With postcards and loudspeakers against food corruption - ambient accountability works!
(image cc Globe-trotter at wts wikivoyage)
A fascinating RCT in Indonesia yields amazing evidence that ambient accountability has enormous potential to complement or even replace conventional anti-corruption interventions. How to reduce high leakage rates in the distribution of subsidized rice for the poor in Indonesia? Mailing information about entitlements to eligible beneficiaries made a huge difference and led to significantly more subsidized food reaching the intended beneficiaries. Adding more specific information on prices etc. further drove up fulfillment. And, most interestingly, posting list of entitled recipients publicly across the village and communicating related information via mosque loudspeaker led to another significant increase in flows to the poor. So beyond proving true impact, what else do these numbers tell us with regard to ambient accountability:
- first: people equipped with the right information can indeed haggle successfully with corrupt officials and prevent them from all too blatant corruption
- second: the more detailed and targeted the information intervention, the better the outcome for the beneficiaries
- third: going public with entitlement information generates one of these super-interesting feedback loops around what game theorists tend to call second order or higher order beliefs: with the information visibly posted out in the public now I know for sure that other previously cheated beneficiaries also know and the corrupt officials know that all the people they are cheating know and also know that they know. The result: even significantly less leakage (on top of the two effects described above) AND a higher degree of collective action- villages that benefited from this type of public posting experienced public protest related to those issues more frequently.
So a clear validation of the power of ambient accountability, as well as a confirmation of the plausibility of some of the central transmission mechanisms that I had posited as working assumptions re. how targeted information interventions in the built environment might work.
(Banerjee, Abhijit, et al. The Power of Transparency: Information, Identification Cards and Food Subsidy Programs in Indonesia. No. w20923. National Bureau of Economic Research, 2015).
