Analysing smallholders behaviour on Sumatra: An ex ante policy analysis and investigation of experiments external validity under consideration of risk

I: Palm oil production creates negative externalities, e.g., through intensive fertiliser application. Policies to limit externalities need an effective, sustainable and efficient measure We use a business simulation game in a framed field experiment in Indonesia to test ex ante different incentives for reducing such negative externalities. This setting allows inclusion of adequate contextual features, required for reasonable ex ante evaluation of policy measures. The different designs of the test incentives (either a reward or punishment) varied in their magnitude and probability of occurrence but with constant effects on expected income. Results show that participants react differently to these incentives, indicating that the design can contribute significantly to effectiveness, sustainability or efficiency. A high reward with a low probability was found to be the most effective and sustainable incentive. Moreover, for the most efficient design, a low and certain reward is indicated. II:This article compares the use of risk-increasing and risk-reducing production inputs with the experimentally measured risk attitudes of farmers. For this purpose, the Just-Pope production function indicates production inputs’ influence on output risk and a Holt-Laury lottery is used to measure the producers’ risk attitude. We test whether more risk averse farmers use more risk-reducing and less risk-increasing production inputs. Therefore, we apply a unique data set which includes 185 small-scale farmers on the island of Sumatra, Indonesia. The Just-Pope production function indicates that a higher fertiliser usage has a risk-reducing effect, whereas a higher herbicide usage and a bigger plot size have risk-increasing effects. By including the outcome of a Holt-Laury lottery into the analysis, we found the expected result that more risk averse farmers use more (risk-reducing) fertiliser and less (risk-increasing) herbicides. These consistent results reinforces the external validity of measuring risk attitude with the Holt-Laury lottery