Crypto in Diversified Portfolios
Michael Zhao
Last Updated: 1/14/2025 | 28 min. read
Crypto is a unique, high-volatility alternative asset class that may help improve risk-adjusted returns when added to traditional portfolios.
When incorporating crypto into portfolios, a little goes a long way: conventional portfolio optimization techniques suggest an allocation of about 5% could maximize expected risk-adjusted returns. However, allocating to crypto will also likely increase portfolio risk.
An allocation to crypto could potentially substitute for other assets used for return enhancement and/or diversification, including gold and other commodities, small-cap equities and international stocks, or technology shares. The crypto asset class could also potentially replace certain private investments where it would offer more liquidity.
Investors should consider their own circumstances and financial goals before investing in crypto. The asset class should be considered high risk and may not be suitable for investors with near-term capital needs and/or high-risk aversion.
Building a well-diversified portfolio has gotten harder, due to more narrowly concentrated returns in traditional assets, changing correlations, and macro risks. At Grayscale, we believe that the crypto asset class can be one way to help address these challenges. The crypto asset class includes a wide range of technologies with many specific use cases. For example, Bitcoin is a monetary system while Aave is a lending protocol and Bittensor is a platform for building open-source AI. This diversity raises questions for asset allocators: Should crypto assets be thought of as commodities, technology investments, or something else when considering an allocation in investment portfolios?
Grayscale Research believes that crypto should be considered a unique, high-volatility alternative asset class that may help improve risk-adjusted returns when incorporated into traditional portfolios. Although the individual assets are diverse, they are all based on the same underlying technology — the public blockchain — and most can be considered early-stage investments. The main difference with other early-stage investment options like venture capital is crypto’s market structure: blockchain-based assets are liquid and traded globally 24/7, including by many retail investors. These features — early-stage investments traded in liquid public markets and available to anyone — result in a set of statistical properties that can be used to inform portfolio construction.
Behavior of the Crypto Asset Class
1) High volatility. Crypto is a uniquely volatile major asset class. Since early 2017, a market cap-weighted index of investible digital assets[2] has produced annualized volatility of 95%; Bitcoin’s volatility over the same period was 83%. This compares to annualized volatility of 16% for the S&P 500 Index and 19% for the Nasdaq 100 Index. Crypto’s volatility is more comparable to certain high-price-volatility commodities like natural gas or investment strategies that incorporate leverage (Exhibit 1). The crypto asset class should therefore be considered high risk, and this is the first consideration investors should keep in mind before incorporating crypto assets into a portfolio.
Exhibit 1: Crypto is a uniquely volatile major asset class
2) Returns commensurate with volatility. At the same time, Bitcoin and digital assets as a whole (e.g. based on market cap-weighted indexes) have produced returns commensurate with their higher volatility. In other words, crypto assets have compensated investors for the higher risk with higher returns. For example, since early 2017, the same market cap-weighted index of investible crypto assets has produced an annualized return of 69%. Given annualized volatility of 95%, the ratio of return to volatility was about 0.7. Over the same period, the S&P 500 Index produced annualized returns of 14% on volatility of 16%, for a ratio of return to volatility of about 0.9.[3] Over periods of ~10yrs or more, major asset classes have tended to see a ratio of return to volatility in the range of 0.5-1.0, and crypto returns fall within this range (Exhibit 2). Therefore, accounting for its higher volatility, crypto’s return profile is comparable to what we observe for other asset classes.
Exhibit 2: Ratio of return to volatility for crypto similar to other asset classes
3) Low correlation with traditional assets. Higher volatility and higher returns can be achieved with leverage — by borrowing money to invest more capital. However, adding leverage to traditional assets does not affect their correlations, and therefore cannot be used to improve risk-adjusted returns.[4] The crypto asset class is unique in providing high risk and high potential returns while at the same time displaying a low correlation to traditional assets (Exhibit 3). Lower correlations are the “secret sauce” of portfolio diversification, and why data-driven approaches to portfolio optimization will typically see a role for incorporating the crypto asset class.
Exhibit 3: Historically low correlations between crypto and equity market returns
4) High momentum. Compared to stocks and bonds, historical returns for the crypto asset class exhibit relatively high momentum: gains tend to follow gains and losses tend to follow losses. In this way, digital asset returns are more similar to returns for currency and commodity assets, which also exhibit price momentum. The combination of high volatility and high momentum means that the crypto asset class can occasionally produce concentrated periods of large gains, as well as large drawdowns. Investors incorporating crypto into diversified portfolios could potentially consider using trend-following strategies for risk management (for more detail, see The Trend is Your Friend: Managing Bitcoin’s Volatility with Momentum Signals).
Exhibit 4: Bitcoin’s volatility has declined over time
A Little Goes a Long Way
Exhibit 5 shows the results for our baseline simulations. As Bitcoin is added to the classic 60/40 portfolio in small increments, the expected Sharpe Ratio initially rises. The reason is that, although Bitcoin is a volatile asset, it may offer a high return and a lower correlation to traditional assets, and therefore diversification benefits. The Sharpe Ratio continues to rise until Bitcoin reaches approximately a 5% share of the total portfolio, then begins to level off. After that point, increasing the Bitcoin allocation is no longer expected to improve risk-adjusted returns.
Exhibit 5: Adding Bitcoin to traditional portfolio may increase risk-adjusted returns
Because crypto is a volatile asset class, a little goes a long way: even a modest allocation to Bitcoin and/or other digital assets can meaningfully improve both total returns and risk-adjusted returns. For example, Exhibit 6 shows the hypothetical ex-post impact of adding Bitcoin to a traditional 60/40 portfolio in increments of 1%, 3%, and 5%. These results are strictly hypothetical and do not reflect actual results achieved by an investor. When adding Bitcoin to a 60/40 portfolio in these amounts, annualized returns increase by 0.6-3.3 percentage points (pp) over this period, and portfolio Sharpe Ratio increased by 0.05-0.22pp. Although total returns and risk-adjusted returns both increase, it is important to stress that portfolio risk, including maximum drawdown, also increases.
Exhibit 6: Hypothetical impact of 1%-5% Bitcoin allocation on 60/40 portfolio
HYPOTHETICAL SIMULATED PERFORMANCE RESULTS HAVE CERTAIN INHERENT LIMITATIONS. There is no guarantee that the market conditions during the past period will be present in the future. Rather, it is most likely that the future market conditions will differ significantly from those of this past period, which could have a materially adverse impact on future returns. Unlike an actual performance record, simulated results do not represent actual trading or the costs of managing the portfolio. Also, since the trades have not actually been executed, the results may have under or overcompensated for the impact, if any, of certain market factors, such as lack of liquidity. Simulated trading programs in general are also subject to the fact that they are designed with the benefit of hindsight. NO REPRESENTATION IS BEING MADE THAT ANY ACCOUNT WILL OR IS LIKELY TO ACHIEVE PROFITS OR LOSSES SIMILAR TO THOSE SHOWN. PAST PERFORMANCE IS NOT INDICATIVE OF FUTURE RESULTS. We selected the timeframe for our analysis because we believe it broadly constitutes the most complete historical dataset for the digital assets that we have chosen to analyze. For additional risk disclosure pertaining to hypothetical performance see end disclosures.
Naturally, this is just one simulation under a specific set of assumptions, and there is no guarantee that future returns will mirror those of the past. Investors should consider a range of possible outcomes for crypto asset returns before allocating capital. Although an allocation of about 5% may maximize risk-adjusted returns given historical results, a lower amount could be appropriate if, for example, Bitcoin’s returns decline in the future or its correlation with other asset classes increases.
Common Implementation Questions
Although traditional portfolio optimization techniques can be a helpful starting point, investors allocating to the crypto asset class often need to address a variety of practical implementation issues. Here are a few of the most common questions we receive from investors about incorporating crypto into diversified portfolios:
2) When adding to crypto, where in my existing portfolio should I subtract from? In a diversified portfolio, the crypto asset class may provide return enhancement, diversification, or both. It could therefore potentially substitute for existing allocations designed to provide these attributes, including gold and other commodities, small-cap equities and international stocks, or technology shares. Crypto could also potentially substitute for certain private investments (e.g., private equity, venture capital, or infrastructure) where it would also offer more portfolio liquidity. The crypto asset class will likely offer higher risk and higher potential return than other asset classes, potentially contributing to capital efficiency for the portfolio as a whole.
Exhibit 7: Bitcoin has experienced four major drawdowns in its 16-year history
4) Should I consider Bitcoin a defensive asset? Grayscale Research believes that Bitcoin can be considered a store of value: we expect it to maintain its value in real terms (i.e., accounting for inflation) over the long term because it is scarce and not dependent on any specific institution. However, Bitcoin today should not be considered a defensive asset that could dampen portfolio volatility during economic recessions or other periods of heightened risk aversion. Bitcoin is a high-volatility asset that is positively correlated to equity markets and typically declines when risk aversion rises. Moreover, at least currently, gold and short-term U.S. Treasury bonds tend to outperform Bitcoin during increases in geopolitical risk.
The Role of Crypto
It is important to stress that investing in digital assets may not be suitable for everyone. In this analysis we considered a representative investor holding a classic 60/40 portfolio. However, in practice there are many types of portfolios serving different financial needs. For example, some investors hold portfolios of low-volatility assets because their capital is earmarked for upcoming expenses (e.g., home purchase or college tuition). Crypto is a high-volatility asset class which may produce strong returns over time, but which can also lose significant value over the short term. Additionally, some investors favor income-producing assets like fixed income securities or dividend-paying stocks. Although some crypto assets generate income, in most cases the earnings will be low compared to the asset’s volatility — crypto is an asset class primarily held for capital appreciation.
While there are some exceptions, Grayscale Research analysis suggests that a traditional balanced portfolio may achieve higher risk-adjusted returns with a moderate allocation to crypto — perhaps ~5% of total financial assets. Because crypto is a high-risk/high-return-potential asset class with a low correlation to stocks, crypto assets can potentially help investors overcome some of the portfolio construction challenges they face today. Allocating crypto does not change other conventional thinking about portfolio construction, including to reduce portfolio volatility when approaching retirement, to use tax-advantaged accounts when possible, and to avoid trying to time the market.
To determine the appropriate share of Bitcoin in a portfolio we simulated expected returns using Monte Carlo methods. Specifically, for each incremental addition of Bitcoin to a 60/40 portfolio, ranging from 0% to 20%, we simulated 1,000 random portfolios over 60 month periods (i.e. 26,000 five-year horizon simulations). We then took an average of the return, volatility, and Sharpe Ratio for each Bitcoin share of the portfolio (e.g. the average Sharpe Ratio when Bitcoin’s share of the portfolio was 0%, 1%, 2%, and so on).
A critical element of this analysis was considering the implications of crypto’s correlation with other asset classes—as this determines its diversification benefit in a portfolio context. To adjust the correlations between Treasury bonds, the S&P 500, and Bitcoin, we employed the Iman-Conover method. This method is useful for its ability to alter the correlation structure of a set of variables while preserving their individual marginal distributions.
The key steps involved in this method are:
Rank Transformation: Each variable's values are replaced with their ranks, transforming the data into a uniform distribution.
Normalization and Conversion to Z-Scores: These ranks are normalized and then converted into z-scores, aligning them with a standard normal distribution.
Correlation Adjustment: We then adjust these z-scores using the Cholesky decomposition of our target correlation matrix, aligning the data with our desired correlation structure.
Transformation Back to Original Scale: Finally, the adjusted z-scores are transformed back to the original scale of the data, ensuring that the original distribution characteristics of each variable are maintained.
The Iman-Conover is helpful for our analysis as it allows for a realistic simulation of how different assets, especially Bitcoin, interact within a portfolio under various correlation scenarios. For our baseline statistics, we selected Bitcoin returns data post-2014. This choice is based on the observation that around 2014, the distribution of Bitcoin returns became more stable and approximated a normal distribution. For stocks and bonds, the baseline distributions were measured since 1980.
Note On Hypothetical Simulated Performance Results
HYPOTHETICAL SIMULATED PERFORMANCE RESULTS HAVE CERTAIN INHERENT LIMITATIONS. There is no guarantee that the market conditions during the past period will be present in the future. Rather, it is most likely that the future market conditions will differ significantly from those of this past period, which could have a materially adverse impact on future returns. Unlike an actual performance record, simulated results do not represent actual trading or the costs of managing the portfolio. Also, since the trades have not actually been executed, the results may have under or over-compensated for the impact, if any, of certain market factors, such as lack of liquidity. Simulated trading programs in general are also subject to the fact that they are designed with the benefit of hindsight. NO REPRESENTATION IS BEING MADE THAT ANY ACCOUNT WILL OR IS LIKELY TO ACHIEVE PROFITS OR LOSSES SIMILAR TO THOSE SHOWN. PAST PERFORMANCE IS NOT INDICATIVE OF FUTURE RESULTS.
The hypothetical simulated performance results are based on a model that used inputs that are based on assumptions about a variety of conditions and events and provides hypothetical, not actual results. As with all mathematical models, results may vary significantly depending upon the value of the inputs given, so that a relatively minor modification of any assumption may have a significant impact on the result. Among other things, the hypothetical simulated performance calculations do not take into account all aspects of the applicable asset’s characteristics under certain conditions, including characteristics that can have a significant impact on the results.
Further, in evaluating the hypothetical simulated performance results herein, each prospective investor should understand that not all of the hypothetical assumptions used in the model are described herein, and conditions and events that are not accounted for by the model may have a significant adverse effect on the performance of the assets described herein.
Prospective investors should consider whether the behavior of these assets should be tested based on different and/or additional assumptions from those included in the information herein.
IN ADDITION TO OTHER DIFFERENCES, PROSPECTIVE INVESTORS IN A PRODUCT SHOULD NOTE THE FOLLOWING POTENTIALLY SIGNIFICANT DIFFERENCES BETWEEN THE ASSUMPTIONS MADE IN THE HYPOTHETICAL SIMULATED PERFORMANCE RESULTS INCLUDED HEREIN AND THE CONDITIONS UNDER WHICH A PRODUCT WILL PERFORM, WHICH COULD CAUSE THE ACTUAL RETURN OF SUCH PRODUCT TO DIFFER CONSIDERABLY FROM RETURNS SET FORTH BY THE HYPOTHETICAL SIMULATED PERFORMANCE, TO BE MATERIALLY LOWER THAN THE RETURNS AND TO RESULT IN LOSSES OF SOME OR ALL OF THE INVESTMENT BY PROSPECTIVE INVESTORS:
IN ADDITION, THE HYPOTHETICAL SIMULATED PERFORMANCE RESULTS DO NOT ASSUME ANY GAINS OR LOSSES FROM TRADING AND THEREFORE DO NOT REFLECT THE POTENTIAL LOSSES, COSTS AND RISKS POSED BY TRADING AND HOLDING ACTUAL ASSETS. The hypothetical simulated performance results do not reflect the impact the market conditions may have had upon a portfolio were it in existence during the historical period selected. The hypothetical simulated performance results do not reflect any fees incurred by a portfolio. If such amounts had been included in the hypothetical simulated performance, the results would have been lowered.
AS A RESULT OF THESE AND OTHER DIFFERENCES, THE ACTUAL RETURNS OF A PORTFOLIO MAY BE HIGHER OR LOWER THAN THE RETURNS SET FORTH IN THE HYPOTHETICAL SIMULATED PERFORMANCE RESULTS, WHICH ARE HYPOTHETICAL AND MAY NEVER BE ACHIEVED. Reasons for a deviation may also include, but are by no means limited to, changes in regulatory and/or tax law, and generally unfavorable market conditions.
Index Definitions:
The Bloomberg US Treasury Index measures US dollar-denominated, fixed-rate, nominal debt issued by the US Treasury. The FTSE/Grayscale Crypto Sectors (CSMI) measures the price return of digital assets listed on major global exchanges. The Bloomberg Barclays US Corporate Bond Index measures the investment grade, fixed-rate, taxable corporate bond market. The S&P Cryptocurrency Broad Digital Market (BDM) Index is designed to track the performance of digital assets listed on recognized open digital exchanges that meet minimum liquidity and market capitalization criteria. The MSCI All Country World Index (ACWI) is a global equity index that measures the equity performance in both the developed and emerging markets. The Bloomberg US Corporate High Yield Bond Index measures the USD-denominated, high yield, fixed-rate corporate bond market. The S&P GSCI is an index of 24 exchange-traded futures contracts that represent a large portion of the global commodities market. The S&P 500 Total Return Index is a total return index that reflects both changes in the prices of stocks in the S&P 500 Index as well as the reinvestment of the dividend income from its underlying stocks.
[1] In its early history, Bitcoin also exhibited evidence of positive skew or a “fat right tail”. We do not consider the implications of positive skew in this report, but did discuss it in a prior report on crypto in portfolios.
[2] Throughout this report we use a composite index of two market cap-weighted crypto price indexes, and the choice of start date is based on data availability. The FTSE/Grayscale Crypto Sectors Index Series begins in April 2019. Prior to that, the only other available broad market cap-weighted index of crypto price returns is the S&P Cryptocurrency Broad Digital Market Index, which begins in February 2017. In this report, our measures of crypto asset class returns are based on the S&P Broad Digital Market Index through March 2019 and the FTSE/Grayscale Crypto Sectors Index Series thereafter.
[3] Source: Bloomberg, Grayscale Investments. Data as of December 31, 2024. Past performance not indicative of future results.
[4] Risk parity strategies, which typically apply leverage to low-volatility asset classes like Treasury bonds, can theoretically improve risk-adjusted returns.
[5] These are the basic concepts of Modern Portfolio Theory, a common approach to portfolio construction.
[6] A simplification, as many modern portfolios would likely include some alternative asset classes.
[7] Yield on cash assumed to be zero in calculation of Sharpe Ratio.
[8] As discussed in footnote #2, this is February 2017 to the present.
[9] Source: Bloomberg, Grayscale Investments. Based on the FTSE/Grayscale Crypto Sectors family of indexes and the S&P Cryptocurrency Broad Digital Market Index.
[10] Source: Coin Metrics, Grayscale Investments. Data as of December 31, 2024. Past performance not indicative of future results.
Share This
Like what you see?
Subscribe below for the latest research from Grayscale.
