The numbers came in early and they were not subtle. By the time the Independent Electoral and Boundaries Commission was done counting on the night of May 14, President William Ruto’s United Democratic Alliance had taken all three seats on offer — the Emurua Dikirr parliamentary seat in Narok, the Porro ward seat in Samburu, and the Endo ward seat in Elgeyo Marakwet. UDA’s David Keter is the new Member of Parliament for Emurua Dikirr. For a political class that had spent two weeks insisting the by-elections were really a referendum on the Ruto–Gachagua war, the verdict was about as ambiguous as a slammed door.
Former Deputy President Rigathi Gachagua had personally pitched camp in Emurua Dikirr, framing the contest as the first electoral test of his post-impeachment Democracy for the Citizens Party (DCP). His script was familiar: that the Mountain has broken with Ruto, that the broad-based government is illegitimate, and that any UDA candidate is a Trojan horse for a regime in retreat. The script does not survive a clean sweep. If the by-elections were a stress test for the Gachagua opposition project, the project failed on its first day in the lab.
It is worth being precise about what was contested. None of the three seats sits in Mt Kenya proper. Emurua Dikirr is a Kipsigis-majority Maa neighbourhood; Porro and Endo are in the North Rift heartland where UDA was always favoured. The Ruto camp, predictably, is treating the results as proof that the President’s base is intact and that his by-election firepower — cabinet secretaries, mobilisation budgets, security deployments — still moves voters. The Gachagua camp will, just as predictably, argue that the playing field was tilted and that his organisation had not yet matured. Both readings contain some truth. Neither rescues DCP from the optics of a 0-for-3 outing.
For UDA, the more interesting question is internal. The party went into these by-elections without the political enforcers from Mt Kenya who, in 2022, drove turnout in places like Emurua Dikirr through alliances with Kalenjin networks. That UDA can still win, and win comfortably, in a Maa-adjacent seat suggests that Kithure Kindiki, Aden Duale and the new generation of Rift Valley operators have rebuilt the field machine that Gachagua once claimed as his own. The TIFA poll released earlier this month, which placed Ruto at 24% in the 2027 preference race and Kindiki at 59% as preferred running mate, no longer looks like a paid advertisement. It looks like a forecast.
The opposition has a different problem to grapple with. With Kalonzo Musyoka at 19%, Fred Matiang’i at 14%, Edwin Sifuna at 10% and Gachagua himself at 9%, the anti-Ruto vote is mathematically a majority — three-quarters of the country, in the latest poll, is not for Ruto. The trouble is that there are four candidates trying to be the alternative, and now one of them has just been outpolled in a contest he chose to make his stand. By-elections do not predict 2027 outcomes, but they do redistribute political capital. After May 14, Gachagua has less of it.
The deeper lesson is one Kenya keeps relearning. Anger at a sitting government is a necessary but insufficient condition for unseating it. Without a single opposition vehicle, without a candidate who can fundraise across regions, and without a credible economic alternative to Kenya Kwanza’s tax-and-borrow mix, the protest vote will keep dispersing in by-elections like these. UDA’s strategists understand this. That is why they competed hard for ward seats that other parties would have written off. Each one is a low-cost rehearsal for 2027 — a way to test the data, the agents, the polling-station discipline. By that measure, the day was a successful dress rehearsal.
What to watch next: whether Gachagua restructures DCP’s field operation before the August round of pending by-elections; whether Kalonzo and Matiang’i finally agree on a primary process for an opposition flag-bearer; and whether the Ruto camp now formalises Kindiki as running mate while political momentum is on its side. The by-election season is not over. But the Mountain revolt, on present evidence, is more a hashtag than a movement.
