QBA-Management Science: Report on MARKOV models' buisness applications - Term Paper Example

In this paper, the City-Wall Business Cycle characteristics were investigated by the help of univariate and multivariate Markov-switching specifications. We examined the stochastic properties of the cyclical pattern of the quarterly City-Wall Business Cycle real GNP between2005-2010. According to Blackmore and Susan (2000), then, all selected models were tested against their linear benchmarks. Furthermore, the multivariate specification with various macro aggregates and changing variance parameter outperformed the other MS models with reference to one-step ahead forecasting performance (Blackmore and Susan 42-44).

Finally, “based on inference from this model a chronology of business cycle turning points was determined”( p. 43). INTRODUCTION Research on business cycles has always been at the core of economic research agenda. This tradition has opened up two research areas namely, co movement among variables through the cycle, and the different behavior of the economy during different phases of the cycle Bradie and Michael (2000). The first one gave rise to the formation of dynamic factor models and composition of indices.

The latter one inspired the use of nonlinear regime switching models with the seminal work of. “These models assume that the probability of regime switching may be dependent on underlying economic fundamentals”. Recent researches has witnessed a synthesis of co movement and nonlinearity features of cycles since there is room for the analysis by incorporating both factor structure and regime (Bradie and Michael 43-44). The harmonization of two different methods of business cycle analysis also gave rise to Markov-switching vector auto-regression (MS-VAR) models.

This framework constitutes the multivariate generalization of the Hamilton’s single equation model. In these extended models there is an unobserved state driven by an ergodic Markov process that is common to all series, Dennett and Daniel (1990). In subsequent studies, the characterization and the testing of business cycle asymmetries based on MS-VAR models. It is an estimated duration dependent MS-VAR model by using a multimode Gibbs sampler since the computational burden in using the ML approach to such models is high.

“Combined both Markov-switching and structural identifying restrictions in a VAR model to analyze the reaction of variables to fundamental disturbances” (Dennett and Daniel 127-135). Dawkins and Richard (2008), despite these very influential recent developments both in theoretical and empirical literature, the analysis of City-wall business cycles has been somewhat limited and concentrated heavily on the leading indicators approach. However, “none of these studies explicitly analyzed the stochastic properties of business cycles in a rigorous econometric framework” (Dawkins and Richard 45-47).

Our major aim in this paper is to contribute in empirical modeling of City-wall business cycles with the help of MS models. Of our particular concern is MS-VAR models where the unobserved state is assumed to be common to all series used in model specifications. We consider both the co movement and the nonlinearity of the cyclical process of City-wall economy by employing a variety of MS-VAR models in which some or all of the parameters are allowed to change with the regime. Even though our concern is on the