Accelerating score-driven time series models

Abstract

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal ? Take down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Accelerating Peak Dating in a Dynamic Factor Markov-Switching Model

“…Here we additionally allow for relevant exogenous economic drivers, as suggested by Diebold et al (1994) and Filardo (1994). Furthermore, we consider a simplified analog of the recently developed aGAS framework by Blasques et al (2019), which considers the recent alignment of scores in determining the magnitude of the parameter update. In an empirical application, using the components of TCB's CEI from 1959 until 2020 and the term spread as an exogenous input, we find that the proposed method can significantly improve the real-time signaling power for NBER recessions.…”

“…Third, we adopt the novel accelerated GAS approach (aGAS) described in Blasques et al (2019) and adapt it to our needs. This extension allows for more rapid changes from score information in the time-varying parameter in question if the current score aligns with its immediate predecessor and smaller changes if this is not the case.…”

“…In the accelerated score-driven framework by Blasques et al (2019) the parameter A in (9), which determines how much to update in the direction of the score, is allowed to be timevarying. This is done by reapplying the score-driven approach to this parameter and under appropriate scaling can be seen to account for the autocorrelations of past score innovations.…”

“…This approach limits the number of additional parameters to be estimated and also does not require us to consider an additional (non-linear) link function to ensure (at least) the positivity of a t . Second, because here s f k t is a transform of the (scaled) score of f k,t , it is not straightforward to apply the framework of Blasques et al (2019) directly. Therefore by considering an intuitive measure of alignment, that may be interpreted as a first-order sample correlation, we provide an intuitively clear alternative that adheres to the core rationale of their accelerated approach.…”

“…This framework is therefore a multivariate extension of the methods described by Bazzi et al (2017) for Markov-switching models for univariate time series. Furthermore, we make use of the accelerated score-driven method recently introduced by Blasques et al (2019) and tailor it to our needs. In this approach the magnitude of the parameter update for a given value of the score-function is also made time-varying.…”