This paper, authored by Esther Duflo, Michael Kremer, and Jonathan Robinson, explores the use of fertilizer among farmers in Western Kenya and proposes a model to explain why many farmers fail to take advantage of profitable fertilizer investments. The authors argue that behavioral biases, such as present-bias and procrastination, may limit farmers' ability to commit to fertilizer use. They model farmers as facing small fixed costs of purchasing fertilizer and assume that some farmers are stochastically present-biased, underestimating their future impatience. The model predicts that small, time-limited discounts on fertilizer, offered just after harvest, can significantly increase fertilizer use compared to larger discounts offered later in the season or no discounts at all. The authors conducted a randomized field experiment in collaboration with International Child Support (Kenya) to test these predictions. The results show that offering free delivery to farmers early in the season increased fertilizer use by 46 to 60 percent, compared to a 50 percent subsidy offered later in the season. Calibration of the model suggests that small, time-limited discounts can yield higher welfare than either laissez-faire policies or heavy subsidies, by helping present-biased farmers commit to fertilizer use without causing overuse among those with standard preferences. The paper concludes by discussing the potential for scaling up such interventions in a cost-effective manner.This paper, authored by Esther Duflo, Michael Kremer, and Jonathan Robinson, explores the use of fertilizer among farmers in Western Kenya and proposes a model to explain why many farmers fail to take advantage of profitable fertilizer investments. The authors argue that behavioral biases, such as present-bias and procrastination, may limit farmers' ability to commit to fertilizer use. They model farmers as facing small fixed costs of purchasing fertilizer and assume that some farmers are stochastically present-biased, underestimating their future impatience. The model predicts that small, time-limited discounts on fertilizer, offered just after harvest, can significantly increase fertilizer use compared to larger discounts offered later in the season or no discounts at all. The authors conducted a randomized field experiment in collaboration with International Child Support (Kenya) to test these predictions. The results show that offering free delivery to farmers early in the season increased fertilizer use by 46 to 60 percent, compared to a 50 percent subsidy offered later in the season. Calibration of the model suggests that small, time-limited discounts can yield higher welfare than either laissez-faire policies or heavy subsidies, by helping present-biased farmers commit to fertilizer use without causing overuse among those with standard preferences. The paper concludes by discussing the potential for scaling up such interventions in a cost-effective manner.