Modeling Financial Return Dynamics via Decomposition

Anatolyev, Stanislav; Gospodinov, Nikolay

doi:10.1198/jbes.2010.07017

Cited by 54 publications

(44 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We use the simulated returns to compute the sample volatility and skewness against 200 points of α. Figure 1, the solid lines are the levels of sample moments estimated from the U.S. return data, and the dot lines are the estimated dependence parameter, α = 0.087, from the constant decomposition model of Anatolyev and Gospodinov (2010). As seen, the sample volatility and skewness request the corresponding dependence around 2.5 and 0.9, respectively, much higher than the estimated constant dependence parameter.…”

Section: Constant Decomposition Modelmentioning

confidence: 98%

“…Nonzero thresholds may be considered for e.g., transaction costs, and targeting rates, etc. Christoffersen and Diebold (2006) and Anatolyev and Gospodinov (2010) consider the leading case, c = 0. The return decomposition can be rewritten with the leading case as:…”

Section: Methodological Frameworkmentioning

confidence: 99%

“…ξ t+1|t = E t |r t+1 | I(r t+1 > 0) is the conditional expectation forecast for the cross product of absolute returns and indicators. Anatolyev and Gospodinov (2010) nd that the expectation of the cross product happens to be weakly conditional dependence. Thus, one might ignore the dependence by assuming the conditional independence between absolute returns and indicators aŝ…”

Section: Conditional Mean Forecastsmentioning

confidence: 99%

“…As the proxies for the unobserved volatility process, the realized variance is an estimator of integrated variance plus a jump component while the latter is unaected by the presence of jumps. For the detail formula, see e.g., Barndor-Nielsen and Shephard (2004,2006,2007), and Anatolyev and Gospodinov (2010), among others.…”

Section: Datamentioning

confidence: 99%

“…Christoersen and Diebold (2006) show that volatility dependence produces sign dependence to help forecast the market timing, which is not likely to be detected via traditional methods due to the special nonlinear nature of sign dependence and the dynamics between absolute returns and signs. Anatolyev and Gospodinov (2010) show promising out-of-sample forecasting results through the modeling of the constant nonlinear dependence between absolute returns and signs.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Modeling time-varying skewness via decomposition for out-of-sample forecast

Liu

2015

International Journal of Forecasting

View full text Add to dashboard Cite

This paper models time-varying skewness for nancial return dynamics. We decompose nancial returns into the product of the absolute returns and signs, so-called the intriguing decomposition. The joint distribution between the decomposed components is modeled through a copula function with marginals. Allowing the copula dependence parameter time-varying, we estimate the dynamic nonlinear dependence between absolute returns and signs, which governs timevarying skewness for out-of-sample forecast of nancial returns.The empirical results in this paper show that the proposed models with dynamic dependence obtain better gains of out-of-sample forecast, and suggest the robust strategy for a risk-averse investor in response to the market timing. This paper also explores the sources of the forecasting performance via a recently developed econometric pin-down approach. Beyond the pure statistical sense, we nd that the forecasts of time-varying skewness trace closely to NBER-dated business-cycle phases.

show abstract

Section: Constant Decomposition Modelmentioning

confidence: 98%

Section: Methodological Frameworkmentioning

confidence: 99%

Section: Conditional Mean Forecastsmentioning

confidence: 99%

Section: Datamentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Modeling time-varying skewness via decomposition for out-of-sample forecast

Liu

2015

International Journal of Forecasting

View full text Add to dashboard Cite

show abstract

Direction‐of‐Change Financial Time Series Forecasting using a Similarity‐Based Classification Model

Skabar

2013

Journal of Forecasting

View full text Add to dashboard Cite

A similarity‐based classification model is proposed whereby densities of positive and negative returns in a delay‐embedded input space are estimated from a graphical representation of the data using an eigenvector centrality measure, and subsequently combined under Bayes' theorem to predict the probability of upward/downward movements. Application to directional forecasting of the daily close price of the Dow Jones Industrial Average over a 20‐year out‐of‐sample period yields performance superior to random walk and logistic regression models, and on a par with that of multilayer perceptrons. A feature of the classifier is that it is parameter free, parameters entering the model only via the measure used to determine pairwise similarity between data points. This allows intuitions about the nature of time series to be elegantly integrated into the model. The recursive nature of eigenvector centrality makes it better able to deal with sparsely populated input spaces than conventional approaches based on density estimation. Copyright © 2013 John Wiley & Sons, Ltd.

show abstract

On the directional predictability of equity premium using machine learning techniques

Iworiso

Vrontos

2019

Journal of Forecasting

View full text Add to dashboard Cite

This paper applies a plethora of machine learning techniques to forecast the direction of the U.S. equity premium. Our techniques include benchmark binary probit models, classification and regression trees (CART), along with penalized binary probit models. Our empirical analysis reveals that the sophisticated machine learning techniques significantly outperformed the benchmark binary probit forecasting models, both statistically and economically. Overall, the discriminant analysis classifiers are ranked first among all the models tested. Specifically, the high dimensional discriminant analysis (HDDA) classifier ranks first in terms of statistical performance, while the quadratic discriminant analysis (QDA) classifier ranks first in economic performance. The penalized likelihood binary probit models (Least Absolute Shrinkage and Selection Operator, Ridge, Elastic Net) also outperformed the benchmark binary probit models, providing significant alternatives to portfolio managers.

show abstract

Modeling Financial Return Dynamics via Decomposition

Cited by 54 publications

References 38 publications

Modeling time-varying skewness via decomposition for out-of-sample forecast

Modeling time-varying skewness via decomposition for out-of-sample forecast

Direction‐of‐Change Financial Time Series Forecasting using a Similarity‐Based Classification Model

On the directional predictability of equity premium using machine learning techniques

Contact Info

Product

Resources

About