skip to content

Keynes Fund

 

Prof. Andrew Harvey - Dynamic Models for volatility and heavy tails (JHLC) and (JHLH)

Published Papers

EGARCH models with fat tails, skewness and leverage, Harvey, A.C. and G. Sucarrat (2014) Computational Statistics and Data Analysis, Vol 26, pp. 320-338 

Abstract: 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 distribution of the maximum likelihood estimator, are set out. Evidence for skewness in a conditional t-distribution is found for a range of returns series, and the model is shown to give a better fit than comparable skewed-t GARCH models in nearly all cases. A two-component model gives further gains in goodness of fit and is able to mimic the long memory pattern displayed in the autocorrelations of the absolute values.

[full text link]


Filtering with Heavy Tails, Harvey, A.C. and A. Luati. (2014) Journal of the American Statistical Association, Vol 109, pp. 1112-1122.

Abstract: An unobserved components model in which the signal is buried in noise that is non-Gaussian may throw up observations that, when judged by the Gaussian yardstick, are outliers. We describe an observation-driven model, based on a conditional Student’s t-distribution, which is tractable and retains some of the desirable features of the linear Gaussian model. Letting the dynamics be driven by the score of the conditional distribution leads to a specification that is not only easy to implement, but which also facilitates the development of a comprehensive and relatively straightforward theory for the asymptotic distribution of the maximum likelihood estimator. The methods are illustrated with an application to rail travel in the United Kingdom. The final part of the article shows how the model may be extended to include explanatory variables.

[full text link]


Time series models with an EGB2 conditional distribution, M. Caivano and Andrew Harvey (2014), Journal of Time Series Analysis, Vol 35(6), pp. 558-571

Abstract: A time-series model in which the signal is buried in noise that is non-Gaussian may throw up observations that, when judged by the Gaussian yardstick, are outliers. We describe an observation-driven model, based on an exponential generalized beta distribution of the second kind (EGB2), in which the signal is a linear function of past values of the score of the conditional distribution. This specification produces a model that is not only easy to implement but which also facilitates the development of a comprehensive and relatively straightforward theory for the asymptotic distribution of the maximum-likelihood (ML) estimator. Score-driven models of this kind can also be based on conditional t distributions, but whereas these models carry out what, in the robustness literature, is called a soft form of trimming, the EGB2 distribution leads to a soft form of Winsorizing. An exponential general autoregressive conditional heteroscedastic (EGARCH) model based on the EGB2 distribution is also developed. This model complements the score-driven EGARCH model with a conditional t distribution. Finally, dynamic location and scale models are combined and applied to data on the UK rate of inflation.

[full text link]


Computation of Maximum Likelihood Estimates for Score Driven Models for Positive Valued Observations, Philipp Andres, (2014), Computational Statistics and Data Analysis, Vol 76, pp. 34-42.

Abstract: Recently, the Dynamic Conditional Score (DCS) or Generalized Autoregressive Score (GAS) time series models have attracted considerable attention. This motivates the need for a software package to estimate and evaluate these new models. A straightforward to operate program called the Dynamic Score (DySco) package is introduced for estimating models for positive variables, in which the location/scale evolves over time. Its capabilities are demonstrated using a financial application.

[full text link]


Testing against changing correlation, Harvey, A.C. and S. Thiele (2016), Journal of Empirical Finance, Vol 38, pp. 575-89.  

Abstract: A test for time-varying correlation is developed within the framework of a dynamic conditional score (DCS) model for both Gaussian and Student t-distributions. The test may be interpreted as a Lagrange multiplier test and modified to allow for the estimation of models for time-varying volatility in the individual series. Unlike standard moment-based tests, the score-based test statistic includes information on the level of correlation under the null hypothesis and local power arguments indicate the benefits of doing so. A simulation study shows that the performance of the score-based test is strong relative to existing tests across a range of data generating processes. An application to the Hong Kong and South Korean equity markets shows that the new test reveals changes in correlation that are not detected by the standard moment-based test.

[full text link]


Robust time series models with trend and seasonal components, Michele Caivano, Andrew Harvey and Alessandra Luati (2016), Journal of the Spanish Economic Association, Vol 7(1), pp. 99-120

Abstract: We describe observation driven time series models for Student-t and EGB2 conditional distributions in which the signal is a linear function of past values of the score of the conditional distribution. These specifications produce models that are easy to implement and deal with outliers by what amounts to a soft form of trimming in the case of t and a soft form of Winsorizing in the case of EGB2. We show how a model with trend and seasonal components can be used as the basis for a seasonal adjustment procedure. The methods are illustrated with US and Spanish data.

[full text link]


Volatility Modelling with a Generalized t-distribution, Harvey, A.C. and R-J. Lange (2017), Journal of Time Series Analysis (forthcoming).

Abstract: Exponential generalized autoregressive conditional heteroscedasticity models in which the dynamics of the logarithm of scale are driven by the conditional score are known to exhibit attractive theoretical properties for the t distribution and general error distribution. A model based on the generalized t includes both as special cases. We derive the information matrix for the generalized t and show that, when parameterized with the inverse of the tail index, it remains positive definite in the limit as the distribution goes to a general error distribution. We generalize further by allowing the distribution of the observations to be skewed and asymmetric. Our method for introducing asymmetry ensures that the information matrix reverts to the usual case under symmetry. We are able to derive analytic expressions for the conditional moments of our exponential generalized autoregressive conditional heteroscedasticity model as well as the information matrix of the dynamic parameters. The practical value of the model is illustrated with commodity and stock return data. Overall, the approach offers a unified, flexible, robust, and effective treatment of volatility.

[full text link]


Working Papers

Modeling the Interactions between Volatility and Returns, Harvey, A.C. and R-J. Lange (2015) CWPE 1518

Abstract: Volatility of a stock may incur a risk premium, leading to a pos- itive correlation between volatility and returns. On the other hand the leverage effect, whereby negative returns increase volatility, acts in the opposite direction. We propose a reformulation and extension of the ARCH in Mean model, in which the logarithm of scale is driven by the score of the conditional distribution. This EGARCH-M model is shown to be theoretically tractable as well as practically useful. By employing a two component extension we are able to distinguish between the long and short run effects of returns on volatility. The EGARCH formulation allows more flexibility in the asymmetry of the response (leverage) and this enables us to find that the short-term response is, in some cases, close to being anti-asymmetric. The long and short run volatility components are shown to have very different effects on returns, with the long-run component yielding the risk pre- mium. A model in which the returns have a skewed t distribution is shown to fit well to daily and weekly data on some of the major stock market indices.

[download pdf] PDF logo


Modeling Time Series with Zero Observations, Harvey, A.C. and R. Ito (2017), Nuffield College Economics Working Paper 2017-W01, Oxford University

Abstract: We consider situations in which a significant proportion of obser- vations in a time series are zero, but the remaining observations are positive and measured on a continuous scale. We propose a new dy- namic model in which the conditional distribution of the observations is constructed by shifting a distribution for non-zero observations to the left and censoring negative values. The key to generalizing the censoring approach to the dynamic case is to have (the logarithm of) the location/scale parameter driven by a filter that depends on the score of the conditional distribution.  An exponential link function means that seasonal effects can be incorporated into the model and this is done by means of a cubic spline (which can potentially be time- varying). The model is fitted to daily rainfall in northern Australia and compared with a dynamic zero-augmented model.

[download pdf] PDF logo