The aim of this final report is to present and expand on the topics covered in Le’s (2023) "Forecasting VaR and ES in emerging markets: The role of time-varying higher moments". This paper discusses the role of time-varying higher moments in conditional volatility models in emerging markets. In particular, an analysis of Value at Risk (VaR) and Expected Shortfall (ES) forecasts from a sample of eight generalized autoregressive conditional heteroskedasticity (GARCH) models for 10 emerging markets is performed. These are evaluated through backtests on absolute performance and loss functions on relative performance. The inclusion of time-varying skewness and kurtosis leads to better forecasts, mainly during crisis periods, where traditional model specifications often underestimate the tail risk in markets.
Lo scopo di questa relazione finale è presentare ed approfondire gli argomenti trattati dall'articolo di Le (2023) "Forecasting VaR and ES in emerging markets: The role of time-varying higher moments". Questo paper tratta del ruolo dei momenti dinamici di ordine superiore nei modelli di volatilità condizionale nei mercati emergenti. In particolare, viene effettuata un'analisi delle previsioni di Value at Risk (VaR) e Expected Shortfall (ES) da un campione di otto modelli GARCH (generalized autoregressive conditional heteroskedasticity) per 10 mercati emergenti. Queste vengono valutate attraverso backtest sulle prestazioni assolute e funzioni di perdita sulla performance relativa. L'inclusione di asimmetria e curtosi dinamiche porta a previsioni migliori, principalmente nei periodi di crisi dove le specifiche dei modelli tradizionali spesso sottostimano il rischio di coda nei mercati.
Asimmetria e curtosi dinamici nella previsione di VaR e ES: il caso dei mercati emergenti
GUZZA, DAVIDE
2023/2024
Abstract
The aim of this final report is to present and expand on the topics covered in Le’s (2023) "Forecasting VaR and ES in emerging markets: The role of time-varying higher moments". This paper discusses the role of time-varying higher moments in conditional volatility models in emerging markets. In particular, an analysis of Value at Risk (VaR) and Expected Shortfall (ES) forecasts from a sample of eight generalized autoregressive conditional heteroskedasticity (GARCH) models for 10 emerging markets is performed. These are evaluated through backtests on absolute performance and loss functions on relative performance. The inclusion of time-varying skewness and kurtosis leads to better forecasts, mainly during crisis periods, where traditional model specifications often underestimate the tail risk in markets.File | Dimensione | Formato | |
---|---|---|---|
Guzza_Davide.pdf
accesso aperto
Dimensione
985.32 kB
Formato
Adobe PDF
|
985.32 kB | Adobe PDF | Visualizza/Apri |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/77681