A1290
Title: A panel extension for noncausal models
Authors: Kevin Cecere Palazzo - Vrije Universiteit (Netherlands) [presenting]
Sean Telg - Vrije Universiteit Amsterdam (Netherlands)
Siem Jan Koopman - VU Amsterdam (The Netherlands)
Francisco Blasques - VU University Amsterdam (Netherlands)
Abstract: Noncausal autoregressive models offer a useful framework to model speculative bubbles in economics and finance, thanks to their ability to generate explosive phenomena. Standard literature on common features for multivariate noncausal models implicitly assumes that common features occur at the same time. A panel extension of noncausal models is considered in order to accommodate for joint modeling of financial bubbles when explosive behaviors are not synchronous across units. Parameters are proposed for estimation by approximate maximum likelihood (AML) and least squares (LS), and the asymptotic behavior of the proposed estimators is established. Properties of limit behavior of the proposed estimators are analyzed in finite and infinite variance frameworks. A testing procedure is derived to test the validity of panel extension in noncausal framework, and the potential usefulness of the results is illustrated in an empirical application on cryptocurrencies.
