5 min de lectureAs growing numbers of investors take Bitcoin more seriously, it raises the question: what is an appropriate amount of Bitcoin in a portfolio? The real answer depends on the investor’s appetite for risk and their portfolio investment style, but here we attempt to answer this question for an institutional multi-asset investor with what we believe to be a moderate appetite for risk.
A modest increase in an investment portfolio risk budget of 100 basis points would suggest a portfolio weighting of 3.6%, with a 120 basis point risk budget suggesting a 4% weight.
Varied back-test periods highlight a consistency in results suggesting the subjectivity in cherry-picking specific time periods isn’t the concern many believe it to be.
Bitcoin being part of a relatively new asset class, where volatility is likely to decline as acceptance and understanding of it proliferates, implies both its future volatility and historical volatility should be taken into consideration when calculating potential portfolio weightings.
Portfolio Allocation Styles
There are varied ways in which an investor may choose to allocate a portion of their portfolio to Bitcoin, although most use varied forms of historic volatility to determine allocation sizes. Alpima has published compelling analysis on understanding how much to allocate to Bitcoin based on differing portfolio styles.
Their results highlight that a maximum diversification portfolio would have the highest allocation due to Bitcoin’s low correlation to other asset classes. We highlighted similar results in our previous work Bitcoin’s role in an investment portfolio. However, applying no constraints to a maximum diversification portfolio would lead to very large weighting in bitcoin, resulting in very high volatility. Understandably, due to bitcoin’s high volatility, a minimal variance portfolio would allocate nothing. But there are challenges in using historical volatility in that choosing different date ranges can lead to very different outcomes with the data range chosen potentially being extremely subjective.
There are other approaches, such as the Black-Litterman model, which requires the investor to input their assumed growth rate and confidence in that prediction, and then apply mean-variance optimisation (MVO) to maximise expected returns and suggest an allocation weight. The problem with this approach is that a high confidence in your prediction is required to allocate in meaningful quantities, this is particularly difficult given how young and volatile Bitcoin is. However, Adam Grealish of Betterment has done some interesting work on applying the Black-Litterman model to Bitcoin.
The Black-Litterman approach highlights that high expectations of outperformance (+40% versus equities) coupled with a confidence level of 50% would suggest a 6% allocation. Unfortunately, we believe it is unreasonable to expect that level of outperformance every year, and furthermore, at lower confidence levels the suggested allocations fall rapidly. At a 5% expectation of outperformance, the suggested allocation at any confidence level is negligible.
Coinshares’ Risk Budgeting Approach
To ascertain how much bitcoin would be appropriate in a portfolio we created a database of daily returns starting from October 2015 when it was first financialised (available as an ETP). We created a traditional balanced portfolio with 60% equities and 40% bonds and then added varied amounts of bitcoin, detracting from both equities and bonds equally (see our more detailed work here).
Previous analysis highlights keeping a fixed portfolio allocation and rebalancing on a quarterly basis helps mitigate Bitcoin’s volatility challenges with only moderate detriment to returns. Rebalancing has also helped improve risk-adjusted returns and have significantly reduced maximum drawdowns.
Being acutely aware that investors have varied appetites for risk, our model portfolio can give guidance on volatility targeting. As indicated in the chart below, when looking at data since October 2015, an additional 100 basis points (1%) of annualised risk suggests a portfolio weight of 3.6%.
The chart highlights a fairly linear path at levels past 4% requiring an additional 120 (1.20%) of annualised risk, with a 200 basis point addition to the volatility budget suggesting a 5% allocation.
A fair critique of this analysis would be our back-test period taking a favourable period of time for Bitcoin. However, we are encouraged by how little the portfolio risk changes when the back-test periods are varied. The chart below highlights how varying the back-test length from 5 years through to just 1 year for a 4% Bitcoin weighting creates a variance of just 45 basis points around the longer-term 100 basis point increase in volatility.
We also witnessed a big decline in the additional volatility in late 2019, which we believe was due to an increase in volatility of equities and bonds rather than bitcoin. It demonstrates that if there are significant volatility spikes in equities and bonds, which is quite plausible given their high valuations, that bitcoin would look increasingly attractive from a risk perspective.
Applying a fixed volatility budget of 100 basis points, instead of a weighting target of 4% bitcoin highlights how little the model portfolio’s suggested weighting varies over the varied back-test periods. The model’s suggested weightings vary from 3.2% to 5.1%.
We are encouraged by the consistent results from varied back-test periods when trying to ascertain what would be an optimal Bitcoin weighting in a portfolio. A modest increase in an investment portfolio risk budget of 100 basis points would suggest a portfolio weighting of 3.6%, but we appreciate investors’ appetite for risk is subjective.
Bitcoin being part of a relatively new asset class, where volatility and returns are both likely to decline as acceptance and understanding of it proliferates, implies both its expected future volatility and historical volatility should be taken into consideration when calculating potential portfolio weightings. Current trends suggest bitcoin’s volatility will continue to decline overtime, and therefore implied allocations in bitcoin are likely to increase too.
