SAFEX Brent Crude Oil futures prices

This research investigates whether a whitening filter approach can improve the modelling of oil futures prices by reducing serial dependence while preserving the essential structure of the data. Using SAFEX Brent crude oil futures price data as the primary empirical dataset, the study combines state-space reconstruction, spectral whitening, Markov-switching regime identification, and Ornstein--Uhlenbeck stochastic modelling to examine how model performance varies across different market conditions. Time-delay embedding is employed to recover the underlying dynamics of the series, while recurrence-based diagnostics are used to characterize nonlinear behaviour and regime differences. The whitened Ornstein--Uhlenbeck specification is then compared with the conventional benchmark to assess their relative adequacy in representing the observed price dynamics. The findings show that the usefulness of whitening is regime-dependent, with stronger performance in more stable market conditions and weaker performance in highly volatile or chaotic environments, thereby underscoring the importance of adaptive modelling strategies in oil futures markets.