Do cryptocurrencies make sense in asset allocation?

An increasing number of institutional investors ha ve expressed interest in cryptocurrencies and envisage allocating a proportion of their portfolios to these rather unconventional assets. Amongst the most attractive characteristics presented by these assets, their decorrelation to traditional asset classes and potential returns are often cited, while obstacles to investment for institutional investors are of course the excessive volatility displayed by cryptocurrencies as well as the absence of fundamentals and fair value models. We observe first of all that , even though cryptocurrencies are decorrelated from traditional assets, their potential as a hedge during major drawdowns is not demonstrated by an examination of past events. Then, it is clear that, when considering the relation between volatility and returns, cryptocurrencies and traditional assets are miles apart: where some see excessive realised volatility, others see a potential for earning decent returns. In reality, their risk-adjusted return has converged for some time now towards l evels for traditional asset classes , exception made of the exuberance displayed during several spells. From the standpoint of the portfolio theory, the introduction of cryptocurrencies in a multi-asset portfolio would appear to le a d to a significant improvement in theoretical asset allocation efficiency. We have attempted to replicate a realistic, but simplified analysis framework for a portfolio manager wanting to add cryptocurrencies to a portfolio, using first a naïve allocation method, then applying optimisation methods commonly used in the industry. At first sight, the results of the backtests appear promising: there is an improvement in Sharpe ratios, while risk indicators (drawdown, CVaR , etc.) appear unchanged, even better as a result of adding cryptocurrencies to portfolio. Allocating 0.5% to 1% to cryptocurrencies therefore appears to procure benefits to asset managers looking to diversify their portfolios without being exposed to significant drawdowns. However, caution remains in order over the addition of cryptocurrencies, as backtests were performed on a rather singular test sample: 2020 was a year when cryptocurrencies performed much better than traditional assets. Furthermore, their contribution to portfolio risk is far too significant when not monitored ex-ante.