Impacts of Climate Information on Coffee Farms in Jamaica

Evidence based mainly on models suggests that climate information, and seasonal forecasts in particular, can improve outcomes in agriculture. Few studies have examined empirically if, how, and under what conditions farmers benefit from availability of seasonal forecasts. While models may capture accurately the response of yields to particular decisions made on the basis of seasonal forecasts, they may not reflect well the farmers’ decision whether or not to use the available forecasts, the actual management decisions that farmers make, and the impact of those decisions on yields under the variety of constraints present on the entire distribution of farms in particular geographic, environmental, and economic contexts.

This study examines the impacts of an intervention developed in collaboration with a number of stakeholders in Jamaica, including farmers, that provides improved weather forecasts as well as seasonal rainfall total and temperature forecasts to farmers who grow Jamaica’s exclusive Blue Mountain coffee. The coffee is a major source of export revenue in Jamaica. Farmers report that seasons are becoming less predictable, making management of the crop difficult. Coffee has also been affected by a recent intensification in outbreaks of coffee leaf rust, which seems to be influenced by both weather and climate. Seasonal forecasts may improve farmers’ ability to prepare for droughts, seasons with substantial moisture, which are favorable to leaf rust, and seasons with a combination of moisture and temperature that either present opportunities or threats to the coffee crop. In addition to providing forecasts, the intervention built capacity among small to medium farmers to use the climate information to manage coffee.

The evaluation relies on a difference-in-difference comparison of changes in outcomes over time among coffee farmers who received the capacity building and the climate information and changes in outcomes among a control group of farmers who did not receive the intervention. The data document changes from baseline, before the intervention began, to follow-up, after the intervention was completed. The difference-in-difference estimator identifies the change in outcomes caused by the climate information under certain assumptions.

One year of access to the incipient climate information service has raised awareness of seasonal forecasts among small to medium farmers in communities that received the information by 12 percentage points relative to the control group. Perhaps most notably, the service has increased the use of seasonal forecasts to make management decision on the coffee farms by 8 percentage points relative to the control group. Both effects are statistically significant. Especially the latter effect reflects a sizeable willingness among resource-constrained small to medium farmers to experiment with using new, untested information on their farms. The percentage of farmers who report using the new information increased from 0 to about 30% among farmers who participated in the capacity-building workshops. Furthermore, this effect occurred during a challenging year, in which the international price of Jamaican Blue Mountain coffee declined sharply. Resource constraints impede farmers’ ability to invest and to adjust the timing of investments, which are the main mechanisms through which seasonal forecasts benefit farmers.

The project period was too short to observe the impact of the new information on yields and profits. However, coffee yields on farmers' main coffee plots did increase more over the year in communities that received the information than in control communities, with the difference being just marginally significant. In addition, the only large farmer among 15 large farmers whom we interviewed who consistently uses seasonal forecasts to manage his coffee farm reports that his motivation for continued use is a clear impact on profitability.

The results indicate that small to medium farmers, with at most primary level education, are willing to experiment with the use of seasonal climate forecasts, which are challenging to understand, despite severe resource constraints. The use of the forecasts begins, not surprisingly, among a small percentage of farmers, the farmers use the information to help make a range of different decisions, and the decisions to which farmers apply the information depend at least partly on the resources at the farmers’ disposal. How the early adopters use the new information and how they are connected to the rest of the farming community is likely to influence the evolution of the role that climate information plays in the coffee sector. More data that document the evolution of use and impacts of climate information over time are essential for understanding the likely impacts of mature climate information services on agricultural outcomes. We investigate lessons that the study offers for designing climate information services and conditions under which these lessons apply.