Modeling of soybean yield using symmetric, asymmetric and bimodal distributions: implications for crop insurance

作者：Duarte Gislaine V.^*; Braga Altemir; Miquelluti Daniel L.; Ozaki Vitor A.

来源：Journal of Applied Statistics, 2018, 45(11): 1920-1937.

DOI：10.1080/02664763.2017.1406902

• Asymmetry
• bimodality distribution
• crop insurance
• heavier tails
• risk management

Over the years, many papers used parametric distributions to model crop yields, such as: normal (N), Beta, Log-normal and the Skew-normal (SN). These models are well-defined, mathematically and also computationally, but its do not incorporate bimodality. Therefore, it is necessary to study distributions which are more flexible in modeling, since most of crop yield data in Brazil presents evidence of asymmetry or bimodality. Thus, the aim of this study was to model and forecast soybean yields for municipalities in the State of Paran, in the period from 1980 to 2014, using the Odd log normal logistic (OLLN) distribution for the bimodal data and the Beta, SN and Skew-t distributions for the symmetrical and asymmetrical series. The OLLN model was the one which best fit the data. The results were discussed in the context of crop insurance pricing.

• 出版日期2018