A1940
Title: Regime causality
Authors: Florian Ielpo - Centre Economie de la Sorbonne (France) [presenting]
Jerome Collet - Lombard Odier Investment Managers (Switzerland)
Abstract: A new class of embedded Markov switching models is introduced, allowing for one time series to see its regimes influence the probability of occurrence of other regimes. This mirrors the intuition that certain economic circumstances can have an influence on financial markets' behavior. We detail this new model alongside the necessary econometrics to estimate its parameters. We then show via Monte Carlo tests the precision with which its parameter can be estimated as a function of the sample size. Finally, we provide two empirical applications, one which ties to the selection of economic data when building a nowcasting indicator, and another testing the ability of widely used economic data to influence returns on equities.