The Tana River clashes of 2012-13 were among Kenya's worst local conflicts since 2007-08. On one day in August 2012, Capital FM reported that 52 people died, 34 of them women and 11 children. Al Jazeera said that the communities were fighting over land, in a semi-arid area, and that changes to political boundaries before the 2013 election had sparked anger. A summary of the whole series gives 118 dead and more than 13,500 displaced. It's a clear case of old disputes meeting new politics.
Key points:
- The violence joined farming and herding disputes with election politics.
- Revenge attacks followed for months.
- Women and children were among the dead.
- The warning signs were there, including earlier killings.
What happened?
In August 2012, attackers raided homes at Reketa in Tana River, killing people with blades and setting homes on fire, as Capital FM reported. A police commander gave the toll of 52, and said 34 were hacked to death and 14 burnt in their houses.
More attacks followed. In December, Al Jazeera reported at least 32 people killed when a village at Kipau was raided. Kenya Red Cross teams evacuated the seriously wounded to Malindi.
Who was involved?
Reports describe clashes between Pokomo and Orma communities, the Pokomo mainly being subsistence farmers and the Orma semi-nomadic herders, according to Al Jazeera. Both sides suffered. We name them here because reports and inquiries do, and that doesn't make any person guilty.
What caused it?
Land and water lay underneath. Al Jazeera describes a semi-arid area with communities fighting over land. In such places, farmers need the riverbank, and herders need it too when the rains fail.
Politics added fuel. The same report says that political boundaries in the area had been changed ahead of the March 2013 election, which angered the communities. With devolution about to begin, control of a new county was at stake.
Was there warning?
Signs were visible. One account says three people were killed about ten days before the attack of 22 August, and that houses were burnt in retaliation. In many conflicts, such a sequence shows up well before the worst day.
Good early warning asks: what happened last week? Who has been hurt? Who is angry? Who's moving? The list of indicators was built for such moments.
Why did it last so long?
Revenge drives repeated violence. Each raid gave the other side a reason to strike. Weak trust in security agencies made people look to their own.
Elections kept tension high. Candidates and groups had reasons to keep their people together, and fear does that very well.
What did it cost?
The human cost was terrible. Women and children were most of those killed in the August attack, which should shape how anyone thinks about protection. Homes were burnt, animals were stolen, and families fled.
The wider cost was to trust. Neighbours who once traded feared each other. Rebuilding that takes far longer than ending the fighting.
What did it show about elections?
It showed that election violence doesn't only come after results. The AFSC study finds that violence peaks during party nominations, campaigns and after voting. Tana River's worst days came months before the vote. Timing matters.
It also showed that local boundaries are a risk. Any change to electoral or county lines in a tense place needs careful handling.
What did it show about farmers and herders?
Peaceful sharing needs rules. Farmers and herders have shared the land for generations, usually with agreements on timing and access. Once those break down, each season brings new risk.
The Constitution points the way. Article 60 encourages communities to settle land disputes through recognised local initiatives. Rebuilding those local agreements is core peace work.
What should a county learn?
Plan with both livelihoods in mind. Identify shared resources, such as water points and riverbanks. Hold joint meetings before each dry season. Agree on what happens when animals enter crops.
Don't leave land questions until after the election. The AFSC study lists historical grievances among the structural drivers of election violence.
What should security agencies learn?
Respond to small incidents early. A deployment after a massacre comes too late. Local leaders should have a direct line to commanders. Forces must be seen as fair by both sides, and protect both.
What should peace actors learn?
Listen to women. Those most at risk often see the danger first. Keep channels open between communities, even when things are at their worst. Include young people who might be recruited.
Keep records too. Good documentation makes it possible to learn, to support survivors, and to hold people to account.
What should journalists learn?
Report carefully. In a tense place, naming communities as attackers can bring revenge. Say what's confirmed, and who said it. Follow the story after the bodies are buried. The Media Council guidelines call for conflict-sensitive reporting during elections.
Could this happen again?
The conditions remain in many places: shared resources, drought, close elections, and old grudges. ACLED recorded clashes on the Garissa-Tana River and Kitui-Tana River borders in 2023. So the lessons of 2012 still apply.
How can you use this case?
Use it in training. Ask a county team to walk through the months before August 2012. What would they have seen? Who would they have called? What could have been done? Then ask the same questions about their own area today.
Read next
- Conflict early warning system for Kenya: the TI pack page for this work
- Conflict and peacebuilding in Kenya: all the pages in this section
- County boundary disputes
- Election hotspots in Kenya
- Conflict early warning indicators
- Takwimu Intelligence (TI) home page
- National cohesion in Kenya: the guide to this library
- The TI platform and how we check our numbers
Where TI fits
Takwimu Intelligence (TI) keeps a county's incident record and resident surveys together, so small events are visible before they grow.
To see it on your own work, book a walk-through.