Last updated on September 11, 2022
My article with Chris Boyd titled “Why Not Insure Prices? Experimental Evidence from Peru” is now out in the Journal of Economic Behavior & Organization.
If you are interested in any of micro-insurance (or “index” insurance), the consequences of price volatility and risk on producers, the impacts of agricultural insurance on producers, or even just in lab-in-the-field experiments with agricultural producers, you can access this article for free here until October 29, 2022.
Here is the abstract:
In a competitive market, a profit-maximizing producer’s total revenue is determined both by the quantity of output she chooses to produce and by the price at which she can sell that output. Of these two variables, only output is in part or wholly within the producer’s control, price being entirely determined by market forces. Given that, it is puzzling that the literature studying the effects of providing insurance to producers in low- and middle-income countries has ignored price risk entirely, focusing instead on insuring output. We run an artefactual lab-in-the-field experiment in Peru to look at the effects of insurance against output price risk on production. We randomize the order of three games: (i) a baseline game in which price risk is introduced at random, (ii) the baseline game to which we add mandatory insurance against price risk sold at an actuarially fair premium, and (iii) the baseline game to which we add voluntary insurance against price risk sold at the actuarially fair premium, but for which we offer a random 0-, 50- or 100% discount to exogenize take-up. Our results show that, on average, (i) price risk does not significantly change production relative to price certainty and (ii) neither does the provision of compulsory insurance against price risk, but the introduction of voluntary price risk (iii) causes the average producer on the market to produce more in situations of price risk than in situations of price certainty, and (iv) causes the average producer on the market to produce more in situations of price certainty than in cases where there is no insurance or where insurance is mandatory. When looking only at situations where there is price risk, (v) this is due almost entirely to the insurance rather than to selection into purchasing the insurance. Our findings further suggest that (vi) even in the absence of the discount, the insurance against price risk would have a large (i.e., 70%) take-up rate.
One research question our paper leaves open and which I would like to see someone seize upon is the notion that, since price and quantity both combine into total revenue, might producers react differently to two insurance products that are analytically equivalent (i.e., which both provide the same coverage for the same premium) but which insure different things, viz. one insuring price, and one insuring quantity? In the paper, we speculate that people respond differently to insurance against price risk (possibly because price is entirely beyond their control, and possibly because price is a clear and verifiable “index”) than they do to insurance against quantity risk (possibly because quantity is partly within their control, and possibly because the usual quantity index is somewhat nebulous and hard for any individual to verify–and that’s when there aren’t more than one index involved!)
My suspicion is that both locus-of-control (i.e., are we insuring something people feel they have some control over, like quantity, or are we insuring something people fill they have no control over, like prices?) and verifiability of the index probably both matter for insurance take-up, but only the latter matters in terms of how people behave once they are insured.
Also, with 52 pages of journal space given the several appendices required for an experimental paper, this article also has the dubious distinction of being the longest peer-reviewed journal I have ever published.