2005Constitutional referendum
Raila Odinga, Kalonzo Musyoka, William Ruto and KANU joined against President Kibaki's draft. The Orange side carried every region except Mt Kenya: the coalition was simply bigger.
About MOASS
Kenya's votes have followed the arithmetic of its communities.
MOASS, the Mother of All Spreadsheets, has called Kenya's national votes since the 2005 referendum. It doesn't poll anyone. It counts who is registered where, how each community has voted, and which leaders are standing together, then adds it up, ward by ward.
2005first call
6national votes
2referendums
The record
Official results of every national vote since 2005. In each one, the outcome follows from which communities stood together. The dashed line marks 50%.
2007Presidential
ODM's 'Pentagon' (Luo, Kalenjin, much of Luhya, Coast and North Eastern) against PNU's Mt Kenya base. Kalonzo took Ukambani (8.9%) as a third force, which kept the race close. The official result was disputed and followed by post-election violence.
2010Constitutional referendum
Kibaki and Raila campaigned together for Yes; Ruto led No. The No vote was concentrated in the Kalenjin Rift Valley, as the arithmetic predicted.
2013Presidential
Jubilee joined the two largest blocs, Kikuyu and Kalenjin, behind Kenyatta–Ruto (the 'tyranny of numbers'). Their high registration and turnout cleared 50% by a few thousand votes.
2017Presidential (August)
The same Jubilee alliance against NASA (Raila, Kalonzo, Mudavadi, Wetang'ula). The Supreme Court annulled the result over process; Kenyatta won the October rerun after Raila withdrew.
2022Presidential
The blocs were reshuffled: Kenyatta backed Raila, but Mt Kenya voters went with Ruto and Gachagua. Ruto kept the Rift Valley and won most of Mt Kenya, enough to clear 50% narrowly.
Why ethnic arithmetic works in Kenya
This isn't a judgment about any community, and no community votes as one person. It is a description of a pattern that has held in every national vote since multiparty politics returned, and one that research on Kenyan exit polls (for example Bratton and Kimenyi, 2008) has measured directly.
1
Communities vote for their own, by wide margins
When a community's leader is on the ballot, its voters back him overwhelmingly, often 85 to 95%. That is far more predictable than voting on issues, and it holds from ward to ward.
2
No community can win alone
The largest community is under a fifth of the country, so every winner needs a coalition. Once you know who is allied with whom, including the running mate, most of the result is already decided.
3
Elections are decided in the swing communities
Luhya, Kamba, Kisii, Coast, Maasai, North Eastern and parts of Meru and Embu split between camps. MOASS spends most of its effort here and in the mixed towns, where the outcome is actually contested.
4
Registration and turnout are the multipliers
A bloc is only worth what it registers and turns out. Regional turnout patterns have been stable since 2013, so the register, not the opinion poll, is the best guide to how heavy each bloc will be.
5
The rules reward geography
A first-round win needs over 50% and at least 25% in 24 counties. Parliament is won seat by seat. Both depend on where communities live, which MOASS maps down to wards and polling streams.
6
Leaders move blocs, but not without limits
2022 showed both sides: Mt Kenya followed Ruto and Gachagua rather than Kenyatta's endorsement. So MOASS models endorsements as partial transfers and does not assume a whole community follows any one leader.
How the 2027 model is built
From polling streams up
27,187 polling centres and 45,931 streams, each with its IEBC 2022 register and Form 34A turnout. The 2027 register is IEBC's latest release (24.7 million, April 2026).
Every ward's community mix
Estimated for all 1,450 wards from the 2019 census, then checked against polling-centre names, MCA winners across cycles and the 2022 vote, with local knowledge where the data is thin.
Scenarios and the 2027 mood
Alliances move votes as partial transfers. On top sit what is new for 2027: lower turnout, the urban Gen Z wave, the Mt Kenya realignment and, on every page, the historical incumbency effect in state-aligned remote counties.
What it can't see
- Ethnicity is a strong pattern, not destiny. Cost of living, the 2024 youth protests and local candidates all move votes, and MOASS models them only as adjustments.
- Ward community mixes are estimates. The census reports ethnicity by county and sub-county, not by ward.
- Mixed urban wards, especially in Nairobi, are the hardest to call and carry the most uncertainty.
- Alliances in Kenya can change late. Every scenario page is a "what if", not a forecast of which alliance will form.
- The incumbency advantage on every page models a historical pattern in state-aligned remote counties. It is not a finding that any result was rigged.
Where 2027 stands
Ruto projected to win outright. Ruto's side holds 215 of 337 National Assembly seats and 36 of 47 Senate seats in the snapshot, where every candidate runs separately.