The paper aims at comparing forecast ability of VAR/VEC models with a non-changing covariance matrix and two classes of Bayesian Vector Error Correction – Stochastic Volatility (VEC-SV) models, which combine the VEC representation of a VAR structure with stochastic volatility, represented by the Multiplicative Stochastic Factor (MSF) process, the SBEKK form or the MSF-SBEKK specification.

Based on macro-data coming from the Polish economy (time series of unemployment, inflation and interest rates) we evaluate predictive density functions employing of such measures as log predictive density score, continuous rank probability score, energy score, probability integral transform. Each of them takes account of different feature of the obtained predictive density functions.

This paper presents some new results on exogeneity in models with latent variables. The concept of exogeneity is extended to the class of models with latent variables, in which a subset of parameters and latent variables is of interest. Exogeneity is discussed from the Bayesian point of view. We propose sufficient weak and strong exogeneity conditions in the vector error correction model (VECM) with stochastic volatility (SV) disturbances. Finally, an empirical illustration based on the VECM-SV model for the daily growth rates of two main official Polish exchange rates: USD/PLN and EUR/PLN, as well as EUR/USD from the international Forex market is presented. The exogeneity of the EUR/USD rate is examined. The strong exogeneity hypothesis of the EUR/USD rate is not rejected by the data.

A Bayesian stochastic volatility model with a leverage effect, normal errors and jump component with the double exponential distribution of a jump value is proposed. The ready to use Gibbs sampler is presented, which enables one to conduct statistical inference. In the empirical study, the SVLEDEJ model is applied to model logarithmic growth rates of one month forward gas prices. The results reveal an important role of both jump and stochastic volatility components.

In the paper we present and apply a Bayesian jump-diffusion model and stochastic volatility models with jumps. The problem of how to classify an observation as a result of a jump is addressed, under the Bayesian approach, by introducing latent variables. The empirical study is focused on the time series of gas forward contract prices and EUA futures prices. We analyse the frequency of jumps and relate the moments in which jumps occur to calendar effects or political and economic events and decisions. The calendar effects explain many jumps in gas contract prices. The single jump is identified in the EUA futures prices under the SV-type models. The jump is detected on the day the European Parliament voted against the European Commission’s proposal of backloading. The Bayesian results are compared with the outcomes of selected non-Bayesian techniques used for detecting jumps.

The study aims at a statistical verification of breaks in the risk-return relationship for shares of individual companies quoted at the Warsaw Stock Exchange. To this end a stochastic volatility model incorporating Markov switching in-mean effect (SV-MS-M) is employed. We argue that neglecting possible regime changes in the relation between expected return and volatility within an ordinary SV-M specification may lead to spurious insignificance of the risk premium parameter (as being ’averaged out’ over the regimes).Therefore, we allow the volatility-in-mean effect to switch over different regimes according to a discrete homogeneous two- or
three-state Markov chain. The model is handled within Bayesian framework, which allows to fully account for the uncertainty of
model parameters, latent conditional variances and state variables. MCMC methods, including the Gibbs sampler, Metropolis-Hastings algorithm and the forward-filtering-backward-sampling scheme are suitably adopted to obtain posterior densities of interest as well
as marginal data density. The latter allows for a formal model comparison in terms of the in-sample fit and, thereby, inference on the
’adequate’ number of the risk premium regime