Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Univariate GARCH Models for Forecasting Real Estate Volatility and Risk Prediction

Martin, Maximilian LU and Cagetti, Dario LU (2023) NEKN02 20231
Department of Economics
Abstract
The “Fundamental Review of the Trading Book” calculates the capital requirements using Expected Shortfall (ES) backtests since Value at risk (VaR) neglects losses in the tail. Additionally, real estate investments have increased significantly, which warrants an exploration of the asset class. This study applies eight different univariate GARCH models to six distinct property indices; Global, Asia Pacific, Europe, Latin America, Mid-East & Africa, and North America using three distributions, namely skewed normal distribution (snorm), skewed Student’s t-distribution (sstd) and skewed generalized error distribution (sged). Subsequently, the variance is forecasted one day ahead, which is used to derive VaR and ES. This approach is repeated at... (More)
The “Fundamental Review of the Trading Book” calculates the capital requirements using Expected Shortfall (ES) backtests since Value at risk (VaR) neglects losses in the tail. Additionally, real estate investments have increased significantly, which warrants an exploration of the asset class. This study applies eight different univariate GARCH models to six distinct property indices; Global, Asia Pacific, Europe, Latin America, Mid-East & Africa, and North America using three distributions, namely skewed normal distribution (snorm), skewed Student’s t-distribution (sstd) and skewed generalized error distribution (sged). Subsequently, the variance is forecasted one day ahead, which is used to derive VaR and ES. This approach is repeated at two confidence levels and the resulting models are evaluated by backtesting procedures. Applying asymmetric conditional volatility models improves forecast accuracy in most cases and it emerges that the assumed distribution influences the results heavily. Specific recommendations are dependent on index makeup and its return structure. Estimating and backtesting ES adds value to this paper, as this is the requirement in new Basel regulations. (Less)
Please use this url to cite or link to this publication:
author
Martin, Maximilian LU and Cagetti, Dario LU
supervisor
organization
course
NEKN02 20231
year
type
H1 - Master's Degree (One Year)
subject
keywords
Real estate, GARCH, Expected Shortfall, Value at Risk, Backtesting
language
English
id
9123441
date added to LUP
2023-11-24 08:57:01
date last changed
2023-11-24 08:57:01
@misc{9123441,
  abstract     = {{The “Fundamental Review of the Trading Book” calculates the capital requirements using Expected Shortfall (ES) backtests since Value at risk (VaR) neglects losses in the tail. Additionally, real estate investments have increased significantly, which warrants an exploration of the asset class. This study applies eight different univariate GARCH models to six distinct property indices; Global, Asia Pacific, Europe, Latin America, Mid-East & Africa, and North America using three distributions, namely skewed normal distribution (snorm), skewed Student’s t-distribution (sstd) and skewed generalized error distribution (sged). Subsequently, the variance is forecasted one day ahead, which is used to derive VaR and ES. This approach is repeated at two confidence levels and the resulting models are evaluated by backtesting procedures. Applying asymmetric conditional volatility models improves forecast accuracy in most cases and it emerges that the assumed distribution influences the results heavily. Specific recommendations are dependent on index makeup and its return structure. Estimating and backtesting ES adds value to this paper, as this is the requirement in new Basel regulations.}},
  author       = {{Martin, Maximilian and Cagetti, Dario}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Univariate GARCH Models for Forecasting Real Estate Volatility and Risk Prediction}},
  year         = {{2023}},
}