Weather Forecasts, Expected Profitability, and Farmer Behavior

Farmers respond more strongly to weather forecasts in regions where [they] tend to be more accurate.

Agricultural output in developing nations is strongly dependent on the weather, and as a result agricultural profits are heavily affected by the accuracy of weather forecasts. Farmers without access to good insurance markets act conservatively, investing less on their farms and choosing crop mixes and cultivation techniques that reduce the volatility of farm profits but also lower expected profits.

In Forecasting Profitability (NBER Working Paper No. 19334), authors Mark Rosenzweig and Christopher Udry use newly available data on farmers in India to estimate how the returns to planting-stage investments vary by rainfall. The sensitivity of investment returns to weather suggests that making weather more predictable has the potential for increasing profitability.

The authors find that weather forecasts significantly affect farmer investment decisions and that these responses account for a substantial fraction of the year-to-year investment variability. The accuracy of weather forecasts varies across regions within India, and farmers respond more strongly to weather forecasts in regions where these forecasts tend to be more accurate. The farmers' use of the forecasts increases both average profit levels and the variability of profits. Farmers with access to weather forecasts earn higher profits than farmers without such forecasts when rainfall realizations are high, but they earn less under adverse rainfall conditions.

The authors' analysis suggests that Indian farmers, on average, underinvest at planting time. The investment level that maximizes expected profits over the full distribution of rainfall forecasts is roughly three times the observed average level.

Little attention has been paid to directly improving farmers' capacity to deal with weather fluctuations by improving the accuracy of weather forecasts. In part this is because of the lack of estimates of the sensitivity of investment returns to weather and other common random events, as prior studies have typically provided estimates based on data from a single crop season or a particular locality. The authors assess how profit levels would be affected by rising forecast skill under both current climate conditions and under the conditions predicted by widely-used climate models. They find that even modest skill improvements would substantially increase average profits per farmer. The effect would be somewhat greater if the climate becomes warmer.

--Les Picker