kth.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Clustering Economic Regimes with Jump Models for Asset Management
KTH, School of Industrial Engineering and Management (ITM). KTH, School of Engineering Sciences (SCI), Mathematics (Dept.). Industriell Ekonomi.
KTH, School of Industrial Engineering and Management (ITM). KTH, School of Engineering Sciences (SCI), Mathematics (Dept.). Industriell Ekonomi.
2025 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

This thesis proposes an innovative approach for identifying economic regimes using Jump Models (JMs), a family of clustering and optimization methods. Traditional regime detection methods, such as Markov Switching Models (MSMs), often face challenges in capturing the complexity and dynamic nature of modern financial markets. By using a JM combined with a systematic variable selection and transformation process, this study improves regime identification by directly optimizing regime separability and return profiles, while explicitly penalizing unnecessary regime shifts through a jump penalty. The model is applied to U.S. macroeconomic and financial data from 2003 to 2025, using variables such as interest rates, macroeconomic variables, market data, and commodity prices. Regimes are evaluated based on their statistical properties (e.g., volatility, correlations, drawdowns) and their implications for portfolio performance. This approach allows for a nuanced identification of economic states by accommodating temporal complexities and penalizing unnecessary regime shifts. Furthermore, the regime assignment is performed using an online classifier to ensure out-of-sample validity. The results demonstrate that this method offers more stable and economically interpretable regime definitions than traditional approaches. In out-of-sample tests, a regime-aware portfolio constructed using our framework outperformed regimeagnostic benchmarks, including the Mean-Variance portfolio, by achieving a Sharpe ratio of 0.68 compared to 0.51, with a lower market beta compared to the S&P 500. These improvements support the use of JM-based regime classification for strategic asset allocation and risk management by institutional investors.

Abstract [sv]

Denna uppsats presenterar ett innovativt ramverk för att identifiera och analysera ekonomiska regimer med hjälp av Hoppmodeller (JMs), en samling klustringsoch optimeringsmetoder. Traditionella modeller för regimskiften, som till exempel Hidden Markov Models (HMM), har ofta svårt att fånga den komplexitet och dynamik som kännetecknar verkliga makrofinansiella system. Genom att kombinera hoppmodeller med en systematisk variabelselektion och datatransformering förbättras regimidentifieringen. Regimernas separabilitet och avkastningsprofil optimeras samtidigt som onödiga regimskiften uttryckligen straffas via en hopp-penalty. Modellen tillämpas på amerikansk makro- och marknadsdata mellan 2003 och 2025, där indikatorer som räntor, makroekonomiska variabler, marknadsdata och råvarupriser ingår. Regimerna utvärderas både statistiskt (t.ex. volatilitet, korrelationer, drawdowns) och utifrån deras påverkan på portföljprestanda. Dessutom utförs regimklassificering med en online algoritm för att säkerställa utvärdering out-ofsample. Resultaten visar att denna metod ger mer stabila regimdefinitioner jämfört med traditionella tillvägagångssätt. Den regimmedvetna portföljen baserad på vår modell överträffade en statisk Mean-Variance-portfölj med Sharpe-kvot på 0,68 jämfört med 0,51, samt ett lägre marknadsbeta mot S&P 500. Detta stärker modellens relevans för strategisk tillgångsallokering och riskhantering för institutionella investerare.

Place, publisher, year, edition, pages
2025. , p. 53
Series
Examensarbete INDEK
Keywords [en]
Time Series, Clustering, Economic Regimes, Regime Switching, Asset Allocation, Financial Market Regimes, Statistical Jump Models, Regime Identification
Keywords [sv]
Tidsserier, Klustring, Ekonomiska Regimer, Regimskiften, Tillgångsallokering, Finansiella Marknadsregimer, Statistiska Hoppmodeller, Regimidentifiering
National Category
Computer and Information Sciences Engineering and Technology Mathematical sciences Economics and Business
Identifiers
URN: urn:nbn:se:kth:diva-364103OAI: oai:DiVA.org:kth-364103DiVA, id: diva2:1963774
External cooperation
Fjärde AP-fonden (AP4)
Subject / course
Applied Mathematics and Industrial Economics
Educational program
Master of Science in Engineering - Industrial Engineering and Management
Presentation
2025-05-19, KTH, Lindstedtsvägen 25, Stockholm, 17:30 (English)
Supervisors
Examiners
Available from: 2025-06-04 Created: 2025-06-04 Last updated: 2025-06-04Bibliographically approved

Open Access in DiVA

fulltext(2477 kB)38 downloads
File information
File name FULLTEXT01.pdfFile size 2477 kBChecksum SHA-512
dded0cea4d12b70f4bf7438a6edbaabee86390b0a807c8aab4007f4fdaee8afa69b3dbae393633b1e75a144c6988af17bb4c9e17b4ded99f90af3aedb4f84a77
Type fulltextMimetype application/pdf

Search in DiVA

By author/editor
Möller, AnnaHellgren, Erik
By organisation
School of Industrial Engineering and Management (ITM)Mathematics (Dept.)
Computer and Information SciencesEngineering and TechnologyMathematical sciencesEconomics and Business

Search outside of DiVA

GoogleGoogle Scholar
Total: 38 downloads
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 165 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf