Posterior-Predictive Evidence on US Inflation Using Extended Phillips Curve Models with Non-Filtered Data

Abstract

Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in AbstractChanging time series properties of US inflation and economic activity, measured as marginal costs, are modeled within a set of extended Phillips Curve (PC) models. It is shown that mechanical removal or modeling of simple low frequency movements in the data may yield poor predictive results which depend on the model specification used. Basic PC models are extended to include structural time series models that describe typical time varying patterns in levels and volatilities. Forward and backward looking expectation components for inflation are incorporated and their relative importance is evaluated. Survey data on expected inflation are introduced to strengthen the information in the likelihood. Use is made of simulation based Bayesian techniques for the empirical analysis. No credible evidence is found on endogeneity and long run stability between inflation and marginal costs. Backward-looking inflation appears stronger than forward-looking one. Levels and volatilities of inflation are estimated more precisely using rich PC models. The extended PC structures compare favorably with existing basic Bayesian vector autoregressive and stochastic volatility models in terms of fit and prediction. Tails of the complete predictive distributions indicate an increase in the probability of deflation in recent years.

Parallelization Experience with Four Canonical Econometric Models Using ParMitISEM

“…The third application we consider is Bayesian inference of the IV model applied to Card (1995) data on education and income, also analyzed in Baştürk et al (2016). The fourth and final model is Bayesian inference of the structural form of the New Keynesian Phillips Curve (NKPC) model capturing the relationship between marginal cost and inflation, applied to quarterly inflation and marginal costs in the US, also analyzed in Baştürk et al (2014b). Due to the automatic and flexible nature of the MitISEM algorithm, all applications use a single parallel implementation of MitISEM, where only the target density has to be adjusted according to the application.…”

“…As shown in Baştürk et al (2014b), solving for the unobserved inflation expectations on the right hand side of (14) leads to the following model which is highly non-linear in structural parameters:…”

Interconnections between Eurozone and US Booms and Busts Using a Bayesian Panel Markov-Switching VAR Mode

“…However, overparameterization may lead to overfitting, especially in macroeconomic applications, where time series are characterized by a low number of observations, slowly changing means and time-varying variances (see, e.g., Basturk et al (2014)). These issues call for the use of a Bayesian approach to modeling and estimation, since it allows for inclusion of parameter restrictions with different degrees of prior belief.…”

“…They also suggest to consider MS models with regime-dependent volatility. We follow Krolzig (2000), Billio et al (2012) and Basturk et al (2014) and specifically assume that both unit-specific intercepts, a i (s i t ), and covariance matrices, Σ i (s i t ), are driven by the regime-switching variables {s i t } t but that the autoregressive coefficients A ip,k = A ip , ∀k remain constant.…”

“…Our paper also extends Kaufmann (2010), where a panel of univariate Markov-switching (MS) regression models is considered, by constructing a multivariate panel MSVAR structure for the country-specific time series. We build on basic model structures of Hamilton (1989) and Krolzig (2000) and consider Markov-switching dynamics for low and high frequency components that are specified as conditional means and covariance matrices of country-specific equations (see also Billio et al (2012), Basturk et al (2014) and Billio et al (2013b)). We further build on Kaufmann (2015) and use an endogenous time-varying transition mechanism to model the transition matrix of the country-specific Markov-chains.…”