kth.sePublications KTH
Operational message
There are currently operational disruptions. Troubleshooting is in progress.
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
Robust Portfolio Optimization with Correlation Penalties
KTH, School of Engineering Sciences (SCI), Mathematics (Dept.), Mathematical Statistics.
2023 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesisAlternative title
Robust portföljoptimering med korrelationsstraff (Swedish)
Abstract [en]

Robust portfolio optimization models attempt to address the standard optimization method's high sensitivity to noise in the parameter estimates, by taking an investor's uncertainty about the estimates into account when finding an optimal portfolio. In this thesis, we study robust variations of an extension of the mean-variance problem, where an additional term penalizing the portfolio's correlation with an exogenous return sequence is included in the objective. Using a normalized risk factor model of the asset returns, estimations are done using EMA filtering as well as exponentially weighted linear regression. We show that portfolio performance can significantly improve with respect to a range of metrics, such as Sharpe ratio, expected shortfall and skewness, when using appropriate robust models and hyperparameters. We further show that extending the optimization problem with a correlation penalty can notably reduce portfolio correlation with an arbitrary return sequence, with only a small impact on other performance metrics.

Abstract [sv]

Robust portföljoptimering är en metod för att reducera vanliga portföljmodellers höga känslighet för brus i parameterskattningar, genom att ta en investerares osäkerhet kring skattningarna i åtanke när en optimal portfölj tas fram. I denna rapport studeras robusta varianter av ett utökat mean-variance-problem, där en straffterm för portföljens korrelation med en exogen avkastningsserie lagts till. Skattningarna bygger på en riskfaktor-modell för avkastningarna, och använder EMA-filter kombinerat med exponentiellt viktad linjär regression. Vi visar att en portföljs prestanda kan förbättras avsevärt med avseende på ett flertal prestandamått, till exempel Sharpe-kvot, expected shortfall och skevhet, vid användning av lämpliga robusta modeller och hyperparametrar. Vi visar också att inkludering av ett korrelationsstraff i optimeringsproblemet kan ge noterbara reduceringar i portföljens korrelation med en godtycklig avkastningsserie, med liten effekt på andra prestandamått.

Place, publisher, year, edition, pages
2023. , p. 71
Series
TRITA-SCI-GRU ; 2023:072
Keywords [en]
Portfolio Optimization, Portfolio Allocation, Robust Optimization, Correlation, Risk Factor Model, EMA Filtering, Weighted Linear Regression
Keywords [sv]
Portföljoptimering, Portföljallokering, Robust optimering, Korrelation, Riskfaktor-modell, EMA-filtrering, Viktad linjär regression
National Category
Other Mathematics
Identifiers
URN: urn:nbn:se:kth:diva-340415OAI: oai:DiVA.org:kth-340415DiVA, id: diva2:1816894
External cooperation
Lynx Asset Management AB
Subject / course
Mathematical Statistics
Educational program
Master of Science - Applied and Computational Mathematics
Supervisors
Examiners
Available from: 2023-12-04 Created: 2023-12-04 Last updated: 2023-12-04Bibliographically approved

Open Access in DiVA

fulltext(3339 kB)751 downloads
File information
File name FULLTEXT01.pdfFile size 3339 kBChecksum SHA-512
fde84878378719475db37822b4de19e33a83c40ff069dfd6a5add91c9db25d7d61ee17724ff42d6986bc521e363d19713ac11210f5c78065260c475840531e74
Type fulltextMimetype application/pdf

By organisation
Mathematical Statistics
Other Mathematics

Search outside of DiVA

GoogleGoogle Scholar
Total: 751 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: 568 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