Generalized Autoregressive Score Models in R: The GAS Package

Abstract

This paper presents the R package GAS for the analysis of time series under the generalized autoregressive score (GAS) framework of Creal, Koopman, and Lucas (2013) and Harvey (2013). The distinctive feature of the GAS approach is the use of the score function as the driver of time-variation in the parameters of non-linear models. The GAS package provides functions to simulate univariate and multivariate GAS processes, to estimate the GAS parameters and to make time series forecasts. We illustrate the use of the GAS package with a detailed case study on estimating the time-varying conditional densities of financial asset returns.

Markov-Switching GARCH Models in R: The MSGARCH Package

“…The economically relevant question is then to evaluate which of the VaR forecasts are most accurate in terms of correctly predicting the α-quantile loss such that we expect to have a proportion α of exceedances. We use the R package GAS (Ardia et al 2019b) to compute the p values of two backtesting hypothesis tests of correct conditional coverage of the VaR: the conditional coverage (CC) test by Christoffersen (1998) and the dynamic quantile (DQ) test by Engle and Manganelli (2004). Both tests aim at determining if the VaR forecasts achieve correct unconditional coverage and if the violations of the VaR are independent over time.…”

“…Since then, multiple extensions of the GARCH scedastic function have been proposed to capture additional stylized facts observed in financial and economic time series, such as nonlinearities, asymmetries, and long-memory properties; see Teräsvirta (2009) for a review. According to the Time Series Analysis (Hyndman 2019) and Empirical Finance (Eddelbuettel 2019) task views at https://CRAN.R-project.org/web/views, the following implementations of univariate GARCH-type models are available in the R (R Core Team 2018) programming language: bayesGARCH (Ardia and Hoogerheide 2010), fGarch (Wuertz, Chalabi, Miklovic, Boudt, and Chausse 2016), GAS (Ardia, Boudt, and Catania 2019b), gets (Pretis, Reade, and Sucarrat 2018), GEVStableGarch (Sousa, Otiniano, Lopes, and Diethelm 2015), lgarch (Sucarrat 2015), rugarch (Ghalanos 2017) and tseries (Trapletti and Hornik 2017). In GARCHtype models, the conditional volatility is driven by shocks in the observed time series.…”

Downside Risk Evaluation with the R Package GAS

“…Dist is a character equal to the label of the distribution. For instance, SKST is identified as "sstd", ST as "std", and N as "norm"; see Table 1 of Ardia et al (2019) for the list of distributions and associated labels available in the GAS package.…”

“…This increasingly popular class of score-driven models has been introduced by Creal et al (2013) and Harvey (2013). Ardia et al (2019) describe the general functionality implemented in the GAS package, but do not cover the functionality useful for the estimation and backtesting of Value-at-Risk (VaR) and Expected Shortfall (ES), which are the two leading risk measures used in finance. The aim of this paper is to show how the functions available in the GAS package can be used for VaR and ES evaluation, prediction, and backtesting.…”

“…GAS models are also referred to as Score Driven (SD) models and Dynamic Conditional Score (DCS) models and have been used extensively for financial risk management purposes. For a comparison between the accuracy of VaR and ES estimates obtained by the GAS approach against alternative volatility models, we refer the reader to Bernardi and Catania (2016), Gao and Zhou (2016), Ardia et al (2018) and Ardia et al (2019).…”