Genaro Sucarrat

International Journal of Forecasting 37(4) Doi: https://doi.org/10.1016/j.ijforecast.2021.03.008

Volatility proxies like realised volatility (RV) are extensively used to assess the forecasts of squared financial returns produced by volatility models. But are volatility proxies identified as expectations of the squared return? If not, then the results of these comparisons can be misleading, even if the proxy is unbiased. Here, a tripartite distinction is introduced between strong, semi-strong, and weak identification of a volatility proxy as an expectation of the squared return. The definition implies that semi-strong and weak identification can be studied and corrected for via a multiplicative transformation. Well-known tests can be used to check for identification and bias, and Monte Carlo simulations show that they are well sized and powerful—even in fairly small samples. As an illustration, 12 volatility proxies used in three seminal studies are revisited. Half of the proxies do not satisfy either semi-strong or weak identification, but their corrected transformations do. It is then shown how correcting for identification can change the rankings of volatility forecasts.

The R Journal 13(1) p. 276-291 Doi: https://doi.org/10.32614/rj-2021-057

The garchx package provides a user-friendly, fast, flexible, and robust framework for the estimation and inference of GARCH(p, q, r)-X models, where p is the ARCH order, q is the GARCH order, r is the asymmetry or leverage order, and ’X’ indicates that covariates can be included. Quasi Maximum Likelihood (QML) methods ensure estimates are consistent and standard errors valid, even when the standardized innovations are non-normal or dependent, or both. Zero-coefficient restrictions by omission enable parsimonious specifications, and functions to facilitate the non-standard inference associated with zero-restrictions in the null-hypothesis are provided. Finally, in the formal comparisons of precision and speed, the garchx package performs well relative to other prominent GARCH-packages on CRAN.

Samfunnsøkonomen 134(3) p. 13-20

Årlige prognoser av norsk økonomi er av stor viktighet for beslutningstakere. Dette gjelder spesielt stortingspolitikerne som vedtar Statsbudsjettet basert på prognosene i Nasjonalbudsjettet. Disse prognosene utarbeides av Finansdepartementet. I dette studiet evaluerer vi presisjonen til et utvalg prognoser i perioden 1999-2018. Vi finner ingen generell støtte for hypotesen om at prognosene til enkle modeller er mer presise enn de til nasjonalbudsjettet. Videre finner vi at Nasjonalbudsjettets prognoser er: Generelt litt mer presise enn de til enkle modeller, på nivå med prognosene til Norges Bank og SSB, og at de i gjennomsnitt treffer det de sikter på.

The R Journal 12(2) p. 388-401 Doi: https://doi.org/10.32614/rj-2021-024

Abstract General-to-Specific (GETS) modelling provides a comprehensive, systematic and cumulative approach to modelling that is ideally suited for conditional forecasting and counterfactual analysis, whereas Indicator Saturation (ISAT) is a powerful and flexible approach to the detection and estimation of structural breaks (e.g. changes in parameters), and to the detection of outliers. To these ends, multi path backwards elimination, single and multiple hypothesis tests on the coefficients, diagnostics tests and goodness-of-fit measures are combined to produce a parsimonious final model. In many situations a specific model or estimator is needed, a specific set of diagnostics tests may be required, or a specific fit criterion is preferred. In these situations, if the combination of estimator/model, diagnostics tests and fit criterion is not offered in a pre-programmed way by publicly available software, then the implementation of user-specified GETS and ISAT methods puts a large programming-burden on the user. Generic functions and procedures that facilitate the implementation of user-specified GETS and ISAT methods for specific problems can therefore be of great benefit. The R package gets is the first software – both inside and outside the R universe – to provide a complete set of facilities for user-specified GETS and ISAT methods: User-specified model/estimator, user-specified diagnostics and user-specified goodness-of-fit criteria. The aim of this article is to illustrate how user-specified GETS and ISAT methods can be implemented with the R package gets.

Journal of Financial Econometrics 20(2) p. 278-309 Doi: https://doi.org/10.1093/jjfinec/nbaa014

The probability of an observed financial return being equal to zero is not necessarily zero, or constant. In ordinary models of financial return, however, e.g. ARCH, SV, GAS and continuous-time models, the zero-probability is zero, constant or both, thus frequently resulting in biased risk estimates (volatility, Value-at-Risk, Expected Shortfall, etc.). We propose a new class of models that allows for a time varying zero-probability that can either be stationary or non-stationary. The new class is the natural generalisation of ordinary models of financial return, so ordinary models are nested and obtained as special cases. The main properties (e.g. volatility, skewness, kurtosis, Value-at-Risk, Expected Shortfall) of the new model class are derived as functions of the assumed volatility and zero-probability specifications, and estimation methods are proposed and illustrated. In a comprehensive study of the stocks at New York Stock Exchange (NYSE) we find extensive evidence of time varying zero-probabilities in daily returns, and an out-of-sample experiment shows that corrected risk estimates can provide significantly better forecasts in a large number of instances.

Financial mathematics, volatility and covariance modelling p. 336-359 Doi: https://doi.org/10.4324/9781315162737-14

Journal of Statistical Software 86(3) p. 1-44 Doi: https://doi.org/10.18637/jss.v086.i03

This paper provides an overview of the R package gets, which contains facilities for automated general-to-specific (GETS) modeling of the mean and variance of a regression, and indicator saturation (IS) methods for the detection and modeling of outliers and structural breaks. The mean can be specified as an autoregressive model with covariates (an "AR-X" model), and the variance can be specified as an autoregressive log-variance model with covariates (a "log-ARCH-X" model). The covariates in the two specifications need not be the same, and the classical linear regression model is obtained as a special case when there is no dynamics, and when there are no covariates in the variance equation. The four main functions of the package are arx, getsm, getsv and isat. The first function estimates an AR-X model with log-ARCH-X errors. The second function undertakes GETS modeling of the mean specification of an 'arx' object. The third function undertakes GETS modeling of the log-variance specification of an 'arx' object. The fourth function undertakes GETS modeling of an indicator-saturated mean specification allowing for the detection of outliers and structural breaks. The usage of two convenience functions for export of results to EViews and Stata are illustrated, and LATEX code of the estimation output can readily be generated.

Energy Economics 74(August) p. 287-298 Doi: https://doi.org/10.1016/j.eneco.2018.05.017

Electricity prices are characterised by strong autoregressive persistence, periodicity (e.g. intraday, day-of-the week and month-of-the-year effects), large spikes or jumps, GARCH and – as evidenced by recent findings – periodic volatility. We propose a multivariate model of volatility that decomposes volatility multiplicatively into a non-stationary (e.g. periodic) part and a stationary part with log-GARCH dynamics. Since the model belongs to the log-GARCH class, the model is robust to spikes or jumps, allows for a rich variety of volatility dynamics without restrictive positivity constraints, can be estimated equation-by-equation by means of standard methods even in the presence of feedback, and allows for Dynamic Conditional Correlations (DCCs) that can – optionally – be estimated subsequent to the volatilities. We use the model to study the hourly day-ahead system prices at Nord Pool, and find extensive evidence of periodic volatility and volatility feedback. We also find that volatility is characterised by (positive) leverage in one third of the hours, and that a DCC model provides a better fit of the conditional correlations than a Constant Conditional Correlation (CCC) model.

Journal of Multivariate Analysis 153 p. 16-32 Doi: https://doi.org/10.1016/j.jmva.2016.09.010

Journal of Financial Econometrics 16(1) p. 129-154 Doi: https://doi.org/10.1093/jjfinec/nbx032

European Journal of Finance 24(10) p. 809-827 Doi: https://doi.org/10.1080/1351847X.2017.1336452

European Journal of Finance 21(13-14) p. 1195-1213 Doi: https://doi.org/10.1080/1351847X.2012.706906

We propose and study simple but flexible methods for density selection of skewed versions of the two most popular density classes in finance, the exponential power distribution and the t distribution. For the first type of method, which simply consists of selecting a density by means of an information criterion, the Schwarz criterion stands out since it performs well across density categories, and in particular when the Data Generating Process is normal. For the second type of method, General-to-Specific density selection, the simulations suggest that it can improve the recovery rate in predictable ways by changing the significance level. This is useful because it enables us to increase (reduce) the recovery rate of non-normal densities by increasing (reducing) the significance level, if one wishes to do so. The third type of method is a generalisation of the second type, such that it can be applied across an arbitrary number of density classes, nested or non-nested. Finally, the methods are illustrated in an empirical application.

Computational Statistics & Data Analysis 100 p. 582-594 Doi: https://doi.org/10.1016/j.csda.2015.12.005

Computational Statistics & Data Analysis 76 p. 320-328 Doi: https://doi.org/10.1016/j.csda.2013.09.022

An EGARCH model in which the conditional distribution is heavy- tailed and skewed is proposed. The properties of the model, including unconditional moments, autocorrelations and the asymptotic distribu- tion of the maximum likelihood estimator, are obtained. Evidence for skewness in conditional t-distribution is found for a range of returns series and the model is shown to give a better t than the correspond- ing skewed-t GARCH model.

The R Journal 5(2) p. 138-148 Doi: https://doi.org/10.32614/rj-2013-034

Oxford Bulletin of Economics and Statistics 74(5) p. 716-735 Doi: https://doi.org/10.1111/j.1468-0084.2011.00669.x

Mathematical and statistical methods for actuarial sciences and finance p. 279-286 Doi: https://doi.org/10.1007/978-88-470-2342-0_33

Journal of Economic Methodology 17(1) p. 53-75

International Journal of Forecasting 26(4) p. 885-907

Economics 3(8) p. 1-33

High-Frequency Financial Econometrics p. 7-29

Empirical Economics 30 p. 889-911 Doi: https://doi.org/10.1007/s00181-005-0005-x