Search

Change-point (CP) VAR models face a dimensionality curse due to the proliferation of parameters that arises when new breaks are detected. We introduce the Sparse CP-VAR model which determines which parameters truly vary when a break is detected. By doing so, the number of new parameters to be estimated at each regime is drastically reduced and the break dynamics becomes easier to be interpreted. The Sparse CP-VAR model disentangles the dynamics of the mean parameters and the covariance matrix. The former uses CP dynamics with shrinkage prior distributions, while the latter is driven by an infinite hidden Markov framework. An extensive simulation study is carried out to compare our approach with existing ones. We provide applications to financial and macroeconomic systems. It turns out that many off-diagonal VAR parameters are zero for the entire sample period and that most break activity is in the covariance matrix. We show that this has important consequences for portfolio optimization, in particular when future instabilities are included in the predictive densities. Forecasting-wise, the Sparse CP-VAR model compares favorably to several time-varying parameter models in terms of density and point forecast metrics.

Change-point (CP) processes are one flexible approach to model long time series. We propose a method to uncover which model parameters truly vary when a CP is detected. Given a set of breakpoints, we use a penalized likelihood approach to select the best set of parameters that changes over time and we prove that the penalty function leads to a consistent selection of the true model. Estimation is carried out via the deterministic annealing expectation-maximization algorithm. Our method accounts for model selection uncertainty and associates a probability to all the possible time-varying parameter specifications. Monte Carlo simulations highlight that the method works well for many time series models including heteroskedastic processes. For a sample of fourteen hedge fund (HF) strategies, using an asset-based style pricing model, we shed light on the promising ability of our method to detect the time-varying dynamics of risk exposures as well as to forecast HF returns.

We propose a new volatility process in which parameters vary over time according to an artificial neural network (ANN). We prove the process’s stationarity as well as the global identification of the parameters. Since ANNs require economic series as input variables, we develop a shrinkage approach to select which explanatory variables are relevant to forecast volatility. Empirically, the proposed model favorably compares with other flexible processes in terms of in-sample fit on six financial returns. It also delivers accurate short-term volatility predictions in terms of root mean squared errors and the predictive likelihood criterion. For long-term forecasts, it can be competitive with the Markov-switching generalized autoregressive conditional heteroskedastic (MS-GARCH) model if appropriate exogenous variables are used. Since our new type of time-varying parameter (TVP) process is based on a universal approximator, the approach can readily revisit and potentially improve many standard TVP applications.

Structural break time series models, which are commonly used in macroeconomics and finance, capture unknown structural changes by allowing for abrupt changes to model parameters. However, many specifications suffer from an over-parametrization issue, since typically all parameters have to change when a break occurs. We introduce a sparse change-point model to detect which parameters change over time. We propose a shrinkage prior distribution, which controls model parsimony by limiting the number of parameters that change from one structural break to another. We develop a Bayesian sampler for inference on the sparse change-point model. An extensive simulation study based on AR, ARMA and GARCH processes highlights the excellent performance of the sampler. We provide several empirical applications including an out-of-sample forecasting exercise showing that the Sparse change-point framework compares favorably with other recent time-varying parameter processes.

This chapter evaluates Bayesian inference, which refers to the Bayesian statistical method for estimating the parameters of a model and for testing a hypothesis. It relies on subjective statistics and extensively uses Bayes’s theorem. In the early 1990s, Bayesian statistics boomed with the emergence of sampling techniques. These new tools rely on the computational power to sample from (rather than evaluate) the posterior probability. However, the main drawback of the Bayesian approach lies in the computation of the posterior probability. The analytical computation of the posterior probability is a complex problem for any application, and this has limited Bayesian statistics for years.

We construct measures of uncertainty and its dispersion exploiting the heterogeneity of a large set of model predictions. The approach is forward-looking, can be computed in real-time, and can be applied at any frequency. We illustrate the methodology with expected shortfall predictions of worldwide equity indices generated from 71 risk models. We use the new measures in asset pricing, risk forecasting, and for explaining the aggregate trading volume of S&P 500 firms.

A new process—the factorial hidden Markov volatility (FHMV) model—is proposed to model financial returns or realized variances. Its dynamics are driven by a latent volatility process specified as a product of three components: a Markov chain controlling volatility persistence, an independent discrete process capable of generating jumps in the volatility, and a predictable (data-driven) process capturing the leverage effect. An economic interpretation is attached to each one of these components. Moreover, the Markov chain and jump components allow volatility to switch abruptly between thousands of states, and the transition matrix of the model is structured to generate a high degree of volatility persistence. An empirical study on six financial time series shows that the FHMV process compares favorably to state-of-the-art volatility models in terms of in-sample fit and out-of-sample forecasting performance over time horizons ranging from 1 to 100 days. Supplementary materials for this article are available online.

Change-point time series specifications constitute flexible models that capture unknown structural changes by allowing for switches in the model parameters. Nevertheless most models suffer from an ...

Change-point processes are one flexible approach to model long time series. We propose a method to uncover which model parameter truly vary when a change-point is detected. Given a set of breakpoints, we use a penalized likelihood approach to select the best set of parameters that changes over time and we prove that the penalty function leads to a consistent selection of the true model. Estimation is carried out via the deterministic annealing expectation-maximization algorithm. Our method accounts for model selection uncertainty and associates a probability to all the possible time-varying parameter specifications. Monte Carlo simulations highlight that the method works well for many time series models including heteroskedastic processes. For a sample of 14 Hedge funds (HF) strategies, using an asset based style pricing model, we shed light on the promising ability of our method to detect the time-varying dynamics of risk exposures as well as to forecast HF returns.

Change-point (CP) VAR models face a dimensionality curse due to the proliferation of parameters that arises when new breaks are detected. To handle large data sets, we introduce the Sparse CP-VAR model that determines which parameters truly vary when a break is detected. By doing so, the number of new parameters to estimate at each regime is drastically reduced and the CP dynamic becomes easier to interpret. The Sparse CP-VAR model disentangles the dynamics of the mean parameters and the covariance matrix. The former uses CP dynamics with shrinkage prior distributions while the latter is driven by an infinite hidden Markov framework. A simulation study highlights that the framework detects correctly the number of breaks per model parameter, and that it takes advantage of common breaks in the cross-sectional dimension to more precisely estimate them. Our applications on financial and macroeconomic systems highlight that the Sparse CP-VAR model helps interpreting the detected breaks. It turns out that many spillover effects have zero regimes meaning that they are zero for the entire sample period. Forecasting wise, the Sparse CP-VAR model is competitive against several recent time-varying parameter and CP-VAR models in terms of log predictive densities.

Generalized auto-regressive conditional heteroskedastic model with changing parameters is specified using the sticky infinite hidden Markov-chain framework. Estimation by Bayesian inference determines the adequate number of regimes as well as the optimal specification (Markov-switching or change-point). The new provided algorithms are studied in terms of mixing properties and computational time. Applications highlight the flexibility of the model and compare it to existing parametric alternatives.

We present an estimation and forecasting method, based on a differential evolution MCMC method, for inference in GARCH models subjected to an unknown number of structural breaks at unknown dates. We treat break dates as parameters and determine the number of breaks by computing the marginal likelihoods of competing models. We allow for both recurrent and non-recurrent (change-point) regime specifications.We illustrate the estimation method through simulations and apply it to seven financial time series of daily returns. We find structural breaks in the volatility dynamics of all series and recurrent regimes in nearly all series. Finally, we carry out a forecasting exercise to evaluate the usefulness of structural break models.

Change-point time series specifications constitute flexible models that capture unknown structural changes by allowing for switches in the model parameters. Nevertheless most models suffer from an over-parametrization issue since typically only one latent state variable drives the switches in all parameters. This implies that all parameters have to change when a break happens. To gauge whether and where there are structural breaks in realized variance, we introduce the sparse change-point HAR model. The approach controls for model parsimony by limiting the number of parameters which evolve from one regime to another. Sparsity is achieved thanks to employing a nonstandard shrinkage prior distribution. We derive a Gibbs sampler for inferring the parameters of this process. Simulation studies illustrate the excellent performance of the sampler. Relying on this new framework, we study the stability of the HAR model using realized variance series of several major international indices between January 2000 and August 2015.

A new model - the factorial hidden Markov volatility (FHMV) model - is proposed for financial returns and their latent variances. It is also applicable to model directly realized variances. Volatility is modeled as a product of three components: a Markov chain driving volatility persistence, an independent discrete process capable of generating jumps in the volatility, and a predictable (data-driven) process capturing the leverage effect. An economic interpretation is attached to each one of these components. Moreover, the Markov chain and jump components allow volatility to switch abruptly between thousands of states, and the transition matrix of the model is structured in such a way as to generate a high degree of volatility persistence. In-sample results on six financial time series highlight that the FHMV process compares favorably to state-of-the-art volatility models. A forecasting experiment shows that it also outperforms its competitors when predicting volatility over time horizons ranging from one to one hundred days.

Markov-switching models are usually specified under the assumption that all the parameters change when a regime switch occurs. Relaxing this hypothesis and being able to detect which parameters evolve over time is relevant for interpreting the changes in the dynamics of the series, for specifying models parsimoniously, and may be helpful in forecasting. We propose the class of sticky infinite hidden Markov-switching autoregressive moving average models, in which we disentangle the break dynamics of the mean and the variance parameters. In this class, the number of regimes is possibly infinite and is determined when estimating the model, thus avoiding the need to set this number by a model choice criterion. We develop a new Markov chain Monte Carlo estimation method that solves the path dependence issue due to the moving average component. Empirical results on macroeconomic series illustrate that the proposed class of models dominates the model with fixed parameters in terms of point and density forecasts. Appendix available at: https://ssrn.com/abstract=2965668

Sequential Monte Carlo (SMC) methods are widely used for non-linear filtering purposes. Nevertheless the SMC scope encompasses wider applications such as estimating static model parameters so much that it is becoming a serious alternative to Markov-Chain Monte-Carlo (MCMC) methods. Not only SMC algorithms draw posterior distributions of static or dynamic parameters but additionally provide an estimate of the marginal likelihood. The tempered and time (TNT) algorithm, developed in the paper, combines (off-line) tempered SMC inference with on-line SMC inference for drawing realizations from many sequential posterior distributions without experiencing a particle degeneracy problem. Furthermore, it introduces a new MCMC rejuvenation step that is generic, automated and well-suited for multi-modal distributions. As this update relies on the wide heuristic optimization literature, numerous extensions are already available. The algorithm is notably appropriate for estimating Change-point models. As an example, we compare Change-point GARCH models through their marginal likelihoods over time.

This Appendix contains additional empirical results with respect to the published article. In Section 1, the posterior results for the HDP parameters of the IHMS- ARMA models are presented for the U.S. GDP growth rate and inflation series. In Section 2, we report additional in-sample and forecasting results for the same series. In Section 3, some results for a different truncation choice of the number of regimes in the approximate model are reported. Full paper available at: https://ssrn.com/abstract=2965441

Sequential Monte Carlo (SMC) methods are widely used for filtering purposes of non-linear economic or financial models. Nevertheless the SMC scope encompasses wider applications such as estimating static model parameters so much that it is becoming a serious alternative to Markov- Chain Monte-Carlo (MCMC) methods. Not only SMC algorithms draw posterior distributions of static or dynamic parameters but additionally provide an estimate of the normalizing constant. The tempered and time (TNT) algorithm, developed in the paper, combines (off-line) tempered SMC inference with on-line SMC inference for estimating many slightly different distributions. The method encompasses the Iterated Batch Importance Sampling (IBIS) algorithm and more generally the Resample Move (RM) algorithm. Besides the number of particles, the TNT algorithm self-adjusts its calibrated parameters and relies on a new MCMC kernel that allows for particle interactions. The algorithm is well suited for efficiently back-testing models. We conclude by comparing in-sample and out-of-sample performances of complex volatility models.

Does commodity price volatility increase when inventories are low? We are the first ones to document this relationship. To that aim, we estimate asym- metric volatility models for a large set of commodities over 1994-2011. Since inventories are hard to measure, especially for high frequency data, we use positive return shocks as a new original proxy for inventories and find that asymmetric GARCH models reveal a significant inventory effect for many commodities. The results look robust. They hold if we allow the uncondi- tional variance to vary over time and if we relax the parametric form.

GARCH volatility models with fixed parameters are too restrictive for long time series due to breaks in the volatility process. Flexible alternatives are Markov-switching GARCH and change-point GARCH models. They require estimation by MCMC methods due to the path dependence problem. An unsolved issue is the computation of their marginal likelihood, which is essential for determining the number of regimes or change-points. We solve the problem by using particle MCMC, a technique proposed by Andrieu, Doucet, and Holenstein (2010). We examine the performance of this new method on simulated data, and we illustrate its use on several return series.