To include or not to include, that is the question

Published in Insights on 14th January 2021 by Pierre E. Mendelsohn

Would you invest in an asset which quadrupled in value last year? Further, what if I told you it produced 230% p.a. returns on average since 2010 (before fees), lost 28% from its peak this year before recovering somewhat, already experienced several violent bubbles in the last ten years, that its annualised historical volatility is 102%, and maximum drawdown 93%. Not to mention, of course, that it is a digital asset described by TV host John Oliver as “everything you don’t understand about money, combined with everything you don’t understand about computers”.

Well, many people have, and a growing number of institutional investors seem to be taking bitcoin (BTC) seriously, even though its recent price action evokes yet another bubble bursting, and the
FCA is on the record saying that “crypto investors should be prepared to lose all their money”.

While the case for buying BTC on its own may look questionable, the picture is more nuanced within a portfolio context. The growth in BTC interest in the last year or so was fuelled in part by a growing number of institutional investors choosing to allocate a percentage of their broader portfolios to BTC, so much so that there is talk of mainstream adoption as gold 2.0, a digital version of gold. To some, the recent surge is yet another bubble while, to others we are in still the early days of a price discovery process for a radically new kind of asset class.

In a market running high on emotions and ephemeral news, we thought it would be helpful to look at a few data points, especially around what systematic allocation engines are telling us about BTC allocation in the context of multi-asset portfolios. To this end, we asked the following questions:

Firstly, assuming you have no view on the future of BTC, how much would a risk-based model suggest allocating to BTC?

Using a USD-denominated basket of ten highly liquid sub-asset classes and adding BTC to it, the answer is between 0% and 2.6% depending on the allocation engine one choses, with an historical range of 0-7%.

Chart 1.1

Of course, factoring in views on BTC’s future performance can produce very different results. The MDP allocation engine, for example, allocates zero to BTC if BTC is forecast to fall significantly from current levels.

While the Minimum Variance logic results in a zero BTC allocation today, it did contain some small allocation in the past. This suggests that BTC, despite its extreme volatility, brought sufficient diversification benefit to warrant an allocation in the portfolio, albeit below 0.5%, a few times in the last nine years.

Chart 2

Secondly, what would a small BTC allocation have done to a multi-asset portfolio’s risks and returns?

BTC’s numbers are so extreme that even a small, intermittent allocation produced noticeable results.

For example, here is the MDP portfolio with and without BTC.

Cahrt 3Cahrt 3.1

A few observations stand out:

  • Historical results are very dependent on the period chosen, with the best results produced in the last 5Y, and the worst results in the last 3Y.
  • While BTC historically increased the simplified Sharpe ratio of the portfolio, it also increased drawdowns by 1.14%, i.e. 14% on a relative basis, a risk some investors will not be prepared to accept.

It is important to note that this analysis is based on the XBTUSD index. It does not include transaction and management fees, which can be substantial.

This does not alter the key point, which is that even a small BTC allocation caused noticeable differences in historical returns. A BTC allocation in the portfolio helped increase the risk /return trade-off in the portfolio, especially so in the last five years or more, but it also increased drawdowns.   

History being no guide to future performance, the likelihood that adding BTC to the portfolio will result in outperformance going forward is, of course, far from assured.

Tools such as forward-looking Monte Carlo simulations can help provide additional colour as to the impact of BTC on the future distribution of portfolio returns. Critical points to consider when doing such work include the return distribution and correlation assumptions for BTC. This leads to the third question:

Thirdly, what is the correlation of BTC with other assets, and how did it change over time?

This question is important as it relates to the diversification benefit BTC may bring to multi-asset portfolios.       

Since July 2012, BTC has displayed a small, negative correlation to US treasuries across the yield curve, and a very small positive correlation to equities and other sub-asset classes, USD corporates and TIPs included. This can be seen in the bottom half of the chart below (blue rectangles).

Over the last six months, BTC has shown similarly low negative correlations to US treasuries, but notably higher correlations to other assets, with a 32% correlation to gold and a 27% correlation to US equities, arguably making it a less effective diversifier that in years past. This can be seen in the top half of the chart below (red rectangles).

Chart 4

A daily performance contribution chart over the last year corroborates this point. The days in the last twelve months when BTC had the worst performance were also down days for US and international equities. This is why BTC increased, rather than decreased, drawdowns in the portfolio. Bitcoin’s behaviour during the COVID-induced period of extreme market turbulence in March and April 20 is the culprit.

Chart 5

Conclusion

The reputational risks of investing in BTC, or at least the perception thereof, seem to have subsided somewhat in recent years, but new risks have emerged along with increased regulatory scrutiny. Maturation and a sharp rise in total market capitalisation are the most obvious differences this time around compared to previous surges.

A further BTC implosion is entirely possible from here, which would add this episode to the list of
previous BTC bubbles, and cause BTC newcomers and some short-term speculators to be burnt.

For longer-term investors and asset allocators, the story is different. Risk-based asset allocation models do not allocate to BTC because of its recent performance and abundant news feed. They do so because, historically, BTC was an uncorrelated asset class which brought a measurable diversification benefit to multi-asset portfolios.

Going forward, BTC will attract further interest only if (a) it does not self-destruct, (b) it can remain uncorrelated to traditional assets and, therefore, continue to bring an observable diversification benefit to multi-asset portfolios.  If so, and this is not a small if, we would expect it to gradually become less volatile as it becomes more widely accepted.

The initial growth of BTC was fuelled by short-term investors and speculators. Its long-term success from here depends on its durable adoption by longer-term investors and asset allocators. 


Please contact us at info@alpima.com to find out more about ALPIMA, and how we help professional investors design, tune, analyse, implement and manage customised investment strategies.


 

 

Share this article

Copy link
Link copied to clipboard