Thursday, January 21, 2010

Lit in Review: Behavioral fertilizer subsidies

Duflo, Kremer, and Robinson have an excellent working paper out, "Nudging Farmers to Use Fertilizer: Theory and Experimental Evidence from Kenya." Being a Duflo paper, you know this will be experimental economics. There is also a behavioral model, a couple regressions behind the scenes, and some calibration of the model based on the experiments.

The gains from using even small amounts of fertilizer are enormous - 70% per year in actual western Kenyan conditions they estimate in another paper - but utilization remains limited. The usual policy solution to this problem has been heavy input subsidies: India spends 0.75% of GDP on them and Zambia spends almost 2% of the government budget on them, for instance.

Among the problems with these subsidies are that they tend to benefit wealthier farmers more than poor and the political connected more than the marginalized, promote overuse, promote further government intervention, and are a strain on the budget. The removal of Malawi's subsidies happened to coincide with a drought and several other bad things, so fertilizer subsidies got a big boost. But is there a better way?

Duflo, Kremer, and Robinson use a model from behavioral economics that assumes that some people are more patient than others and we tend to overestimate just how patient we are (stochastic hyperbolic discounting is the official term). Patient people save and invest in fertilizer, impatient people never intend to, but these stochastically hyperbolic discounters who think they are patient intend to save, but may or may not invest in fertilizer when the time comes. There's a fair amount of evidence from the behavioral literature that this affects westerners, so why not people in developing countries also? In fact, their pre-intervention statistics show that a large number of farmers switch back and forth between using fertilizer and not using it.

Among the difficulties farmers face is that it's a 30 minute hike to go get the fertilizer. So they sent a field officer round to the farms with an offer to buy a fertilizer voucher at the regular price, but with free delivery. They could pay in cash (66% did) or in maize.

They find:
  • About 16% are patient, 13% impatient, and 71% are more interesting.
  • This temporary delivery subsidy (quite small, really) at the time of harvest is enough to get the interesting people to commit their savings to fertilizer, avoiding the time inconsistency problem. Fertilizer use is increased 46-60%
  • Offering free delivery AND a 50% subsidy later in the year also increases fertilizer use, but by less and at a higher cost.
  • "The calibration suggests that a 'paternalistic libertarian' approach of small, time-limited discounts could yield higher welfare than either laissez faire policies or heavy subsidies."
They mention that other policies might get similar results, such as a fine-tuned subsidy [yeah, good luck getting that out of the political process!] and improved financial access so they have a "soft commitment" to fertilizer purchase, but could get their money out if they needed it [but you need the financial infrastructure to do it.]


  1. What I would like to know is, how does sending a field officer around to the farms manage to make 71% of the people "more interesting?"

    Does the field officer also teach line dancing?

  2. (Sorry, Derrill! Just proves I really *do* read these things!)