Forecasting crude oil prices: A scaled PCA approach

A-Tier
Journal: Energy Economics
Year: 2021
Volume: 97
Issue: C

Authors (4)

Score contribution per author:

1.005 = (α=2.01 / 4 authors) × 2.0x A-tier

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

Abstract

In this paper, we employ a novel dimension reduction approach, the scaled principal component analysis (s-PCA), to improve the oil price predictability with technical indicators. The empirical results show that the s-PCA model outperforms various competing models both in- and out-of-sample. From a market timing perspective, an oil futures investor can realize a larger Sharpe ratio using the s-PCA approach than using the competing models and Buy-and-Hold strategy. Furthermore, we investigate the driving forces behind the superior performance of the s-PCA model from a loading perspective. We illustrate that the s-PCA model can identify technical indicators with strong predictive power and put relatively large loadings on them when constructing diffusion indexes. Finally, our results are robust to a series of settings.

Technical Details

RePEc Handle
repec:eee:eneeco:v:97:y:2021:i:c:s0140988321000943
Journal Field
Energy
Author Count
4
Added to Database
2026-01-29