Portfolio Allocation for Cryptocurrencies: From Modern Portfolio Theory to Higher-Moment Utility A Comparative Analysis of Traditional, Tailor-Made, and Smart Beta Approaches

This thesis investigates portfolio allocation for cryptocurrencies under non-normal return distributions, moving beyond classical mean–variance optimization. Using daily data for major crypto assets from 2018–2024, and benchmarking correlations against representative equity and sovereign bond indices, we document distributional features—fat tails, skewness, volatility clustering—that challenge Gaussian assumptions. We evaluate four classes of allocation methods: (i) Traditional approaches based on Modern Portfolio Theory ; (ii) Tailor-made criteria that emphasize downside and tail risk ; (iii) Smart Beta strategies adapted to crypto; and (iv) utility-based optimization under higher moments following Jondeau’s framework, via a Taylor expansion of expected utility.