OeNB-Freitagsseminar with Philippe Goulet Coulombe

Maximally Forward-Looking Core Inflation

OeNB-Freitagsseminar with Philippe Goulet Coulombe, Université du Québec à Montréal
Core inflation measures are at the heart of post-Pandemic monetary policy debates. Typically, they are either built from heuristics (like exclusion- or trimming-based approaches) or unsupervised learning methods, like factor models. Then, in a post-estimation step, their empirical merits are assessed—with perhaps the most desired quality being that the core series should be as indicative as possible of future inflation conditions. We create a new core inflation measure that is explicitly designed to succeed at that goal. Precisely, we write a nonnegative ridge regression problem that optimizes CPI/PCE subcomponent weights such that the aggregate is maximally predictive of future headline inflation. Ordering subcomponents according to their rank in each period switches the algorithm to learning a supervised optimally trimmed inflation. In an extensive out-of-sample forecasting experiment for the US, Euro area, and Canada, we find substantial improvements for signaling medium-term inflation developments in both the pre- and post-Covid years. Those coming from the optimally trimmed version are particularly striking and are attributable to a highly asymmetric trimming which contrasts traditional indicators enforcing symmetry. We also find that this metric was indicating first upward pressures on underlying inflation as early as mid-2020.

Date
Friday, 15 December 2023 | Start: 11:00 AM | End: 12:30 PM             

Venue
The event is planned both, online via Webex and onsite at the Oesterreichische Nationalbank, Otto-Wagner-Platz 3, 1090 Vienna, 3rd floor (M3080).

Please register by 13 December 2023, at the latest.