Mean Variance Portfolio Optimization including cryptocurrencies and time-varying conditional correlation
Author
Dalgaard, Mads Skinnerup
Term
4. semester
Publication year
2024
Submitted on
2024-06-03
Pages
30
Abstract
Dette speciale undersøger, om tilføjelse af kryptovalutaer til en traditionelt diversificeret portefølje kan forbedre diversifikation og samlet performance. Med en mean–variance-optimeringsramme (MVO) konstrueres 12 porteføljer, der afspejler forskellige investorsentimenter, og de vurderes med Sharpe-ratio, efficiente fronter og subjektive vægtbegrænsninger. For at adressere MVO’s antagelse om konstante korrelationer estimeres desuden tidsvarierende betingede korrelationer mellem to indeks og to kryptovalutaer via en DCC‑GARCH‑model. Resultaterne viser, at porteføljer med tre kryptovalutaer opnår bedre risikojusterede afkast end traditionelle porteføljer og porteføljer med kun Bitcoin; de kan også levere højere afkast, dog til højere risiko. DCC‑GARCH‑analysen påviser markant tidsvariation i korrelationer, der spænder fra negative til relativt høje positive værdier, hvilket understreger, at diversifikationsfordele kan ændre sig betydeligt over tid. Specialet drøfter implikationer og begrænsninger ved metoden, foreslår potentielle metodiske udvidelser og belyser adfærdsøkonomiske faktorer i kryptomarkeder, der kan føre til suboptimale investeringsbeslutninger.
This thesis examines whether adding cryptocurrencies to a traditionally diversified portfolio can improve diversification and overall performance. Using a mean–variance optimization (MVO) framework, the study builds 12 portfolios reflecting different investor sentiments and evaluates them with the Sharpe ratio, efficient frontiers, and subjective weight constraints. To address MVO’s assumption of constant correlations, it also estimates time‑varying conditional correlations between two indices and two cryptocurrencies using a DCC‑GARCH model. The results indicate that portfolios including three cryptocurrencies deliver superior risk‑adjusted returns compared with traditional portfolios and those holding only Bitcoin; they can also achieve higher returns, albeit with higher risk. The DCC‑GARCH analysis reveals pronounced time variation in correlations, ranging from negative to relatively high positive values, underscoring that diversification benefits can change substantially over time. The thesis discusses implications and limitations of the approach, proposes potential methodological extensions, and highlights behavioral finance factors in cryptocurrency markets that may lead to suboptimal investment decisions.
[This summary has been generated with the help of AI directly from the project (PDF)]
Documents