Skip to content
All library documents

Estimating Bitcoin Portfolio Weights with Bootstrapped Return Distributions

Article Systematic trading blog (Rob Carver)
Read the original

Summary

The author questions whether Bitcoin’s positive skew alone justifies very large portfolio allocations, using a published allocation claim as a starting point. The post compares the intuition behind holding Bitcoin with the appeal of lottery-like payoffs, then frames allocation as a joint result of return distributions and investor preferences. It uses historical daily returns for Bitcoin, equities, and bonds, with the equity and bond exposures represented by futures and Bitcoin returns adjusted for short-term Treasury yields.

Instead of choosing a utility function, the author proposes maximizing a selected percentile of bootstrapped geometric returns. Repeated samples with replacement create distributions of possible portfolio growth outcomes; the chosen percentile represents differing tolerance for uncertainty. The author argues this differs from mean-variance optimization and tends to favor crypto more when maximizing compound growth. Summary statistics are provided for returns, volatility, correlations, Sharpe ratios, and skew, but the supplied text ends before presenting the allocation results. The approach therefore remains a proposed empirical framework here, with sampling uncertainty and the inability to separate skew preference from risk preference acknowledged as limitations.

Key ideas

  • The post examines how return skew and investor preferences may affect Bitcoin portfolio weights.
  • It proposes bootstrapping historical returns to estimate distributions of portfolio geometric returns.
  • The investor’s uncertainty preference is represented by the percentile of geometric returns being optimized.
  • The author notes that maximizing compound growth can produce larger crypto allocations than mean-variance approaches.
  • The excerpt supplies descriptive statistics but omits the allocation results, so its empirical conclusions are incomplete.

Tags

CryptoPortfolio constructionBacktestingStatisticsRisk management

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.