An interesting paper by James Habyarimana and William Jack of Georgetown Univ.
The usual debate is what is better – Should state regulate to ensure safety etc for consumers? Or should consumers be made more aware and let them decide. Ideally one should have a bit of both but econs want to know which is more effective. Moreover, in case of limited resources one needs to choose the most effective policy option.
The case here is of road safety. They do a random experiment to understand whether state regulation is better or a consumer empowerment intervention. Their findings show it is latter:
This paper compares the relative impact of two road safety interventions in the Kenyan minibus or matatu sector: a top down set of regulatory requirements known as the Michuki Rules and a consumer empowerment intervention. We use very detailed insurance claims data on three classes of vehicles to implement a difference-in-differences estimation strategy to measure the impact of the Michuki Rules. Despite strong political leadership and dedicated resources, we find no statistically significant effect of the Michuki Rules on accident rates.
In contrast, the consumer empowerment intervention that didn’t rely on third party enforcement has very large and significant effects on accident rates. Our intent-to-treat estimates suggest reductions in accident rates of at least 50%. Our analysis suggests that in institutionally weak environments, innovative consumer-driven solutions might provide an alternative solution to low quality service provision.
They could have written the paper a little better. Becomes confusing towards the middle..
The consumer awareness mission is nice bit. They put stickers on these vans on safe driving etc and reward the drivers (via a lottery) to keep the stickers pasted on the vans:
Institutional weakness and corruption may compromise the effectiveness of a variety of reform efforts, especially those that rely on third‐party enforcement. In the case of public transportation, an alternative to top‐down campaigns like the Michuki rules is to empower passengers to demand higher quality services directly, not by threatening to report a bad driver, but simply by openly complaining to him.11 To motivate the potential impact of such a strategy, we argue that complaints to the driver represent contributions to a local public good, and that a collective action problem among passengers could arise accordingly. Multiple equilibria can exist in such environments, characterized by different aggregate levels of public good provision. This suggests that lowering the resource or psychic costs of complaints, for example by making them appear more legitimate and thereby giving passengers a voice, could lead to discrete changes in the intensity of consumer monitoring and enforcement, and perhaps meaningful changes in safety and outcomes.
To test these ideas, we conducted a randomized control trial of an intervention aimed at empowering matatu passengers to exert pressure on drivers to drive more safely.12 The intervention was simple and cheap: stickers with evocative messages intended to motivate passengers to take demonstrative action ‐ to “heckle and chide” a dangerous driver ‐ were placed in about half of roughly 2,300 recruited matatus. The stickers included graphic images of injuries, and text in English and Kiswahili encouraging passengers to “Don’t just sit there! Stand up! Speak up!”
An initial small pilot in fall 2007 was compromised when we discovered that stickers were quickly being removed from treated vehicles. In response to this, we reissued stickers and scaled up to the full sample in early 2008, but in an attempt to ensure higher rates of compliance we ran a weekly lottery amongst drivers of participating treatment matatus. Each week three prizes of 5,000, 3,000 and 2,000 Kenyan Shillings were awarded to drivers (about $60 – roughly a week’s wages, $35, and $25) if their vehicle was found to have all stickers intact upon inspection by our field staff.
Interesting. How we usually do not think of such issues in designing experiments and surprises crop up from somewhere..
