Application of log-periodic power law to oil spot market

Oil price crashes and bubble regimes in oil spot markets have a substantial impact on the real economy. One model that has been proposed for detecting bubble regimes and forecasting the most probable date of the crash is the log-periodic power law. The model aims to detect log-periodic oscillations...

Täydet tiedot

Bibliografiset tiedot
Päätekijä: Virtanen, Lauri
Muut tekijät: Jyväskylän yliopiston kauppakorkeakoulu, Jyväskylä University School of Business and Economics, Jyväskylän yliopisto, University of Jyväskylä
Aineistotyyppi: Pro gradu
Kieli:eng
Julkaistu: 2025
Aiheet:
Linkit: https://jyx.jyu.fi/handle/123456789/102582
_version_ 1833317277931405312
author Virtanen, Lauri
author2 Jyväskylän yliopiston kauppakorkeakoulu Jyväskylä University School of Business and Economics Jyväskylän yliopisto University of Jyväskylä
author_facet Virtanen, Lauri Jyväskylän yliopiston kauppakorkeakoulu Jyväskylä University School of Business and Economics Jyväskylän yliopisto University of Jyväskylä Virtanen, Lauri Jyväskylän yliopiston kauppakorkeakoulu Jyväskylä University School of Business and Economics Jyväskylän yliopisto University of Jyväskylä
author_sort Virtanen, Lauri
datasource_str_mv jyx
description Oil price crashes and bubble regimes in oil spot markets have a substantial impact on the real economy. One model that has been proposed for detecting bubble regimes and forecasting the most probable date of the crash is the log-periodic power law. The model aims to detect log-periodic oscillations in price data which reveal a bubble regime and enable the model to forecast the most probable crash date. This thesis looks at 4 known bubble regimes in Brent and WTI spot oil markets to test if the model is successful at detecting the bubble regimes and if it can produce meaningful estimations of the most probable crash date. The data for this study was gathered from LSEG Workspace and this study presents a novel machine learning implementation of the LPPL model. The results show that the machine learning implementation of LPPL has significant challenges in calibrations to real world data. This study also presents contradictory evidence to prior studies. In this study the model was not able to confirm the existence of oil price bubble in 2001-2008 which has been detected by prior research using LPPL. A period in 1990 was also found where there exists a low confidence bubble regime in WTI but not in Brent which is contradictory to prior research. This research provides a novel implementation of the LPPL model with machine learning frameworks and presents difficulties with this implementation. Öljyn hintaromahdukset sekä hintakuplat vaikuttavat huomattavasti reaalitalouden toimintaan. Eräs malli jota on ehdotettu näiden hintakuplien tunnistamiseen sekä todennäköisimmän romahduksen päivämäärän ennustamiseen on logperiodinen potenssilaki. Tämä malli koettaa havaita log-periodisia oskillaatioita hintadatassa havaitakseen hintakuplia sekä tuottamaan merkityksellisiä ennusteita kaikista todennäköisimmästä ajankohdasta romahdukselle. Tämä tutkimus kattaa 4 tunnettua hintakuplaa Brent sekä WTI öljylaatujen spot-markkinoilla testatakseen mallin kykyä tunnistaa hintakupla sekä pystyykö se tuottamaan merkityksellisen ennusteen todennäköisimmästä ajankohdasta hinnan romahdukselle. Tutkimuksen data kerättiin LSEG Workspace palvelusta ja tutkimuksessa esitellään uusi tapa implementoida LPPL malli koneoppimiskirjastoja hyödyntäen. Tutkimuksen tulokset osoittavat, että LPPL mallin implementointi koneoppimiskirjastoja hyödyntäen sisältää huomattavia haasteita mallin kalibroinnissa oikealla hintadatalla. Tutkmus myös esitttää ristiriitaisia tuloksia aiempien tutkmusten kanssa. Tässä tutkimuksessa malli ei kyennyt tunnistamaan ja täten validoimaan aikaisemman LPPL tutkimuksen havaitsemaa hintakuplaa vuosien 2001-2008 välillä. Malli havaitsi myös vuodelta 1990 hintakuplan, joka esiintyi ainoastaaan WTI öljylaadussa mikä on aiempien tutkimustulosten vastaista. Tämä tutkimus esittää uuden implementointitavan LPPL mallille hyödyntäen koneoppimis kirjastoja sekä tuo esille haasteita tämän implementoinnin totetuksessa.
first_indexed 2025-05-20T20:05:15Z
format Pro gradu
free_online_boolean 1
fullrecord [{"key": "dc.contributor.advisor", "value": "Lehkonen, Heikki", "language": null, "element": "contributor", "schema": "dc", "qualifier": "advisor"}, {"key": "dc.contributor.author", "value": "Virtanen, Lauri", "language": null, "element": "contributor", "schema": "dc", "qualifier": "author"}, {"key": "dc.date.accessioned", "value": "2025-05-20T10:57:35Z", "language": null, "element": "date", "schema": "dc", "qualifier": "accessioned"}, {"key": "dc.date.available", "value": "2025-05-20T10:57:35Z", "language": null, "element": "date", "schema": "dc", "qualifier": "available"}, {"key": "dc.date.issued", "value": "2025", "language": null, "element": "date", "schema": "dc", "qualifier": "issued"}, {"key": "dc.identifier.uri", "value": "https://jyx.jyu.fi/handle/123456789/102582", "language": null, "element": "identifier", "schema": "dc", "qualifier": "uri"}, {"key": "dc.description.abstract", "value": "Oil price crashes and bubble regimes in oil spot markets have a substantial impact on the real economy. One model that has been proposed for detecting bubble regimes and forecasting the most probable date of the crash is the log-periodic power law. The model aims to detect log-periodic oscillations in price data which reveal a bubble regime and enable the model to forecast the most probable crash date.\n\nThis thesis looks at 4 known bubble regimes in Brent and WTI spot oil markets to test if the model is successful at detecting the bubble regimes and if it can produce meaningful estimations of the most probable crash date. The data for this study was gathered from LSEG Workspace and this study presents a novel machine learning implementation of the LPPL model.\n\nThe results show that the machine learning implementation of LPPL has significant challenges in calibrations to real world data. This study also presents contradictory evidence to prior studies. In this study the model was not able to confirm the existence of oil price bubble in 2001-2008 which has been detected by prior research using LPPL. A period in 1990 was also found where there exists a low confidence bubble regime in WTI but not in Brent which is contradictory to prior research.\n\nThis research provides a novel implementation of the LPPL model with machine learning frameworks and presents difficulties with this implementation.", "language": "en", "element": "description", "schema": "dc", "qualifier": "abstract"}, {"key": "dc.description.abstract", "value": "\u00d6ljyn hintaromahdukset sek\u00e4 hintakuplat vaikuttavat huomattavasti reaalitalouden toimintaan. Er\u00e4s malli jota on ehdotettu n\u00e4iden hintakuplien tunnistamiseen sek\u00e4 todenn\u00e4k\u00f6isimm\u00e4n romahduksen p\u00e4iv\u00e4m\u00e4\u00e4r\u00e4n ennustamiseen on logperiodinen potenssilaki. T\u00e4m\u00e4 malli koettaa havaita log-periodisia oskillaatioita hintadatassa havaitakseen hintakuplia sek\u00e4 tuottamaan merkityksellisi\u00e4 ennusteita kaikista todenn\u00e4k\u00f6isimm\u00e4st\u00e4 ajankohdasta romahdukselle. \n\nT\u00e4m\u00e4 tutkimus kattaa 4 tunnettua hintakuplaa Brent sek\u00e4 WTI \u00f6ljylaatujen spot-markkinoilla testatakseen mallin kyky\u00e4 tunnistaa hintakupla sek\u00e4 pystyyk\u00f6 se tuottamaan merkityksellisen ennusteen todenn\u00e4k\u00f6isimm\u00e4st\u00e4 ajankohdasta hinnan romahdukselle. Tutkimuksen data ker\u00e4ttiin LSEG Workspace palvelusta ja tutkimuksessa esitell\u00e4\u00e4n uusi tapa implementoida LPPL malli koneoppimiskirjastoja hy\u00f6dynt\u00e4en.\n\nTutkimuksen tulokset osoittavat, ett\u00e4 LPPL mallin implementointi koneoppimiskirjastoja hy\u00f6dynt\u00e4en sis\u00e4lt\u00e4\u00e4 huomattavia haasteita mallin kalibroinnissa oikealla hintadatalla. Tutkmus my\u00f6s esittt\u00e4\u00e4 ristiriitaisia tuloksia aiempien tutkmusten kanssa. T\u00e4ss\u00e4 tutkimuksessa malli ei kyennyt tunnistamaan ja t\u00e4ten validoimaan aikaisemman LPPL tutkimuksen havaitsemaa hintakuplaa vuosien 2001-2008 v\u00e4lill\u00e4. Malli havaitsi my\u00f6s vuodelta 1990 hintakuplan, joka esiintyi ainoastaaan WTI \u00f6ljylaadussa mik\u00e4 on aiempien tutkimustulosten vastaista.\n\nT\u00e4m\u00e4 tutkimus esitt\u00e4\u00e4 uuden implementointitavan LPPL mallille hy\u00f6dynt\u00e4en koneoppimis kirjastoja sek\u00e4 tuo esille haasteita t\u00e4m\u00e4n implementoinnin totetuksessa.", "language": "fi", "element": "description", "schema": "dc", "qualifier": "abstract"}, {"key": "dc.description.provenance", "value": "Submitted by jyx lomake-julkaisija (jyx-julkaisija.group@korppi.jyu.fi) on 2025-05-20T10:57:35Z\nNo. of bitstreams: 0", "language": "en", "element": "description", "schema": "dc", "qualifier": "provenance"}, {"key": "dc.description.provenance", "value": "Made available in DSpace on 2025-05-20T10:57:35Z (GMT). No. of bitstreams: 0", "language": "en", "element": "description", "schema": "dc", "qualifier": "provenance"}, {"key": "dc.format.extent", "value": "75", "language": null, "element": "format", "schema": "dc", "qualifier": "extent"}, {"key": "dc.format.mimetype", "value": "application/pdf", "language": null, "element": "format", "schema": "dc", "qualifier": "mimetype"}, {"key": "dc.language.iso", "value": "eng", "language": null, "element": "language", "schema": "dc", "qualifier": "iso"}, {"key": "dc.rights", "value": "CC BY 4.0", "language": null, "element": "rights", "schema": "dc", "qualifier": null}, {"key": "dc.title", "value": "Application of log-periodic power law to oil spot market", "language": null, "element": "title", "schema": "dc", "qualifier": null}, {"key": "dc.type", "value": "master thesis", "language": null, "element": "type", "schema": "dc", "qualifier": null}, {"key": "dc.identifier.urn", "value": "URN:NBN:fi:jyu-202505204426", "language": null, "element": "identifier", "schema": "dc", "qualifier": "urn"}, {"key": "dc.contributor.faculty", "value": "Jyv\u00e4skyl\u00e4n yliopiston kauppakorkeakoulu", "language": "fi", "element": "contributor", "schema": "dc", "qualifier": "faculty"}, {"key": "dc.contributor.faculty", "value": "Jyv\u00e4skyl\u00e4 University School of Business and Economics", "language": "en", "element": "contributor", "schema": "dc", "qualifier": "faculty"}, {"key": "dc.contributor.organization", "value": "Jyv\u00e4skyl\u00e4n yliopisto", "language": "fi", "element": "contributor", "schema": "dc", "qualifier": "organization"}, {"key": "dc.contributor.organization", "value": "University of Jyv\u00e4skyl\u00e4", "language": "en", "element": "contributor", "schema": "dc", "qualifier": "organization"}, {"key": "dc.subject.discipline", "value": "Master's Degree Programme in Banking and International Finance", "language": "fi", "element": "subject", "schema": "dc", "qualifier": "discipline"}, {"key": "dc.subject.discipline", "value": "Master's Degree Programme in Banking and International Finance", "language": "en", "element": "subject", "schema": "dc", "qualifier": "discipline"}, {"key": "dc.type.coar", "value": "http://purl.org/coar/resource_type/c_bdcc", "language": null, "element": "type", "schema": "dc", "qualifier": "coar"}, {"key": "dc.rights.copyright", "value": "\u00a9 The Author(s)", "language": null, "element": "rights", "schema": "dc", "qualifier": "copyright"}, {"key": "dc.rights.accesslevel", "value": "openAccess", "language": null, "element": "rights", "schema": "dc", "qualifier": "accesslevel"}, {"key": "dc.type.publication", "value": "masterThesis", "language": null, "element": "type", "schema": "dc", "qualifier": "publication"}, {"key": "dc.format.content", "value": "fulltext", "language": null, "element": "format", "schema": "dc", "qualifier": "content"}, {"key": "dc.rights.url", "value": "https://creativecommons.org/licenses/by/4.0/", "language": null, "element": "rights", "schema": "dc", "qualifier": "url"}, {"key": "dc.description.accessibilityfeature", "value": "ei tietoa saavutettavuudesta", "language": "fi", "element": "description", "schema": "dc", "qualifier": "accessibilityfeature"}, {"key": "dc.description.accessibilityfeature", "value": "unknown accessibility", "language": "en", "element": "description", "schema": "dc", "qualifier": "accessibilityfeature"}]
id jyx.123456789_102582
language eng
last_indexed 2025-05-20T20:05:15Z
main_date 2025-01-01T00:00:00Z
main_date_str 2025
online_boolean 1
online_urls_str_mv {"url":"https:\/\/jyx.jyu.fi\/bitstreams\/5337c8e5-f6f4-432c-a1d2-fb01f0e5f32c\/download","text":"URN:NBN:fi:jyu-202505204426.pdf","source":"jyx","mediaType":"application\/pdf"}
publishDate 2025
record_format qdc
source_str_mv jyx
spellingShingle Virtanen, Lauri Application of log-periodic power law to oil spot market Master's Degree Programme in Banking and International Finance
title Application of log-periodic power law to oil spot market
title_full Application of log-periodic power law to oil spot market
title_fullStr Application of log-periodic power law to oil spot market Application of log-periodic power law to oil spot market
title_full_unstemmed Application of log-periodic power law to oil spot market Application of log-periodic power law to oil spot market
title_short Application of log-periodic power law to oil spot market
title_sort application of log periodic power law to oil spot market
title_txtP Application of log-periodic power law to oil spot market
topic Master's Degree Programme in Banking and International Finance
topic_facet Master's Degree Programme in Banking and International Finance
url https://jyx.jyu.fi/handle/123456789/102582 http://www.urn.fi/URN:NBN:fi:jyu-202505204426
work_keys_str_mv AT virtanenlauri applicationoflogperiodicpowerlawtooilspotmarket