Could God Make It In Crypto?

Confessions of a crypto hedge fund manager. Often, uncomfortably narrow is the line between a portfolio manager and a fortune teller. Both are in the business of forecasting the future, both expect remuneration for this task. Where they differ is that at least some portfolio managers are not charlatans, which is perhaps the best that we can say.

But what if you could actually see the future? Just how good of a portfolio manager could you be?

The past year in crypto has been a difficult but necessary one. There is a certain sad irony to the fact that in pursuing the illusion of familiarity by investing in crypto-adjacent companies, rather than crypto directly, so many brave and pioneering institutions actually compounded, rather than reduced, their risk.  And although the ghosts of FTX, Celsius, 3 Arrows Capital etc. will continue to haunt the space for years to come, the cleansing fire of the bear market will bring new and more deserving projects to the fore, leaving the entire space better off for it.

At Firinne Capital we’ve been running a fund for the past eleven months that has generated positive alpha net of fees for our investors over our benchmark. Nevertheless, we are always interested in how much better we could have done. Often, times of market turmoil are opportunities for the savvy investor, but at other times markets can become so pathological that the opportunities to generate capital gains dry up. Running some experiments and attempting to characterize the market through such a lens may yield some useful information.

Back to the original question, what if you could see the future, how well could you have possibly done? Surely your investors would be impressed by your godlike prowess. It’s an interesting experiment to perform, and one that was previously conducted by Wesley Gray back in 2016 on the US stock market and summarized in his article, “Even God would get fired as an Active Investor”. We thought it would be instructive to place an upper bound on just how well somebody could potentially do in the crypto markets, so let’s take a look.

Crypto is easy money?

The general public perception is that cryptocurrencies generated huge returns and easy money for many people. Whilst it’s true that early adopters witnessed a massive run-up in price, there is no guarantee they managed to hold through wild market swings through several orders of magnitude. Regardless, this cohort represents a tiny fraction of all crypto holders.

BTC’s history of severe drawdowns amidst great price appreciation during bull markets.

The violent boom-bust cycles of the crypto markets are almost perfectly designed for retail investors to lose money. Without a thesis for why they should be investing in crypto, the spikiness of returns means that uninformed investors are likely to buy and sell at the worse possible times. The most sensible option for most investors has been to reason about why they are investing and what they expect to gain, then hold for the long term.

To illustrate this point, let’s consider how three portfolio managers of widely differing abilities would have fared in the markets over recent years. Although we are embarking on a thought experiment, let’s inject a bit of reasonability by limiting our universe of assets to the 30 largest assets over the last 3 years, excluding stablecoins. Consider one manager with godlike prescience who manages to pick and hold the top quartile performing asset on a daily basis and repeat this consistently to the present day. Another manager represents mere mortals, and simply finds and holds the median performant asset, and finally a third ill-fated manager who buys and holds the bottom quartile asset on the market, represented by a chimpanzee. Note this is perhaps a little unfair on chimps, who would presumably select at random and therefore have an expected performance of the median.

As can be seen from the graph above, even the worse portfolio manager starting from the beginning 2020 could have made money in crypto, as long as they did nothing but hold almost any crypto asset. Breaking the analysis down on a year-by-year basis shows a more nuanced picture. The graph below shows the median asset performance, in the darker lines, for each of the three years, with 2023 included also. The shaded area around the lines represents the spread between the 75th quantile and the 25th quantile. Note that to be able to contrast the wildly different experiences across these three-plus years, the wealth is normalized to $1 at the beginning of each year and is plotted on a log-scale.

The main result here is that in 2021 and 2020, even a chimpanzee throwing darts at www.coingecko.com could have made money in crypto.

Notably, however, in 2022 not even God could have been successful in generating gains with a long only crypto portfolio.

Another important characteristic of the market is the shift in the dispersion of possible returns heading into 2022, one that we had previously detailed. As the correlation of all assets in the market tends towards 1, the opportunity set of assets with differentiated returns shrinks. If all assets essentially move in tandem, then the only possible bet is the choice to be in the market or not. Even accounting for the distorting effects of the log scale, it’s clear from the graphs that the opportunity set in 2021 was so great that a brilliant manager could have hoped to massively outperform the median. However, in 2022, the best managers were still quite close to the median, and the worst had many opportunities to do very poorly indeed, potentially losing nearly 90% of their wealth.

The perfect portfolio

Let’s now walk through the exercise of figuring out how to gain the best possible exposure to crypto.

Rather than just a single asset, now let us consider the portfolio of an omniscient being subjected to a few reasonable constraints that us mortal portfolio managers prefer to adhere by for the sake of risk management.

So, let’s start from the assumption that you’ve decided that you want an allocation in crypto. Ideally, it’s because you’ve concluded that a peer-to-peer electronic means of value transfer independent of arbitrary government intervention has merit. Or perhaps it’s just because it previously went up and made people rich, not as good a reason, but at the very least it has de-risked the bet somewhat in that you know explosive growth in the space is possible, something that early investors could not be sure of. Should you not agree that there is any merit in the space you may as well stop reading here; this article is not going to try convincing you otherwise. If you do believe, then the remaining question is whether now is the time, or to wait and see. From a seasoned investor’s perspective, the time to buy is when there is blood on the streets, and 2022 was about as bloody a year as there has been in the history of crypto. And if you assume there is value in the space with probably several orders of magnitude of growth still to occur if it is to realise its promise, then the question of whether or not to wait to get a marginally better price, at the risk of missing out on that expected growth seems like a suboptimal choice.

Obviously being able to time a market as volatile as crypto would lead to spectacular returns, but timing crypto would mean timing macro events like Fed rate hikes, inflation and wars. There are signals intrinsic to crypto which may allow for a portion of the portfolio to jump between cash and crypto but allocating a significant portion to such a strategy hardly makes sense conditioned on investor expectations of significant growth in the longer term.

Ruling out market timing, leaves two categories of long only or long biased options, namely VC or Liquid strategies.

Both types of funds will provide the desired beta exposure. Successful VC funds can return many multiples of the market; however this comes with the downside of added idiosyncratic risk. The market might rally but your VC fund invested in FTX, Luna and Celsius returns nothing. You may have gotten the big picture right but failed on the details. Crypto is risky as it is, and VC funds amp up that risk. The very best can do very well, while the rest generally don’t.

A simpler and lower risk alternative is to invest in a fund offering broad market exposure. Which leaves the remaining choice of who is going to manage your money. The first possible choice is, well, nobody at all, as you can simply buy a passive bitcoin or Ethereum ETF and let it ride. Probably not a bad choice, but you will be missing out on all the growth which may accrue to a whole class of assets beyond the big two. Previously, we have written about how inefficient the crypto markets are and therefore may be a good candidate for active management.

Next, assuming that you’ve chosen active management, the remaining question to be answered is whom will you choose to manage your money, what are the goals of their fund and what fees will they charge you for their services. Ideally the fund will only charge you performance fees on alpha earned in excess of a reasonable market benchmark, whilst at the same time offering you broad market exposure. A seemingly sensible but alas all too rare an offering. The remaining choice is what manager will you place your faith in? And what is the best that that manager could possibly do?

To explore this question let’s continue the work developed above. For the purposes of this thought experiment, you once again get to place your faith in God Himself to manage your portfolio. We want to calculate the Sharpe (a measure of risk-adjusted returns), Information Ratio (a measure of risk-adjusted active returns) and portfolio drawdowns that an omniscient portfolio manager – one that can literally see the future, would have returned.

So, let’s try to answer the age-old question, “could God make it in crypto?” Firstly, God would have steered well clear of Celsius, FTX et. al, probably due to their preexisting non-competes with the Devil.  This should not be underestimated as an achievement since many fund managers, sadly, did not.

Let’s then afford God the one-time opportunity to select the best performing assets over the period from Jan 2020 through May 2023. However, let’s apply some sensible constraints for the sake of making reasonable comparisons with us mere mortal fund managers. Obviously, an omniscient portfolio manager could have invested 100% of the assets into the Pepe meme coin for a 100x return, but this is not a realistic portfolio fulfilling the requirements of diversification and broad market exposure. So, let’s impose the liquidity constraint such that the investible universe is limited to assets which featured within the top 30-40 assets by market cap according to CoinGecko.

If we assume that the portfolio must resemble a market benchmark to some degree, then one choice of constraints is that the weight of the omniscient portfolio is allowed to deviate by either 100% of the benchmark weight or by 1%, whichever is larger.

Finally, If we further assume no rebalancing is allowed, we find the performance below.

Such a portfolio would have a modest Sharpe of 1.08, and IR of 0.62 and a max drawdown of 76%. Pretty shocking. We are forced to conclude that even God would have been fired in crypto, as no one is likely to keep a manager around with a 76% drawdown even though it had bested the market.

Nonetheless, it is instructive to see what assets this omniscient portfolio may be holding and what weights would be assigned to each asset. The chart below visually depicts the composition of this portfolio held over time. Broadly speaking, the reds represent active overweights while the cooler blues represent active underweights. God is allowed to make all the selections only once at the beginning of 2020 and the quantities are not rebalanced over the subsequent three years. The asset weights do drift relative to each other, reflecting the differential performance of each asset over time.

We can see that concentration in relatively few assets drove much of the outperformance. More particularly, the omniscient manager needs to be bold enough to invest in some of the mid-sized or smaller cryptoassets. Then again, having that perfect foresight can allow one to be bold.

Surely, a truly omniscient portfolio manager can do better than holding the same portfolio from inception to the present day. As a next step, we allow for rebalancing according to some prespecified periodicity such that the omniscient portfolio manager now has more opportunities to employ their supernatural skill. The plot below shows the active weights in a portfolio with the same thought experiment but allowing God to rebalance the portfolio at the beginning of each calendar year.

In fact, we can continue this experiment with increasing frequency of rebalances, i.e., allowing God to reselect the perfect set of assets for the subsequent holding periods. The outcome of doing so results in the plot below of Sharpe Ratios (Sharpe), Information Ratios (IR) and maximum drawdowns for the different rebalancing frequencies of this series of experiments.

Obviously if you were able to see one week ahead and rebalance accordingly, then you would have had a phenomenal IR and Sharpe. But as the rebalance frequency lengthens, we see a fast drop in these performance metrics from 2020 onwards, a symptom of the extreme volatility the crypto space.

Further breaking performance down annually, we arrive at the charts below.

The omniscient manager would have had very healthy Sharpes and IRs in 2020 and 2021 but would have still experienced some stomach-churning drawdowns. However, the brutality of 2022 is placed in strong contrast, with nothing short of being able to see one week ahead good enough to avoid a negative total return. For example, even being able to see 90 days ahead still yielded quite a negative performance, as the chart below shows.

Conclusion

By introducing a portfolio manager with ability to see the future, we are able to analyze the impact of varying degrees of skill on common performance metrics and place some upper bounds on their values. Applying this analysis over the three-plus years from 2020, which encompasses the strong bull market of 2021 and the severe bear market of 2022, we observe a considerable divergence of outcomes.

Choice of manager is critical. Spectacular returns in 2021 bull markets don’t really say a lot about skill. Assessments of performance ought to include performance during severe market conditions such as 2022. Exceptional performance can often mask hidden unacceptable risks, e.g., FTX.

Exceptional skill doesn’t necessarily translate into exceptional Sharpe or IR in crypto and exceptional performance doesn’t necessarily imply exceptional skill. Investors in crypto need to keep sight of the potential upside of their investments and not be shaken out by volatility. The conclusion that even God would have been fired as a portfolio manager in the equity markets is even truer in crypto.

One possibility to potentially smoothen out the market volatility and boost the performance of these portfolios is to add some market neutral strategies. We will explore these in a future post.