown said...

Hello Silaha
Would you like to republish this prediction on KenyaImagine?

Please email me at wainaina@kenyaimagine.com

 

Post a Comment

<< Home