It’s not every day, or even every decade, that an entirely new asset class is born. Yet, through a combination of computer science, cryptography, economics, and network theory, digital assets have arrived and are proving that they are an asset class unlike any other. As they transform our global financial infrastructure and challenge modern monetary theory,1 we believe digital assets are one of the most exciting investment opportunities of the 21st century.
In this paper, we will demonstrate why we view digital assets as a brand new asset class that can enhance strategic asset allocation and help investors build portfolios with higher risk-adjusted returns. We will provide a few different lenses through which the reader can gain a deeper understanding of the role that digital assets may play in building more efficient portfolios.
As first outlined by Chris Burniske and Adam White in their January 2017 whitepaper titled Bitcoin: Ringing the Bell for a New Asset Class, digital assets provide exposure to unique market opportunities and risks, thus creating a diversifying return stream for investors. As such, they should be considered a component of the optimal beta portfolio2 alongside traditional assets such as equities, bonds, and real estate.
Since the focus of this paper is on portfolio construction, we will not go deep into detail on the investment merits of individual digital assets. However, we encourage you to review the pioneering work of Burniske and White, as well as our previous investment theses for select digital assets here:
Bitcoin & the Rise of Digital Gold (June 2016)
Hedging Global Liquidity Risk with Bitcoin (December 2016)
Into the Ether with Ethereum Classic (August 2017)
The Zcash Investment Thesis (January 2018)
It should be noted that numerous investment opportunities exist outside of the digital assets that we have explored, but our analyses are intended to paint a picture of our general framework for assessing real-world use cases that drive the investability of the asset class.
A Brief Intro to Modern Portfolio Theory
“Diversification is the one free lunch of investing, and when you see a free lunch, the only rational thing to do is eat.” 3
Cliff Asness, Managing Principal and CIO at AQR Capital Management
New asset classes are rare and very powerful because they offer a unique return stream that can provide a diversification benefit. This might seem like a simple concept, but few investors truly appreciate the impact this can have on the return/risk profile of a portfolio, and subsequent wealth creation.
According to Modern Portfolio Theory (“MPT”), by estimating the future returns, volatilities, and correlations of various assets, and pairing different combinations of each, an efficient frontier of portfolios can be constructed, in which the level of return is optimized per unit of risk.4 While it has proven difficult to estimate these parameters for assets over short periods of time, in the long-run, equilibrium risk, return, and correlation values can be estimated more reliably, making this framework incredibly valuable for disciplined investors who stick to a strategic asset allocation.5
We generally subscribe to the notion that the optimal return/risk ratio for a portfolio can be found on the efficient frontier. But contrary to conventional wisdom, we think many of today’s asset allocators are missing out on a “free lunch.” That’s because (i) digital assets represent a brand new investment opportunity that is uncorrelated to other asset classes and (ii) investors are generally under-allocated to this sector. It is our view that the optimal beta portfolio lies somewhere higher than what was previously believed to be the efficient frontier, and digital assets are the proverbial “missing piece of the puzzle.”
By viewing digital assets in this context, we’re simply taking well-established portfolio management principles and applying them to a new class of assets. Throughout the remainder of this paper, we will provide evidence of how diversification can be extended to digital assets and why we believe they are a valuable tool to investors seeking to build sustainable portfolios with higher risk-adjusted returns.
A Brand New Asset Class
Over the past several years, Bitcoin (BTC) has embodied the power of decentralized systems, exhibiting no single point of failure in both a permissionless and censorship-resistant manner. This important proof-of-concept has now led to an explosion in the development of second-generation blockchain protocols, designed to expand upon the capabilities of Bitcoin by modifying its social, economic and/or technological constructs to satisfy entirely different needs. These differences have necessitated a new way to describe the assets that operate these protocols. There are no longer just digital currencies, but also digital commodities and digital tokens, which fall into the broader category of digital assets.
Digital currencies, like Bitcoin, seek to fulfill the role of a decentralized global currency and store-of-value, a necessary alternative to government-monopolized fiat monetary regimes in the post-financial crisis world. Others, like Zcash (ZEC) and Monero (XMR), build upon Bitcoin’s role by offering privacy-enhancing features that further protect the identities of users and their transaction details. Digital commodities, like Ethereum (ETH), fuel decentralized applications (DApps) that can execute complex, condition-based transactions through the use of smart contracts, while assets like Ethereum Classic (ETC) are a hybrid currency and commodity, combining the monetary characteristics that have made Bitcoin successful as a digital store-of-value with the smart contract capabilities of Ethereum. There are also new networks like Filecoin (FIL), billed as “Airbnb for file storage,” that seek to demonstrate the value that distributed networks can provide as a more efficient data storage architecture, with built-in incentives for users who contribute unused file storage capacity. These are just a few examples of how digital assets are functioning today.
Moreover, digital assets are squarely at the intersection of some of the most significant trends reshaping the global economy6, including:
- A new market paradigm, characterized by slow economic growth, low interest rates, and divergent central bank policies.
- Rapid advancements in financial technologies and payment infrastructure, which now make it possible to move, settle, and clear value/assets at the same speed as information in a digital format.
- Regulatory shifts, altering financial industry economics and significantly increasing the cost of compliance and financial operations.
- Demographic shifts, driven by (i) the next generation of investors entering their prime earning years (i.e., millennials) and (ii) baby boomers entering retirement and tapping underfunded pension plans.
The combination of these micro and macro factors has led many to believe that digital assets are among the greatest technological and financial innovations since the advent of the internet itself.
While digital assets have already begun addressing various use cases, this technology will likely evolve in many ways that we cannot possibly imagine. Still, it is becoming clearer that it will be very difficult for a single asset to be optimized for all use cases simultaneously. For that reason, we think it is likely that multiple digital assets will survive, thrive, and complement one another within this new financial and technological architecture, solidifying a bona fide asset class.
Combining the growth opportunities that digital assets offer as a revolutionary technology and the store-of-value characteristics that many of them possess as alternative currencies, digital assets may have the potential to provide both inflation protection and growth exposure concurrently.
The Power of Diversification
Now that we’ve established why we believe digital assets have and will continue to be a unique investment opportunity, we’ll quantitatively reinforce the power of uncorrelated assets and why they are a foundational element of building more balanced portfolios.
Below we show the formula for calculating portfolio risk:
We can use this formula to gain a deeper understanding of the profound impact that uncorrelated assets can have on a portfolio by holding all other factors constant (e.g., volatility, portfolio weight, expected return, etc.) and running a sensitivity analysis to assess how assets with different correlation coefficients affect portfolio risk.
The table below shows the cumulative portfolio risk reduction (i.e., diversification benefit) that an investor can realize when constructing balanced portfolios that include anywhere from two to fifteen assets with different correlation coefficients. For the purpose of this analytical exercise, we define a balanced portfolio as one in which each asset has equal weight and volatility.
For example, if an investor were to construct a balanced portfolio consisting of four assets, where each asset has a correlation of zero with one another (ρ = 0), the risk of that portfolio would be 50% lower than if those same four assets were perfectly correlated (ρ = 1). We’re able to derive this from the formula shown in Figure 2.
We can also create a real world example that reinforces this point in a more intuitive way, by assuming that the excess return and volatility of each asset within the portfolio is 4% and 10%, respectively.
If the assets in the portfolio are perfectly correlated (ρ = 1), the Sharpe ratio of the portfolio would be 0.4 (4% excess return / 10% portfolio risk). In this scenario, there is no diversification benefit, as the risk of the portfolio is no lower than that of the individual assets that comprise it.
However, if those same four assets were uncorrelated (ρ = 0), an investor would be able to structure a portfolio with a Sharpe ratio of 0.8 (4% excess return / 5% portfolio risk). Due to the uncorrelated nature of the assets in this scenario, the risk of the portfolio can be reduced from 10% to 5% (a 50% reduction), resulting in a Sharpe ratio that is twice as high.
This analysis highlights how important uncorrelated assets are in portfolio construction and how investors can build portfolios with higher risk-adjusted returns, not necessarily by finding better performing assets, but by properly combining uncorrelated returns streams.
Based on their unique use cases and applications, it is clear there is a qualitative basis for digital assets to be lowly-correlated with traditional assets. In the next step of our analysis, we will assess whether this concept is quantitatively reinforced in reality, by observed market returns.
Below we examine the relationship that some of the larger, more established digital assets have to traditional assets and each other through a matrix constructed from the average rolling one-month correlations over the last 1.5 years. We selected this 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 the sake of consistency and for comparison purposes, we will use this timeframe throughout the remainder of this paper.
From the previous charts, we can see that the average rolling one-month correlations range from slightly negative to slightly positive, with an average correlation of zero. This provides evidence that digital assets can be considered a diversifying component in multi-asset portfolios. Moreover, many digital assets are imperfectly correlated to one another, which means there may even be diversification benefits within the asset class itself.
In the next section, we will provide examples of how digital assets can be used by portfolio managers as a tool to build portfolios with better risk-adjusted returns.
Building Better Portfolios with Digital Assets
As digital assets promote economic growth in innovative ways, they offer investors an opportunity to build more efficient portfolios.
Slow global growth, secular-high debt burdens, deteriorating effectiveness of monetary policies, and low yielding assets are all contributing to a savings crisis that threatens the economic welfare of future generations. We have entered a low return environment for traditional assets with significant downside risk, rendering it difficult for many investors to achieve their target returns.
In order to resolve these challenges, there are two options available to investors. They can:
- Increase exposure to risky assets already held in their portfolios in an attempt to generate higher returns, which will mean holding more concentrated, less diversified portfolios, with higher risk of ruin; or
- Identify uncorrelated assets with positive expected returns, and use them to build more efficient portfolios.
As an investment that is uncorrelated to other components of investors’ portfolios, digital assets can further enhance a strategic asset allocation.
To gain a deeper understanding of the diversification benefits that digital assets can offer, we conducted a series of portfolio simulations to see how an allocation to Bitcoin and an equal-weighted mix of select digital assets might have impacted the return and risk profile of a portfolio comprised of global equities and bonds (the “Global 60/40”).10
From the chart above, we can see that even small allocations to Bitcoin can significantly enhance the returns of traditional portfolios, like the Global 60/40, without materially increasing volatility. For example:
- Adding a 1% Bitcoin allocation to the Global 60/40 increased the hypothetical simulated cumulative return by 293 bps, without materially increasing volatility to improve risk-adjusted returns by 18%.
- Adding a 3% Bitcoin allocation to the Global 60/40 increased the hypothetical simulated cumulative return by 896 bps, without materially increasing volatility to improve risk-adjusted returns by 45%.
- Adding a 5% Bitcoin allocation to the Global 60/40 increased the hypothetical simulated cumulative return by 1,524 bps, without materially increasing volatility to improve risk-adjusted returns by 101%.
Furthermore, we can run attribution on the hypothetical simulated portfolio returns to identify what proportion of the cumulative period return was driven by the allocation to Bitcoin versus the Global 60/40. This can help a portfolio manager assess whether the return received on an investment was commensurate with the level of risk taken.
As the previous charts show:
- Adding a 1% Bitcoin allocation to the Global 60/40 drove approximately 16% of the hypothetical simulated portfolio’s return for the period under analysis.
- Adding a 3% Bitcoin allocation to the Global 60/40 drove approximately 36% of the hypothetical simulated portfolio’s return for the period under analysis.
- Adding a 5% Bitcoin allocation to the Global 60/40 drove approximately 49% of the hypothetical simulated portfolio’s return for the period under analysis.
From the above, it becomes clear that placing even small amounts of capital at-risk through an allocation to Bitcoin may meaningfully drive portfolio return in certain investment environments.
In the second series of portfolio simulations, we took this one step further. Instead of allocating exclusively to Bitcoin within the digital asset bucket, we invested across an equal-weighted mix of Bitcoin (BTC), Bitcoin Cash (BCH), Ethereum (ETH), Ethereum Classic (ETC), Litecoin (LTC), Ripple (XRP), and Zcash (ZEC). This simulation can provide insight as to whether allocating to a diversified mix of digital assets can improve the risk-adjusted returns of portfolios versus those solely containing Bitcoin.
There are a number of different weighting mechanisms one can use when evaluating the impact of a digital asset mix. For the purpose of this analysis, we opted for simplicity, and a weighting scheme that does not bias (overweight or underweight) any digital asset relative to another.
Looking at the chart above, it appears that portfolios containing allocations to a mix of digital assets performed even better than those only including Bitcoin. For example:
- Adding a 1% digital asset allocation to the Global 60/40 increased the hypothetical simulated cumulative return by 484 bps, without materially increasing volatility to improve risk-adjusted returns by 30%.
- Adding a 3% digital asset allocation to the Global 60/40 increased the hypothetical simulated cumulative return by 1,508 bps, without materially increasing volatility to improve risk-adjusted returns by 76%.
- Adding a 5% digital asset allocation to the Global 60/40 increased the hypothetical simulated cumulative return by 2,610 bps, without materially increasing volatility to improve risk-adjusted returns by 166%.
Given what we know about MPT, this is not all that surprising. Since digital assets are imperfectly correlated with one another, they can be combined to build portfolios with higher risk-adjusted returns.
We conclude our historical analysis by looking at the return contribution from digital assets relative to the capital-at-risk during the period:
As the previous charts show:
- Adding a 1% digital asset allocation to the Global 60/40 drove approximately 24% of the hypothetical simulated portfolio’s return for the period under analysis.
- Adding a 3% digital asset allocation to the Global 60/40 drove roughly 49% of the hypothetical simulated portfolio’s return for the period under analysis.
- Adding a 5% digital asset allocation to the Global 60/40 drove approximately 62% of the hypothetical simulated portfolio’s return for the period under analysis.
It’s worth highlighting that we also ran hypothetical simulations over longer timeframes, dating back to September 2013. In doing so, we used the data available for the digital assets included in this analysis that existed at the time. The results were broadly consistent in that the hypothetical simulated portfolios containing both Bitcoin and an equal-weighted mix of digital assets had higher risk-adjusted returns than the standalone Global 60/40.
Investing for the Future
Examining historical performance is one way to test our hypothesis about the role that digital assets can play in portfolio construction, but it doesn’t tell us much about what we should expect in the future if the risk and return dynamics of the asset class change. We believe that most people would agree that the historical return/risk profile for digital assets is unsustainable and that future equilibrium will likely be a different story. Unfortunately, due to their limited trading history, we don’t yet know what the equilibrium return/risk profile looks like.
While the long-run volatilities of asset classes can differ substantially, there is consistency when looking across their Sharpe ratios. As you can see from the below chart, the equilibrium Sharpe ratio across asset classes is generally between 0.25 – 0.35.16
One way to interpret this consistency is that the relative risk and return preferences of investors are reflected almost uniformly across asset classes over time. In other words, for accepting a certain level of risk (i.e., volatility) investors will generally require the proportionate level of return.
Using this insight, we can make relatively conservative assumptions about the long-term expected risk and return profile of digital assets to estimate the impact that a small allocation (5%) could have on wealth creation when coupled with a Global 60/40 portfolio.
We think it is reasonable to assume that digital assets will continue to be the most volatile asset class for quite some time. However, we assume that the equilibrium volatility will dampen to around 35%. In this scenario, we’d expect digital assets to have an annualized excess return of approximately 10.5%, based on a Sharpe ratio of 0.3.
Although the average correlation of digital assets to other asset classes appears to be around zero today, we think that correlations could increase as more managers allocate to digital assets over time. Our basis for this is that managers must often make decisions about how they allocate between cash and risk assets, particularly in a liquidity crisis. Since digital assets clearly fit into the “risk” bucket, this relationship could drive their correlations to other risk assets higher. Still, we continue to believe that a high proportion of the returns for the digital asset class will be driven by idiosyncratic factors. For this reason, we assume an equilibrium correlation of 0.2.
Now that we’ve established our equilibrium assumptions, we can create an investment scenario that almost all investors face: saving for retirement.
Assuming $100,000 of starting capital and annual contributions of $18,500 per year, we outline what the Global 60/40 (95%) + Digital Assets (5%) hypothetical simulated portfolio would look like from a return/risk perspective, as well as the impact this would have on retirement savings over different time horizons.
Based on the assumed risk, return, and correlation dynamics of the Global 60/40 and digital assets, we determine that the Global 60/40 (95%) + Digital Assets (5%) hypothetical simulated portfolio would generate roughly 30 bps of additional return on an annualized basis when compared to the Global 60/40, at the same level of risk. While this might seem small, the effect of compound returns can make this meaningful over time.
In the following charts, we map out a 40-year timeframe and calculate the surplus that the Global 60/40 (95%) + Digital Assets (5%) hypothetical simulated portfolio would have delivered over the Global 60/40 in dollar-terms. We can also see how the surplus might vary at different points along the curve, relative to an investor’s time horizon.
Furthermore, we can calculate the worst-case scenario for the Global 60/40 (95%) + Digital Assets (5%) hypothetical simulated portfolio relative to the Global 60/40 by assuming that the value of digital assets drops to $0 in Year 1, when the impact on compounding wealth would be greatest.
Below we’ve charted the the maximum shortfall, estimated surplus, and median outcome for the Global 60/40 (95%) + Digital Assets (5%) hypothetical simulated portfolio relative to the Global 60/40 based on our assumptions.
The results were as follows:
- In the maximum shortfall scenario, the Global 60/40 (95%) + Digital Assets (5%) hypothetical simulated portfolio would generate a total value of $3,068,626 versus $3,106,610 for the Global 60/40 at the end of the 40-year period. This indicates a maximum shortfall of $37,984, or roughly 1.2% less than the dollar value of the Global 60/40 at the time of retirement.
- In the surplus scenario, if digital assets perform in line with the return/risk profile we have outlined, the ending value of the Global 60/40 (95%) + Digital Assets (5%) hypothetical simulated portfolio would be $3,345,498 versus $3,106,610 for the Global 60/40, a surplus of $238,889. This is nearly 8% higher than the dollar value of the Global 60/40 at the time of retirement, and more than 6X the absolute value of the maximum shortfall.
- Finally, the median outcome is net positive, and asymmetric relative to the maximum shortfall. Another way to interpret this is that there is higher return potential per unit of capital-at-risk. Given that the median outcome is 2.6X the absolute value of the maximum shortfall, digital assets could be a mispriced risk exposure. For that reason, it may make sense for investors to allocate to these assets if properly sized within their portfolios.
Although we believe this analysis is relatively conservative, we encourage investors to use this framework and test their own set of assumptions.
At Grayscale, we were early investors in digital assets because we have long believed in their potential to capture a share of some of the largest markets in the world (e.g., store-of-value), improve the efficiency of our global financial system, and create business models that democratize information and value in incredible new ways. We also recognized that because of their highly unique set of properties, they offer a distinct return stream, allowing them to play a diversifying role in investor portfolios.
Our motivation for creating this paper was twofold. First, we wanted to stress-test our hypothesis that digital assets fit squarely within Modern Portfolio Theory, a time-tested and proven approach that many investors are using today to build better portfolios. Second, we wanted to share our analytical framework with those who might benefit from understanding it to determine the optimal digital asset allocation within their own portfolios.
It’s still early in the lifecycle of digital assets, but we believe our multifaceted approach to assess their investability makes a compelling case for investors to have some portion of their portfolio allocated to this new asset class. A lot can happen over the next few years, but remember: diversification is a “free lunch” and asset allocation is all about the long-game. We invite you to join us on the journey to a new frontier.