Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Measuring Risk for WTI Crude Oil: An application of Parametric Expected Shortfall

Eriksson, Alexander LU (2015) NEKP01 20151
Department of Economics
Abstract (Swedish)
Oil is the most traded commodity in the world and is an important part in the global economy. The change in the price of oil has an effect on all sectors of the economy, and the ability to capture its risk is an important research topic. This study calculates the risk of one benchmark crude oil (West Texas Intermediate) over the period 1986-2015 by estimating the Value-at-Risk (VaR) and the Expected Shortfall (ES) on daily spot returns. More specifically, this is done by using a GARCH (1, 1) model with the normal distribution, the t-distribution, and the Generalized Error Distribution (GED). The study uses a rolling window to estimate these risk measurements creating 7125 estimates for each distribution in each tail. The normal... (More)
Oil is the most traded commodity in the world and is an important part in the global economy. The change in the price of oil has an effect on all sectors of the economy, and the ability to capture its risk is an important research topic. This study calculates the risk of one benchmark crude oil (West Texas Intermediate) over the period 1986-2015 by estimating the Value-at-Risk (VaR) and the Expected Shortfall (ES) on daily spot returns. More specifically, this is done by using a GARCH (1, 1) model with the normal distribution, the t-distribution, and the Generalized Error Distribution (GED). The study uses a rolling window to estimate these risk measurements creating 7125 estimates for each distribution in each tail. The normal distribution was the worst performing distribution on both ES and VaR according to the backtests. The t-distribution performed good ES estimates; however it was not as accurate when calculating VaR. The GED performed the best when calculating VaR but constantly underestimated ES. The main conclusion is that both GED and the t-distribution are needed when estimating the risk for WTI. (Less)
Please use this url to cite or link to this publication:
author
Eriksson, Alexander LU
supervisor
organization
course
NEKP01 20151
year
type
H2 - Master's Degree (Two Years)
subject
keywords
Risk Expected shortfall VaR Backtesting Crude oil
language
English
id
7791254
date added to LUP
2015-09-14 14:27:26
date last changed
2015-09-14 14:27:26
@misc{7791254,
  abstract     = {{Oil is the most traded commodity in the world and is an important part in the global economy. The change in the price of oil has an effect on all sectors of the economy, and the ability to capture its risk is an important research topic. This study calculates the risk of one benchmark crude oil (West Texas Intermediate) over the period 1986-2015 by estimating the Value-at-Risk (VaR) and the Expected Shortfall (ES) on daily spot returns. More specifically, this is done by using a GARCH (1, 1) model with the normal distribution, the t-distribution, and the Generalized Error Distribution (GED). The study uses a rolling window to estimate these risk measurements creating 7125 estimates for each distribution in each tail. The normal distribution was the worst performing distribution on both ES and VaR according to the backtests. The t-distribution performed good ES estimates; however it was not as accurate when calculating VaR. The GED performed the best when calculating VaR but constantly underestimated ES. The main conclusion is that both GED and the t-distribution are needed when estimating the risk for WTI.}},
  author       = {{Eriksson, Alexander}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Measuring Risk for WTI Crude Oil: An application of Parametric Expected Shortfall}},
  year         = {{2015}},
}