Potential of Reducing Crop Insurance Subsidy Based on Willingness to Pay and Random Forest Analysis

Indonesia has recently piloted a national paddy insurance program. The paddy insurance is heavily subsidized by the government generously covering 80% of the premium. Because the agricultural insurance is still in its infancy in Indonesia, it is imperative to have an insurance scheme that is effective and sustainable. To better understand the existing insurance scheme as well as farmers attitude towards paddy insurance, an extensive survey on the paddy farmers was conducted. Through formal statistical analysis based on the Wilcoxon signed-rank test, farmers’ willingness to pay (WTP) is found to be higher than the current subsidized premium for a satisfactory insurance scheme. This implies that the government subsidy could be lowered from 80% to 72%. Further analysis based on the method of random forest classification allows to identify most important factors for affecting farmers’ WTP. The first four most important factors are willingness to buy crop insurance, choice of insurance plan, satisfaction toward ease of premium administration from previous owned insurance, and farmers’ perception on the priority of insurance company trustworthiness. This is based on criteria of mean decrease of Gini impurity. These analysis and findings provide valuable guidance in revamping the existing paddy insurance program, including exploring the possibility of reducing premium subsidy.