Piccari, Carlo (2024) Three essays on crude oil markets. PhD thesis, Birkbeck, University of London.
|
Text
Piccari C, thesis for library.pdf - Full Version Download (4MB) | Preview |
Abstract
The oil markets have been at the centre of attention of researchers over the past 50 years. Remarkably, the last 10 years on which the three chapters are focused, brought a large shares of events: starting with the shock collapse of prices in 2014 due to the massive arrival of shale oil in the US output, followed by price recovery and another, steeper, crash in 2020 due to the dramatic effect of COVID-19. In the former case supply was the major explanation of the collapse; in the latter demand vanished away because the world had come to a standstill. We firstly analysed the relationship between lagged WTI oil prices and the credit default swaps of 11 oil producers and 7 oil servicing companies of different sizes and examine how it varies through time. Our findings show a significant inverse effect for over 60% of the dataset. Moreover, this relationship grows in magnitude during periods of high volatility. Then, we applied a Variational Mode Decomposition-Neural Network and –Generalised Additive Models ensemble to forecast 5-minute WTI and BRENT prices in 2020 and in 2024. We highlighted the difficulties due to the structural instability caused by the unprecedented drop of WTI prices in 2020, which do not appear when forecasting data observed in other time-periods. Lastly, we proposed a methodology to account for the counter-intuitive features of sentiment analysis of oil-related news articles over a two years period, encompassing the events of the Ukraine war and Gaza conflict. We showed how adopting a scoring system of words co-occurrences and war-related nouns leads to increased predictability power of next day WTI returns, historical and conditional volatility. Topic modelling coupled with our sentiment measures was applied to forecast WTI volatility and returns, displaying higher accuracy over measures obtained via a standard sentiment analyser.
Metadata
| Item Type: | Thesis |
|---|---|
| Copyright Holders: | The copyright of this thesis rests with the author, who asserts his/her right to be known as such according to the Copyright Designs and Patents Act 1988. No dealing with the thesis contrary to the copyright or moral rights of the author is permitted. |
| Depositing User: | Acquisitions And Metadata |
| Date Deposited: | 29 Nov 2024 10:17 |
| Last Modified: | 29 Nov 2024 10:17 |
| URI: | https://eprints.bbk.ac.uk/id/eprint/54616 |
| DOI: | https://doi.org/10.18743/PUB.00054616 |
Statistics
Additional statistics are available via IRStats2.
Archive Staff Only (login required)
