TRAP@NCI

Historical Simulation Value at Risk and Expected Tail Loss: A test of reliability in a modern financial climate

McGrath, Brendan (2016) Historical Simulation Value at Risk and Expected Tail Loss: A test of reliability in a modern financial climate. Masters thesis, Dublin, National College of Ireland.

[img]
Preview
PDF (Master of Science)
Download (906kB) | Preview

Abstract

In 2016, the Basel Committee for Banking Supervision intends to implement mandatory changes to the way in which financial institutions and banks measure the risk associated with capital requirements. This shift will move the focus from Value at Risk models to Expected Tail Loss models. This report set out to determine if this shift is warranted. To do this, Value at Risk and Expected Tail Loss models were created using the Historical Simulation methodology at both 95% and 99% confidence levels. The Expected Tail Loss model was constructed as an extension of the VaR model. These models were then divided into smaller models based on individual years.

All models were then back-tested using simple hypothesis tests in order to establish their reliability. It was found that VaR models are not reliable in periods of market volatility. Expected Tail Loss however was found to be reliable at both the 95% and 99% confidence levels.

These results would seem to support the shift from VaR to Expected Tail Loss to some extent although considering the ETL model is built from the Value at Risk model, it may be more beneficial to use both tools instead of one individual.

Item Type: Thesis (Masters)
Subjects: H Social Sciences > HD Industries. Land use. Labor > HD61 Risk Management
H Social Sciences > HG Finance > Financial Management
Divisions: School of Business > Master of Science in Finance
Depositing User: CAOIMHE NI MHAICIN
Date Deposited: 02 Nov 2016 10:48
Last Modified: 02 Nov 2016 10:48
URI: http://trap.ncirl.ie/id/eprint/2287

Actions (login required)

View Item View Item