Satellite-based index insurance for hedging weather-related production risk in agriculture and forestry

This cumulative dissertation examines the hedging effectiveness of satellite-based weather index insurance for both annual crops, such as rice and wheat, and also perennial olive trees and pine stands. Unlike traditional indemnity-based insurance (such as hail or multi-peril insurance), weather index insurance does not require claims to be settled by an adjuster, thus weather index insurance has lower transaction costs. With weather index insurance, the amount of loss is derived by using an objectively measurable index. As a result, weather index insurance has the advantage of being less affected by moral hazard and adverse selection. Typically, indices measured at weather stations serve as underlying for weather index insurance. However, weather index insurances are affected by a so-called basis risk, which is a low correlation between crop yield and index. Therefore, in addition to considering different crops, the individual papers in this dissertation also examine the different types of basis risk associated with index insurance. In this context, satellite indices have recently been investigated in the literature as underlying for weather index insurance. Satellite indices have the advantage of being available at any location in the world, independent of a weather station and in some cases are free of charge. The first paper shows that satellite indices can explain the relatively high loan default risk of agricultural microcredits in Madagascar. Microfinance loans in developing countries enable smallholder farmers to grow agricultural crops such as rice. However, smallholder farmers are often unable to repay loan installments when production fails due to the weather. Satellite-based index insurance contracts could reduce the loan default risk of agricultural microcredits, allowing microfinance institutions to increase lending on improved terms. As a result, access to credit and thus the economic situation of small farmers in developing countries could be improved. The second paper shows that a higher spatial resolution of satellite indices could improve the hedging effectiveness of satellite-based index insurance for yield risk of winter wheat in Northeast Germany. Furthermore, there is an aggregation effect, whereby with increasing aggregation of data -field level compared to farm level- the hedging effectiveness of satellite-based index insurance contracts also increases. Consequently, in the future, satellite-based index insurance contracts should be based on the highest resolution satellite imagery possible to minimize the basis risk of spatial resolution. However, it is important to ask at what level of data aggregation the contracts are offered and how the aggregation effect affects the fair premium. The third paper shows that satellite-based weather index insurance contracts are also suitable for hedging the olive oil yield of perennial non-irrigated olive trees in Spain, thus providing an alternative to irrigation. Compared to weather station-based insurance contracts, the satellite-based insurance contracts have a lower geographical basis risk. Furthermore, the insurance contracts should also be offered for coverage against extreme weather events to reduce the basis risk of design. The fourth paper applies satellite-based index insurance from perennial olive trees to perennial pine stands. Insurance contracts may also provide a way to hedge the mortality risk of pine stands. Index insurance could both reduce the relatively high insurance premiums of traditional fire and windstorm insurance and allow satellite indices to be determined for larger forest areas worldwide. In addition to the impact analyses conducted in this dissertation, future studies need to address other economic issues such as the willingness-to-pay of both farmers and foresters for such insurance. In addition, future studies should use satellite imagery from the Sentinel 1 and 2 satellites as soon as the data series allows, since these are cloud-independent or have high spatial and temporal resolution.2021-11-1