Artificial Intelligence and Blockchain for Oil Risk Management in the Moroccan Financial Market
DOI:
https://doi.org/10.71420/ijref.v3i9-2.413Keywords:
Artificial intelligence, Blockchain, Oil risk, Moroccan financial market, Volatility , Data governanceAbstract
This article examines how artificial intelligence and a permissioned blockchain can be combined to improve the forecasting and governance of oil-related risk in the Moroccan financial market. The empirical analysis uses daily observations from 24 November 2020 to 24 November 2023 for the MASI, the Oil/Gas and Transport sector indices, Brent, WTI and USD/MAD. Log returns are evaluated out of sample using a chronological 80/20 split. ARIMA(1,0,1), Random Forest, XGBoost and LSTM are compared using MAE, RMSE, directional accuracy and Diebold-Mariano tests. ARCH effects are also examined to characterize conditional heteroskedasticity; the GARCH framework is retained as a theoretical volatility reference and is not included in the point-forecast comparison. Results show that algorithmic complexity does not guarantee superior forecasts: ARIMA records the lowest RMSE for Oil/Gas (0.01480), whereas the naive benchmark is marginally best for the MASI (0.00681) and Transport (0.01257). AI models therefore do not deliver a stable out-of-sample improvement in this sample. Random Forest variable importance nevertheless assigns a larger contribution to oil variables for Transport (54.2%) than for the MASI (35.4%) or Oil/Gas (33.3%). These findings reject a presumption of systematic AI superiority and support strict out-of-sample model governance. Blockchain is presented as a conceptual traceability and auditability layer whose institutional benefits require separate prototype validation.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Soufiane Zakhouni, Achraf El Yadmani, Mohamed Benkhalla, Omar Kharbouch

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.



