The real rate of protection: the income and insurance effects of agricultural policy

Abstract

Agricultural price policies in developed countries aim at protecting farmers against both low and volatile world market prices. However, traditional indicators of protection only refer to the income (level) effect of policy. Following other research, it is argued that public policy can also yield an insurance (stabilizing) effect. In this paper a way to measure these dual effects is proposed. The method is illustrated with wheat market data for the USA and the European Union. Strong evidence is found that the insurance effect is an important component of protection, albeit a small one relative to the income effect. Policy support provided higher income and lower insurance effects in the EU than in the USA. For both markets, policy reforms in the 1990s led to significantly reduced income effects and smaller insurance effects. Without accounting for the influence of policy on income variability, traditional measures of protection will understate the real rate of protection.

Acknowledgements

This research was made possible with funding from the Deutsche Forschungsgemeinschaft, The Ohio State University, and the Austrian–American Fulbright Commission. It was written while Thompson was a Fulbright scholar at the University of Natural Resources and Applied Life Sciences, Vienna, Austria.

Notes

Any measure of policy effects is likely to suffer shortcomings and the validity of any given measure depends upon exactly what aspect of support the instrument intends to represent. Measures of market distortions may imply a quite different picture of support when compared to measures of producer income effects, such as the one considered here.

If it is alternatively assumed that real income is normally distributed and constant absolute risk aversion utility, the simple mean-variance formula of expected utility of income is: E [U (y)] = E(y) − 1/2 A var(y) where y = income and, A = coefficient of absolute risk aversion.

Equation 2 is derived in the Appendix.

The exact value used for the Spearman correlation coefficient is –0.4. This value was used for corn prices by Hennessy et al. (1997). They based their calculation on test plot data for Iowa corn markets. Of course, this value may vary from year to year and certainly may be different across alternative crops, including wheat. This coefficient was found to be a reasonable estimate as our simulations were relatively insensitive to alternative values.