Title: Does money growth predict inflation: Evidence from vector autoregressions using four centuries of data
Authors: Par Osterholm - Orebro university (Sweden) [presenting]
Rodney Edvinsson - Stockholm university (Sweden)
Sune Karlsson - Orebro University (Sweden)
Abstract: New evidence is added to a long-debated macroeconomic question, namely, whether monetary aggregates have predictive power for inflation or, put differently, whether monetary aggregates Granger cause inflation. We study this issue by employing vector autoregressive models to unique data. Using a historical dataset - consisting of annual Swedish data on money growth and inflation ranging from 1620 to 2018 - we conduct analysis both within- and out-of-sample. Using state-of-the-art Bayesian methods, we estimate models with stochastic volatility and also allow for the possibility of drifting parameters. Model selection based on marginal likelihoods indicates that a model with both stochastic volatility and drifting parameters is preferred. Concerning the issue of Granger causality, within-sample analysis provides strong evidence in favour of money growth Granger causing inflation when estimated on the full sample. Our out-of-sample forecast exercise does, however, not support this picture; the point forecast accuracy from the model where inflation is exogenous with respect to money growth tends to be on par with that from the model where it is endogenous. There is, however, a tendency for the model in which inflation is endogenous with respect to money growth to generate better density forecasts.