Show simple item record

dc.contributor.advisorLucey, Brian
dc.contributor.authorRen, Boru
dc.date.accessioned2023-05-26T06:41:01Z
dc.date.available2023-05-26T06:41:01Z
dc.date.issued2023en
dc.date.submitted2023
dc.identifier.citationRen, Boru, Essays in Energy Finance, Trinity College Dublin, School of Business, Business & Administrative Studies, 2023en
dc.identifier.otherYen
dc.identifier.urihttp://hdl.handle.net/2262/102733
dc.descriptionAPPROVEDen
dc.description.abstractThis dissertation is composed of three main chapters related to energy finance, where the first two chapters each contain one essay and the third chapter contains two essays. In the first chapter (published in the Economic Modelling, co-authored with Dr. Brian Lucey), we study the relationship between the news tone, extracted using dictionary-based textual analysis, and the monthly oil prices. We directly measure the sentiment of Financial Times oil news articles from 1 June 2008 to 30 September 2020 using a oil-specific dictionary and the approach - la Loughran et al. (2019) as well as a commonly employed Henry's general financial dictionary. We find non- linear (linear) Granger-causality between news tone computed using the oil-specific (Henry's financial) dictionary and the oil prices. Since the preliminary results show predictive power of the news tone on oil prices, we perform out-of-samples forecasts over short (1-), medium (2-), long (3-month) horizons, controlled by other popular macroeconomic and sentiment variables. Unlike previous studies show that the News ToneHenry is useful in oil price trend forecasting, our results indicate that it has no ability to forecast the actual oil prices across all horizons. Instead, we find that the News ToneOil exhibits strong (weak) forecasting power over short (medium) terms, checked by robustness tests which further consider the Global Real Economic Activity, oil production and supply. We further verify the economic significance of the forecasting models by comparing the performance with those of a naive buy & hold strategy. Our study documents the use of domain-specific dictionary in relevant financial analysis. In the second chapter (published in the Energy Economics, co-authored with Dr. Brian Lucey), we examine the role of renewable energy stocks could play dur- ing cryptocurrency market turmoils from 1 January 2018 to 17 September 2021. Cryptocurrencies could be roughly classified as "dirty" and "clean" types based on the estimated energy consumption, depending on what underlying algorithm (e.g., Proof-of-Work (PoW), and non-PoW) is used. We first analyse the dynamic spillover patters among the renewable energy stocks, cryptocurrencies, S&P 500 (as a proxy for general stock market), and gold. We show that there is only weak connectedness between the renewable energy and cryptocurrency markets, which implies the possibility of renewable energy stocks to provide hedge and diversifi- cation benefits in the future. We further perform statistical analysis to examine the hedge and safe haven property of renewable energy stocks for cryptocurrencies? extreme negative movements and uncertainties, and vice versa. We confirm that renewable energy stocks have not yet become direct long-term hedge for either type of cryptocurrencies. However, it could at least serve as a weak safe haven for both types in extreme bearish markets. Moreover, renewable energy stocks are more likely to be a safe haven for "dirty" than "clean" cryptocurrencies during heightened market uncertainty. By contrast. cryptocurrencies are not general safe havens for renewable energy stocks. This study provides significant implications for investors, policy-makers, and founders of cryptocurrencies. In the third chapter (co-authored with Dr. Brian Lucey), we research the herd behaviour in emerging assets such as cryptocurrencies and renewable energy stocks. In this chapter?s first essay (published in the Finance Research Letters), similar to what we do in the previous chapter, we classify the cryptocurrencies into "dirty" and "clean" types, where we find empirical evidence that herding generally exists only in the dirty cryptocurrency market and is more significant in down than up markets. Moreover, we find that clean cryptocurrencies do herd, but with dirty cryptocurrencies, when the two markets are both in positive condition. The results are robust across value- and equal-weighted portfolios and provide valuable insights to cryptocurrency investors and policy makers. In the second essay which focuses on the renewable energy market in China (published in the Energy Economics), we find that the herds of renewable energy stocks significantly show up in the Chinese exchanges, which on the one hand, contradicts previous literature that declares that such market do not herd. On the other hand, our findings support literature that Chinese stock market is significantly inefficient and immature. We further investigate the asymmetric and time-varying characteristics of such behaviour in the Chinese market. We find that herding asymmetry is more pronounced during bullish markets and among smaller firms. When within-industry herding weakens, large price movements in the overall stock market provide additional trading signals for herding formation in this sector.en
dc.language.isoenen
dc.publisherTrinity College Dublin. School of Business. Discipline of Business & Administrative Studiesen
dc.rightsYen
dc.subjectEnergy Economicsen
dc.subjectHerd Behavioren
dc.subjectSpillover Effectsen
dc.subjectSentiment Analysisen
dc.subjectCryptocurrencyen
dc.titleEssays in Energy Financeen
dc.typeThesisen
dc.type.supercollectionthesis_dissertationsen
dc.type.supercollectionrefereed_publicationsen
dc.type.qualificationlevelDoctoralen
dc.identifier.peoplefinderurlhttps://tcdlocalportal.tcd.ie/pls/EnterApex/f?p=800:71:0::::P71_USERNAME:RENBen
dc.identifier.rssinternalid256253en
dc.rights.ecaccessrightsopenAccess


Files in this item

Thumbnail
Thumbnail

This item appears in the following Collection(s)

Show simple item record