Previously during a podcast chat with Cathy Wood, certain managers stated that during the last 10 years, holding a 2% BTC and 98% cash portfolio would yield greater 4-year rolling risk-adjusted returns than just the S&P500, we did a deep dive, and find the sweet spot to be at 7% BTC and 93% cash. Whilst from a Sharpe and Sortino ratio perspective, holding 1% BTC and 99% cash is already sufficient in performing better than the S&P from a risk-adjusted angle, the actual annualised return is not practical (only 3.06%).
In this study, we try to find the BTC/ Cash allocation that matches the annualised return of S&P, whilst having less volatility and drawdowns.
Our research findings, for the decade between Jan 2014 till Jan 2024:
A portfolio with just 7% BTC, 93% cash money market allocation, roughly equals the return with 100% SPY allocation. Annualised Return: SPY (12.0%) vs 7% BTC (11.8%)
However, with significantly less volatility (standard deviation): SPY (15.1%) vs 7% BTC (11.4%)
Comparing best year, 7% BTC does significantly better: SPY (31.2%), 7% BTC (89.3%)
Comparing worst year, 7% BTC is also significantly safer: SPY (-18.2%), 7% BTC (-4.1%)
Maximum Drawdown, we can see 7% BTC offers 3x the protection: SPY (-23.9%), 7% BTC (-7.1%)
And lastly from downside risk adjusted return (sortino) perspective, the 7% BTC portfolio also does better: SPY (1.2), 7% BTC (3.4)
From an annual return perspective, a 7% BTC / 93% cash portfolio has 5 year of outperformance against a 100% SPY portfolio
A 7% BTC portfolio outperformed in years 2015, 2017, 2018, 2020, 2022, whilst a 100% SPY portfolio produced better return in years 2014, 2016, 2019, 2021, 2022
From a drawdown perspective, a 7% BTC /93% cash portfolio acts as a safer portfolio vs 100% SPY, as evidenced below:
If we were to accept similar volatility (standard deviation), that would mean choosing a 10% allocation to BTC, we can see the strong outperformance below:
With a 10% BTC/90% cash allocation, even though the volatility is similar to holding 100% SPY, the drawdown is still significantly lower: