Should I wait or should I sell? The impact of market information and commercialization advice on farmers’ sales decisions. Guest post by Giulio Schinaia

This is the 30th in this year’s series of posts by PhD students on the job market.

Farmers face many important decisions throughout the growing season. While these decisions determine how much they produce, another crucial decision that affects their income is when and at what price to sell their output.

In my job market paper, I study how farmers alter their sales decisions when they receive market information and commercialization advice during the post-harvest trading season, complementing  previous studies that found mixed results from providing price information alone.

A market advisory system in Guinea-Bissau’s cashew nut market

I evaluate the impact of a new market advisory system for Guinea-Bissau's market for raw cashew nuts. Raw cashews account for over 80% of the country's export revenue and provide the main source of income for a large share of the population.

The new system I examine relies on an international network of market analysts who gather information from major producing countries and deliver weekly advisories to farmers' mobile phones via voice or text messages during the trading season. The messages include current market information (such as farmgate prices and expected trends) and commercialization advice (such as whether to wait or sell based on current offers), in clear, simple language. A typical message would say something like: 'Demand is strong at the harbor and across neighboring Senegal. Farmgate prices range from 350-450 FCFA/kg. Recommendation: If you receive an offer of 400 FCFA/kg, sell at least half your stock. If offered 450 FCFA/kg or more, sell everything.'

To estimate the effects of this market advisory system, we implemented a nationwide two-level clustered randomized control trial across 290 villages with 1,987 farmers, randomly assigning villages to treatment or control. In treated villages, we sampled seven producers and randomized four to receive a one-hour training on market dynamics and advisory use. These farmers then received free advisory messages on their mobile phones during two trading seasons and could call back to listen to weekly advisories through an interactive voice-response system. We also tracked the remaining three farmers in treated villages to study spillovers.

Guinea-Bissau provides a relevant setting for this experiment. Farmers face high-stakes decisions when selling cashews, as most earn their annual revenue from one or two transactions during a four-month trading season. Nearly all sell directly at the farmgate to intermediary buyers rather than traveling to wholesale markets, with limited scope for spatial arbitrage. Importantly, as a price taker in the global cashew market, local prices fluctuate based on international forces. Producers have limited access to information on these fluctuations, while buyers are better informed through their exporter connections.

Commercialization advice and market information increased producers’ prices and revenue

We have three main results:

1. Access to the system modestly increased farmgate prices.  In the first trading season, the treatment effect on prices was small and not statistically significant. Prices were historically low that season, and all producers were more likely to barter their cashews for rice. When accounting for these barter transactions, treated producers increased their 'implicit' price—the value per unit of cashew from all sales and barter exchanges—by 4%. In the second trading season, as overall prices increased, treated producers received prices that were 2% higher on average than control group producers (Figure 1).
2. Treated producers sold their cashews more frequently across both trading seasons. This aligns with the advice they received through the messages, which recommended selling in multiple transactions. Treated producers reported preferring to sell more than once based on the advice received. The advice had (correctly) suggested that prices would increase in the second half of both trading seasons.
3. Treated producers earned higher revenue. This increase in revenue from all sales and barters reflects both higher prices and a shift toward selling more cashews in market transactions rather than consuming them or using them for other purposes. The revenue effects are noisy and amount to 12-22% of the control-group mean, becoming larger and statistically significant in the second trading season. Treated producers were more likely to make productive investments with their sales revenue in the first season, which may have helped increase their output in the second season.

The advice received altered negotiation strategies

Several pieces of evidence suggest that producers adapted their negotiation strategies in response to the advice. First, treated producers in the second season were 7% more likely to report having negotiated a higher price than initially offered. Second, treated producers were 6% more likely to sell to itinerant traders, who can negotiate prices more freely than local buyers, who usually receive a fixed payment for their intermediation. Third, treated producers were less likely to rely on buyers as a source of price information and more likely to report using the messages to guide their sales decisions.

Producers closer to treated villages also benefited from their treated neighbors. Leveraging the randomized spatial allocation of treated villages (Figure 2), we found that an additional treated producer within a five km radius increased the average farmgate price by 1%, affecting producers in control villages closer to treated villages. Treated producers were not more likely to share information outside their village, suggesting these spillovers were likely due to itinerant buyers adjusting their offers upwards after visiting treated villages.

Our findings do not support other behavioral explanations. Producers did not time their sales to capture higher prices. While treatment effects were larger during high-price periods later in the season, we found no increase in the likelihood of selling then. We also found no evidence that effects stemmed from improved record-keeping, changes in risk aversion, or increased transaction salience.

In a nutshell…

Integrating commercialization advice with market information can cost-effectively support farmers' sales decisions, complementing advisory services during the production cycle. While the effects in our study varied across trading seasons, as prices and market conditions changed, our back-of-the-envelope calculations suggest the intervention was cost-effective even at the scale of the experiment: the revenue increases alone outweigh the implementation costs by a factor of at least four.

Giulio Schinaia is a Postdoctoral Scholar at The University of Chicago.

This paper is co-authored with Brais Álvarez Pereira (Nova SBE and NOVAFRICA), Adewusi Mendonça (Ministry of Economy and Finance of Guinea-Bissau), and Dayvikson Raiss Laval Tavares (Bissau Economics Lab).