By AtramUpdated 2026-11-20
This page holds every table from one season of early warnings with one Kenyan lender: ten rain events, six with warnings, three with surveys. Clients completed 1,545 phone surveys.
Warned clients prepared more often in event 8
Three numbers from one season with one lender
| Measure | Value | Sample and date |
|---|---|---|
| Event 8 surveyed clients who prepared after a warning | 74.2%, against 53.5% in the comparison group of the same event | 772 surveyed clients in event 8, of whom 376 in the comparison group, 22 April to 3 May 2026 |
| Reported loss, small businesses in event 8 | USD 80 if they prepared, USD 115 if they did not | Event of 22 April to 3 May 2026 |
| Clients enrolled for warnings | About 10,000 | February to June 2026 |
Atram, season data, Kenya long rains 2026. Method: how we measured the season. The comparison group comes from event 8; event 9's comparison group is too small to show. See the limits below.
The loss row compares clients who prepared with clients who did not, not warned clients with others.
Warnings went out for six of the ten rain events
The season had ten rain events in Kenya's western corridor, between February and June 2026. The first four fell in an automated test phase, and no warning went out. Warnings went out for the last six events, and surveys followed events 5, 8 and 9.
The ten rain events, with reach per event
| Event | Dates, 2026 | Magnitude | Highest warning level | Clients reached | Surveyed clients |
|---|---|---|---|---|---|
| 1 | 14 to 18 February | Low | No warning | None | None |
| 2 | 20 to 25 February | High | No warning | None | None |
| 3 | 26 February to 1 March | Moderate | No warning | None | None |
| 4 | 3 to 7 March | High | No warning | None | None |
| 5 | 18 to 22 March | High | Red | 9,978 | 607 |
| 6 | 25 to 31 March | Low | Orange | 4,182 | Not surveyed |
| 7 | 16 to 17 April | Low | Yellow | 6,401 | Not surveyed |
| 8 | 22 April to 3 May | High | Red | 10,269 | 772, of whom 376 in the comparison group |
| 9 | 8 to 16 May | Low | Yellow | 8,438 | 166 |
| 10 | 29 May to 1 June | Low | Orange | 3,193 | Not surveyed |
Atram, season data, Kenya long rains 2026. Method: how we measured the season. One client can appear in several events, so reach is never summed.
The public toll gives the scale of the same season. During the 2026 long rains, Kenya's Ministry of Interior and the Daily Nation published counts of the dead and the displaced. By 30 March 2026, Kenya's Ministry of Interior counted at least 110 deaths, with 30 counties affected. These figures are national. Atram measured one lender's clients and does not estimate the national picture. On 27 October 2026, we found no published series on the loan repayment of Kenya's smallholder farmers after climate shocks.
Most warned clients prepared in each surveyed event
Share of surveyed clients who prepared after a warning, by event
| Group | Event 5 | Event 8 | Event 9 |
|---|---|---|---|
| Clients who received warnings | 78.1% | 74.2% | 70.3% |
| Comparison group | Not surveyed | 53.5% | Too few to show |
Atram, season data, Kenya long rains 2026. Method: how we measured the season. Under Atram's data rules, cells under 30 respondents are not shown. In event 8, the gap of 20.7 points is unlikely to be chance.
Three questions stay open
- The effect on the lender's book. Whether warned clients repaid better is not measured. That needs loan-level data.
- Recovery losses. The surveys asked what clients lost in the event, not what it cost to trade again.
- A clean comparison. The comparison group mixes pre-selected clients with warned clients who did not recall the warning.
Phone surveys after three rain events measured the season
Phone surveys after three of the six events with warnings. Losses reported in Kenyan shilling bands, converted to one value per band. The method is a first version, shared for feedback. Read how we measured the season, and contest it there.
One microfinance lender in Kenya took part
The warnings ran with a microfinance lender in Kenya and its clients. Atram names a lender only with its written consent.
A lender can map its own portfolio against the rain
Want to know which part of your own portfolio sits where the rain falls? Read El Niño 2026: four steps for Kenyan lenders, or assess your portfolio's exposure with us. The Evidence page gives the result in one screen.
Cite this
Atram (2026). Kenya long rains 2026, season data, method v1. atram.ai/data/kenya-long-rains-2026. Add the date you read it.
Questions on this data: hello@atram.ai.