Forecasts of the real price of oil revisited: Do they beat the random walk?

B-Tier
Journal: Journal of Banking & Finance
Year: 2023
Volume: 154
Issue: C

Authors (2)

Ellwanger, Reinhard (not in RePEc) Snudden, Stephen (Wilfrid Laurier University)

Score contribution per author:

1.005 = (α=2.01 / 2 authors) × 1.0x B-tier

α: calibrated so average coauthorship-adjusted count equals average raw count

Abstract

In macroeconomic forecasting, the real price of oil is traditionally computed as the monthly average price of oil deflated by the price index. Consequently, the no-change forecast used to benchmark forecasts of the real price of crude oil is a monthly average price. We demonstrate that an alternative no-change forecast which reflects the random walk forecast from daily oil prices – the end-of-month price – is significantly more accurate in predicting the real price of oil up to one year ahead. We find that at the one-step-ahead prediction, all existing forecasts that outperform the monthly average no-change forecast perform worse than the end-of-month no-change forecast. The results call into question the usefulness of existing forecasting approaches for the real price of crude oil relative to naive forecasts.

Technical Details

RePEc Handle
repec:eee:jbfina:v:154:y:2023:i:c:s0378426623001619
Journal Field
Finance
Author Count
2
Added to Database
2026-01-29