Eliminate the sting of cryptocurrencies

Introduction

Faced with structurally slower long-term economic growth cause by aging populations, shrinking labor force and weak capital formation and productivity – as well as the impact of the pandemic – central bank balance sheets and government borrowing exploded. The result has been to undermine national currencies through inflation and the potential for competitive currency depreciation. In this scenario, decentralized alternative currencies, free from political manipulation – such as cryptocurrencies – have become increasingly attractive.

However, cryptocurrencies are still a developing asset class and are extremely volatile. This note will show how – using an appropriate trading strategy – one can significantly reduce the volatility of cryptocurrencies while simultaneously improving their returns.

Cryptocurrency performance

The historical returns cryptocurrencies have been staggering. From the beginning of 2018 until May 20, 2022, Bitcoin and Ethereum, for example, were up almost 250% and 550%, respectively, compared to 48% for the S&P 500 (see chart and table below). However, while returns have been stellar, cryptocurrencies have also experienced tremendous volatility – in 2018, Bitcoin fell over 70%.

Cryptocurrency modeling

Unlike stocks, cryptocurrencies are more difficult to value because they do not produce profits or dividends. However, one might still be able to evaluate cryptocurrencies against macroeconomic and market factors if it can be established that there is a meaningful long-term relationship. If so, one can calibrate these relationships to find a pattern that best fits the historical price change of cryptocurrencies, which can then be used to minimize volatility and maximize returns.

A number of factors come to mind that can explain the price variation of cryptocurrencies. Gold is one, as it is also considered an alternative currency. Bonds are another, as a rising yield becomes more attractive relative to another asset that does not generate any income. Stocks could also be an important factor, as a rising stock market reflects growing risk appetite, which is likely to spill over into the cryptocurrency market. The dollar is probably one of the most important factors, since the raison d’être of cryptocurrencies is the fear of currency depreciation. Finally, the VIX volatility index could be important, as cryptocurrencies could act as a hedge against uncertainty.

These factors can be used in a multiple regression which estimates a line of best fit between the price of the cryptocurrency and the factors. The slope of this line on each factor is the beta factor, which gives a measure of the factor’s influence on the cryptocurrency. We can combine these betas to estimate a model price for the cryptocurrency. The charts below compare the actual (log) price of Bitcoin, Ethereum, and Litecoin, with their respective model prices. We can see that over the past five years, the price of the model matches the actual data very closely, with a “goodness of fit” of 0.89 (the maximum being 1.0) for Bitcoin, 0.84 for Ethereum and 0.67 for Litecoin.

Factor contributions

Of the factors used in the model, which are the most important? We can see in the chart below that the largest beta is in the dollar, followed by the S&P500 and bond yields.

However, no matter how big the beta is, if the factor doesn’t move much, it won’t have much influence on the model. So you have to look at how much the factor has also moved, and the combination of the two will give an idea of ​​the contribution of each factor over a given period of time. The following two charts show this, and it can be seen that in the twelve weeks leading up to May 20, 2022, the most important factors that contributed to the price movement of the Bitcoin model, for example, were the yield of bonds of company, the S&P500 and the Dollar.

Cryptocurrency trading strategy

Since there is a proximity long term relationship between the actual price and the model price of the cryptocurrency, then, in theory, any divergence between the two should only be temporary, and the deviation between them should revert to the mean. This is indeed the case as can be seen in the graphs below.

Since spreads are mean reverting, if the actual price is above the pattern price, the cryptocurrency is overvalued and if the actual price is below the pattern price, the cryptocurrency is undervalued. A strategy then presents itself: if the actual price is excessively higher than the model price, sell the cryptocurrency; and if the actual price is excessively lower than the pattern price, buy the cryptocurrency. Trades are settled when the spread reverts to the mean.

Cryptocurrency Strategy Performance

The results of this strategy are very promising, as shown in the chart and table below. An equally weighted portfolio of the three cryptocurrency indices, over the period from January 1, 2018 to May 20, 2022, would have produced an annualized return of 25.9% with an annualized standard deviation of 68.7% (volatility). On the other hand, an equally weighted portfolio of the three cryptocurrencies strategies, over the same period, would have produced an annualized return of 68% with an annualized volatility of 31.7%. Thus, the cryptocurrency strategy would have cut cryptocurrency volatility by more than half and more than doubled cryptocurrency returns in the process.

Conclusion

In a world of policy-induced currency depreciation, decentralized digital assets such as cryptocurrencies are increasingly attractive. But they come with significant volatility. One way to minimize this volatility while improving returns is to find macroeconomic and market factors that influence cryptocurrency prices and trade the relationship between them. The results are very attractive.

Thanks for reading my note; I hope you found it interesting.

Data source: Federal Reserve. The results are estimates.