#854 2009 · Habyarimana & Jack (Zusha! campaign, with Kenyan matatu insurers) · Public health / road safety
Kenya's road-safety campaign didn't train reckless minibus drivers — it gave silent passengers a script to yell at them
the problem
Reckless minibus drivers killed passengers who quietly recognized the danger but never dared object aloud
background
Kenya's matatus, 14-seat minibuses, are the country's dominant transit mode, and road accidents already kill about 1.3 million people a year worldwide, disproportionately in poor countries with weak enforcement. Matatu drivers rarely own their vehicle; they're paid a daily wage plus a cut of the fares they collect, so cramming in more trips costs them nothing personally if a crash happens — the cost falls on passengers, other road users, and insurers. A passenger who felt endangered had no real recourse: no refund, and with thousands of near-identical matatus on the road, no way to build or check a driver's reputation before boarding.
The standard playbook — train drivers, tighten licensing, add traffic police, raise fines — assumed the problem was ignorance or weak deterrence, but Kenya's enforcement capacity was thin and drivers already knew speeding was dangerous. Passengers, meanwhile, individually recognized the danger too and still said nothing — not from ignorance, but because objecting alone, as a stranger surrounded by strangers, risked being the one lone voice with nobody backing them up if it went badly.
what everyone would do
Every standard road-safety lever points at the driver: train them, license them more strictly, fine them, or put more traffic police on the road. Kenya's weak enforcement made all of those slow and expensive, and none of them addressed that drivers already knew speeding was dangerous — they simply weren't the ones who bore the cost of a crash.
what they saw
The one person continuously monitoring a driver's behavior, on every single trip, for free, was already inside the vehicle: the passenger. What stopped them from acting wasn't ignorance of the danger — it was that objecting alone, as a stranger among strangers, risked being the sole voice with no one backing them up. A sticker publicly declaring that speaking up was normal didn't teach anyone anything new; it removed the fear of being the only one who did it.
the move
Researchers papered the inside of thousands of matatus with stickers carrying blunt Swahili-language messages — some framed as a collective call to action, some paired with graphic crash imagery — that explicitly authorized passengers to reprimand a driver for reckless behavior, alongside a small weekly lottery reward that gave drivers a reason to leave the stickers up instead of tearing them down.
why it works
A uniform, visible message inside every matatu works as a coordination device: it signals to each passenger that everyone else in the vehicle has seen the same permission, so objecting stops feeling like a personal risk and starts feeling like an expected, shared response. Once one passenger speaks up with that backing, the driver faces real-time social pressure from paying customers rather than a distant, rarely-enforced rule, and moderates speed accordingly. The researchers confirmed this directly: messages framing the complaint as a collective act consistently beat messages aimed at a lone passenger, while a radio campaign carrying the identical words but without an in-vehicle audience to create that common knowledge measured no effect at all.
the payoff
In the original 2009 pilot, treated matatus were 50% less likely to file any insurance claim over the following year and 60% less likely to file one involving injury or death. Scaled to nearly 12,000 vehicles in 2011-2013, the campaign still cut insurance claims by roughly a quarter to a third, prevented an estimated 140 accidents and about 55 deaths a year, and lowered average driving speeds by 1-2 km/h — at a cost of $10-45 per disability-adjusted life-year saved. A companion radio campaign carrying an identical message aimed at drivers, rather than passengers, showed no measurable effect.
where it breaks
It needs a captive audience that can witness the risky behavior together and plausibly coordinate a response — a driver working alone, or with a single passenger, gets none of this protection. It also needs the person misbehaving to care at least somewhat about in-the-moment social pressure from the people present, and to be unable to retaliate against whoever objects; an anonymous matatu driver who will never see that passenger again has little to fear from speaking up, but a captive employee confronting an identifiable, retaliation-capable superior faces a very different calculation. And the target of the message needs a reason to tolerate the intervention at all — here, a small lottery payment kept drivers from simply ripping the sticker down.
what came after
The scaled program, branded Zusha! ('Protest!' in Swahili), was co-designed with Kenyan matatu insurers and is cited by the World Bank and WHO as a model for road-safety interventions that empower the passenger side of the transaction instead of retraining the driver side. The researchers' own follow-up found messages framed as collective, coordinated protest consistently beat messages aimed at a lone individual — direct evidence the mechanism worked by breaking a social silence, not by informing anyone of a risk they didn't already know.
references
- [1]Results of a large-scale randomized behavior change intervention on road safety in KenyaProceedings of the National Academy of Sciences (PNAS), 2015pmc.ncbi.nlm.nih.gov
- [2]Encouraging Behavioral Change to Improve Road Safety in KenyaAbdul Latif Jameel Poverty Action Lab (J-PAL), MIT, 2015povertyactionlab.org