Can Oil Price Predict Exchange Rate? Empirical Evidence from Deep Learning

Samir Safi, Salisu Aliyu, Kekere Sule Ibrahim, Olajide Idris Sanusi

Research output: Contribution to journalArticlepeer-review

1 Citation (Scopus)

Abstract

This paper critically analyses the predictability of exchange rates using oil prices. Extant literature that investigates the significance of oil prices in forecasting exchange rates remains largely inconclusive due to limitations arising from methodological issues. As such, this study uses deep learning approaches such as Multi-Layer Perceptron, Convolution Neural Network (CNN), and Long Short-Term Memory to predict exchange rates. In addition, the Empirical Mode Decomposition (EMD) of time series dataset was utilized to ascertain its effect on the quality of prediction. To examine the efficacy of using oil prices in forecasting exchange rates, bivariate models were also built. Of the three bivariate models developed, the EMD-CNN model has the best predictive performance. Results obtained show that oil price information has a strong influence on forecasting exchange rates.

Original languageEnglish
Pages (from-to)482-493
Number of pages12
JournalInternational Journal of Energy Economics and Policy
Volume12
Issue number4
DOIs
Publication statusPublished - Jul 19 2022

Keywords

  • Convolution neural network
  • Deep learning
  • Exchange rate
  • Long short-term memory
  • Multilayer perceptron
  • Oil prices

ASJC Scopus subject areas

  • General Energy
  • General Economics,Econometrics and Finance

Fingerprint

Dive into the research topics of 'Can Oil Price Predict Exchange Rate? Empirical Evidence from Deep Learning'. Together they form a unique fingerprint.

Cite this