Assessing the costs of risk management tools: A crop insurance scenario based on a stochastic partial equilibrium model approach
A FAPRI-UK Project
Siyi Feng1, Myles Patton1, Julian Binfield2, and John Davis1
1) Agri-Food and Biosciences Institute 2) University of Missiuri
Following the move within the European Union’s agricultural sector from stable,
administratively determined prices and production linked subsidies to more freely moving prices and decoupled subsidies, agricultural risk management is of increasing concern. Significant increases in global commodity prices have further contributed to volatility. There is growing interest in developing policy programmes aimed at promoting risk management tools, some of which already exist in countries such as the US on a large scale. These programmes operate in varying degrees as a form of insurance. At the same time, they generally involve policy support due to the presence of systematic risks within the relevant sectors. Thus they entail complex design issues and careful assessment. This paper examines a hypothetical scheme that provides protection against falls in crop yield within the UK using a stochastic FAPRI-UK and EU-GOLD modelling system.
Investigation of the level of spatial aggregation.
In particular, two scenarios of crop yields based on which payments are triggered are examined: one that is based on the national average and applied to the four countries within the UK versus one that is based on averages of the four individual countries. Outcomes based on the national averages are largely driven by England which contributes over 90% wheat production of the UK. Wheat productions of the other three countries (i.e. Wales, Scotland and Northern Ireland) are exposed to considerable idiosyncratic risks and therefore protection is limited if a national average is applied. Although together they produce only a small proportion of UK wheat, expected total payments for the four countries is around 20% higher when individual averages are used. In reality, the aggregation level can be different from what has been investigated here. However, our results highlight trade-offs between programme costs and its effectiveness in risk reduction.
The definition of the reference or the normal condition, which also has implications for programme costs and their variability. In particular, the use of Olympic average of historic outcomes (i.e. crop yields within this study) in the preceding years is investigated. Olympic averages of preceding years allow the reference periods to change with policy years (as opposed to fixed periods) and at the same time smooth out extreme variations. Reference of this kind is in line with WTO rules on stabilisation tools for farm income and is also adopted in payment programmes in the US. However, our analysis based on the UK experience shows that the remaining variations that are not smoothed out are potentially large, especially when there are multiple extremes in the reference period.