This thesis deals with a multi-factor stochastic volatility model for crude oil futures, capturing the main characteristics of crude oil markets. We derive the state-space form of this model as well as the Kalman filter algorithm for prediction and estimation. Subsequently, we extend the model by making the model parameters depend on an unobservable Markov chain. The di↵erent states or regimes in which the Markov chain and thus the model can be found are periods of structural changes in the stream of data. The aim of this thesis is to estimate the probabilities of the model being in a certain regime. For this purpose, we generalize the Kalman filter to the Kalman-Hamilton-Kim filter. At this juncture, the modified algorithm provides the sought probabilities in which regime the model is located. A computer-based implementation of a simple version of the model with daily WTI (West Texas Intermediate) crude oil and OVX (Oil Volatility Index) data from 2007 up to 2020 provides estimates for the model parameters as well as the regime-probabilities. Here, a particular focus lies on the historic drop of oil prices in April 2020.     «

This thesis deals with a multi-factor stochastic volatility model for crude oil futures, capturing the main characteristics of crude oil markets. We derive the state-space form of this model as well as the Kalman filter algorithm for prediction and estimation. Subsequently, we extend the model by making the model parameters depend on an unobservable Markov chain. The di↵erent states or regimes in which the Markov chain and thus the model can be found are periods of structural changes in the stre...     »