Inflation Targeting Near 40 across Space and Time: Bayesian Model Comparison of Central Bank Deeds

This paper contributes to the literature by providing a comparative analysis of inflation targeting (IT) across groups of countries and time periods. We estimate via Bayesian techniques the central bank policy preferences of African inflation targeters (AfITs), employing the medium-scale New Keynesian small open economy model under complete asset markets (CAM) proposed by Kam et al. (2009), with application to advanced-country inflation targeters (ACITs), as extended also to incomplete asset markets (IAM) by McKnight et al. (2020), with application to Latin American inflation targeters (LAITs), and including or not real exchange-rate concerns in the social loss function of IT central banks. Our study convincingly selects CAM over IAM in a Bayesian model comparison of 4 model versions and compares the estimated weights of 4 typical IT central bank policy choices for 2 AfITs, 5 LAITs, and 3 ACITs in a common recent sample period, 2009:Q1-2021:Q4, as well as referring back to the sample periods in the cited original studies, starting in the early 1990s or early 2000s and with no or minimal overlap. Our findings confirm that all 10 IT central banks are firmly committed to their price stability mandate in the sense of prioritizing inflation stabilization, with an estimated almost unchanged 40-60% share across space and time. Adapting to real-world global developments as the millennium was unfolding, IT approaching the age of 40 seems to have evolved toward more 'flexible' regimes with increased 'fear of floating', but we also point to nuances or specificities across the 3 groups or 10 countries compared.