Rice Yield Distribution and Risk Assessment in South Asian Countries: A Statistical Investigation
Subject Areas : Environmental policy and managementMahadeb Prasad Poudel 1 , Shwu-En Chen 2 , Raju Ghimire 3
1 - Department of Tropical Agriculture and International Cooperation, 1, Shuefu Road, Neipu, Pingtung 912, Taiwan, R.O.C
Department of Agriculture, Ministry of Agriculture Development, Government of Nepal, Harihar Bhaban, Lalitpur, Nepal
2 - Department of Tropical Agriculture and International Cooperation, 1, Shuefu Road, Neipu, Pingtung 912, Taiwan, R.O.C
3 - Department of Agribusiness Management, National Pingtung University of Science and Technology, 1, Shuefu Road,
Neipu, Pingtung 912, Taiwan, R.O.C
Keywords: Rice, Yield Risk, Yield Distribution, Maximum Likelihood Estimation, South Asia,
Abstract :
In the last decades, rice yields in South Asian countries grew tremendously in one hand and a noticeable yield fluctuation on the other. The objective of this study was to examine the rice yield distributions, estimate yield risks at country level, and compare risks between five countries namely Afghanistan, Bangladesh, Nepal, Sri Lanka, and Pakistan. Anderson Darling (AD) test was applied to test the goodness-of-fit for four distributions by using country level de-trended rice yields from 1961 to 2010. Results showed the Normal distribution was fitted well in Afghanistan and Sri Lanka, whereas the Wei bull distribution in Bangladesh, Nepal, and Pakistan. The average yield risks at 85% of the expected yield were found 5.29, 4.27, 3.86, 1.55, and.15% in Afghanistan, Sri Lanka, Pakistan, Nepal, and Bangladesh, respectively. Wilcoxon signed rank test results of mean absolute percentage differences showed yield risk in Bangladesh was significantly lower than the rest four counties and that in Afghanistan was significantly higher than Nepal and Bangladesh at 0.1 level. The outcome of this study could give policy implications for designing and implementing the risk reducing programs in the countries with higher yield risk.