Page 227 - Contributed Paper Session (CPS) - Volume 3

CPS2002 Atina A. et al.

The impact of weather risk on the estimation of
yield-based agricultural losses and value at risk
using Copula Models
1
1
1,2
1
Atina Ahdika , Dedi Rosadi , Adhitya Ronnie Effendhie , Gunardi
1 Department of Mathematics, Universitas Gadjah Mada, Yogyakarta, Indonesia
2 Department of Statistics, Universitas Islam Indonesia, Yogyakarta, Indonesia

Abstract
Weather risk, such as temperature change, is one of the main factors affecting
agricultural products. Temperature change can significantly affect the
occurrence of agricultural losses which can be measured from the agricultural
yield. An agricultural loss is defined as the difference value between the
estimated and the actual yield at some confidence levels. This paper aims to
identify the dependency structure between temperature change and
agricultural yield using copula functions. The estimation procedure of yield-
based agricultural losses is conducted by simulating joint occurrence between
the two variables and the selected copula parameters. The result shows that
the agricultural losses happened mostly when the temperature is low. Value
at risk in the form of yield-based agricultural losses is also estimated based on
the distribution of the estimated losses.

Keywords
agricultural losses; agricultural yield; copula; temperature change; value at risk

1. Introduction
Indonesia is one of the developing countries whose main livelihood is
farming. Farmers are very susceptible to losses such as crop failure or a
decrease in the price index of agricultural production which can be caused by
weather risk or disease attacks. Many studies have been conducted to estimate
agricultural losses based on the factors that influence them. Vergara et al.
(2008) modelled the impact of catastrophic weather on crop insurance losses.
Dahal & Routray (2011) identified the association between soil variables and
agricultural yield using multiple linear regression. Sellam & Poovammal (2016)
predicted agricultural yield by analysing the relationship between
environmental parameters such as harvest area, annual rainfall, and food price
index using linear regression. Luminto & Harlili (2017) built a weather analysis
to predict rice cultivation time to increase farmers exchange rate using linear
regression.
Along with the advance research in the field of correlation, the relationship
between the variables that affect the risk of agricultural losses is assumed to
not always be linear so that the prediction model based on the Pearson
correlation coefficient, such as a linear regression model, is no longer relevant
216 | I S I W S C 2 0 1 9