Risk is bad

Evidence on insurance market failure as a source of underdevelopment.

We conducted several experiments in northern Ghana in which farmers were randomly assigned to receive cash grants, grants of or opportunities to purchase rainfall index insurance, or a combination of the two. Demand for index insurance is strong, and insurance leads to significantly larger agricultural investment and riskier production choices in agriculture.

The salient constraint to farmer investment is uninsured risk: when provided with insurance against the primary catastrophic risk they face, farmers are able to find resources to increase expenditure on their farms.

Demand for insurance in subsequent years is strongly increasing in a farmer’s own receipt of insurance payouts, and with the receipt of payouts by others in the farmer’s social network. Both investment patterns and the demand for index insurance are consistent with the presence of important basis risk associated with the index insurance, and with imperfect trust that promised payouts will be delivered.

Why am I always surprised when micro 101 models turn out to be true?

A new paper by Karlan, Osei, Osei-Akoto, and Udry.

8 thoughts on “Risk is bad

  1. Chris, you’re right that this is an interesting paper but, being somewhat blunt, the actual story is probably closer to ‘researchers reduce welfare of African farmers by offering them a product that the farmers think will protect them from downside risk, but in actual fact is basically just a lottery ticket’.

    The problem here is basis risk (agricultural revenue loss minus claim payment from indexed product). The ‘insurance’ products offered to these farmers were almost surely terrible. The authors offer *no* statistical analysis of the correlation between their weather derivative and farmer outcomes. Evidence from elsewhere (including Northern Ghana!) finds that weather index insurance is basically a lottery ticket, with very low correlation between losses and claim payments:
    http://elibrary.worldbank.org/content/workingpaper/10.1596/1813-9450-5985 is (by far) the largest study of weather index insurance in a low income country, finding a 13% correlation across 318 products in India.
    – Specifically for Northern Ghana there is already a paper finding low correlation: http://www.ilo.org/public/english/employment/mifacility/download/repaper7.pdf
    – For a (my) theory paper on why we should care about correlation see http://ideas.repec.org/p/oxf/wpaper/572.html
    – If you think index insurance is simple, try playing the game on slide 12 of http://users.ox.ac.uk/~ball2201/pdf/RioExperiements_vFinal.pdf . Is it obvious to you that paying a premium of more than 6 Birr in this game is inconsistent with decreasing absolute risk aversion?

    For a more general overview of what ‘we have learned from all the agricultural microinsurance pilots’ see a recent CSAE blog post: http://blogs.csae.ox.ac.uk/2012/11/what-have-we-learned-from-all-the-agricultural-microinsurance-pilots/