10. Implications of Farmer Information Provision Policies: Heterogeneous Farmers and Market Selection

In Kenya and Mali, an NGO program, Mali Shambani, provides a weekly hour-long radio program that discusses market price trends, current market prices, farming techniques, etc. This free radio program also allows farmers to ask agricultural questions via phone or short messaging services (SMS); see USAID (2011). Likewise, India’s Ministry of Agriculture provides some market information on its website (www.​india.​gov.​in/​topics/​agriculture) and mounts a hotline service (Kisan call centers) to provide advisory service to farmers over the phone. Besides the information provided by the government and NGOs, for-profit enterprises such as Reuters Market Light (RML, reutersmarketlight.com) and Nokia Life Tools (NLT) broadcast customized information via SMS messages to farmers who subscribed to their service at a nominal fee since 2007 and 2009, respectively (Tang and Sheth 2013). The provided information includes local crop prices, customized, localized, and personalized weather forecasts, agricultural news, and crop advisory. The reader is referred to Chen and Tang (2015) for details of different services provided by the governments, non-governmental organizations (NGOs), and for-profit companies such as Nokia Life Tools and Reuters Market Light.

We shall extend our analysis to the case when farmer’s location is not uniform but a symmetric distribution at the origin 0 and show that our main results continue to hold.

Each smallholder farmer is a price taker on the individual level. However, on the aggregate level, the market price drops as the total production quantity increases.

We shall extend our analysis to the case when the distribution is not uniform but a general symmetric distribution.

Following a standard assumption used in the literature, we assume that A is sufficiently large so that the p _i is almost always positive (Gal-Or 1985 and Vives 1984). Also, we assume that u _l and u _r are independent.

We did consider a more general class of provision policy δ that can be specified by two decisions R ⊂ [−0.5,  0.5] and ρ ≤ 1 so that ρ percentage of farmers located within R will receive the market signals. For tractability, we shall focus on the case when R is symmetric about the origin 0. When R consists of a finite number of closed intervals, we show that each provision policy (R, ρ) is dominated by a corresponding provision policy under which R is a continuous interval so that R = [−K, K], where K ∈ [0,  0.5]. We omit the details here, but the reader is referred to the online supporting material for details. Therefore, it suffices for us to focus on a class of provision policy δ = (K, ρ), where K ∈ [0,  0.5] and ρ ∈ [0,  1] throughout this chapter.

For ease of notation, we use τ ^(F0) ≡ 0 to denote the threshold adopted by all farmers under the No Information Policy.

For a farmer who receives no signals, his market selection is based on the comparison between \(E(\hat {\pi }_0(\theta , l))\) and \(E(\hat {\pi }_0(\theta , r)) \), which does not depend on (x _l, x _r). However, after he selects the market to sell in, his ex-post expected profit depends on the market selection of other farmers, which depends on (x _l, x _r) via the threshold τ ^(δ)(x _l, x _r).

The government should choose a threshold structure. Otherwise, we can always find a pair of farmers who sell in different markets, such that if they exchange the markets they sell in, famers’ total profit becomes higher.

For ease of exposition, we shall examine the case when b, t, or α is high enough or β is low enough so that τ ^(C) and τ ^(F1) lie within (−0.5,  0.5) almost surely. However, when the thresholds are truncated at ± 0.5, the expressions are slightly different, but the qualitative characteristics remain the same.

On the contrary, in the models considered by Chen et al. (2013) and Morris and Shin (2002), farmers are price setters who can control the market price by selecting their production quantity. Consequently, the negative externality imposed by each farmer is reduced because each farmer’s decision will have direct impact on his own profit.