Common Risk Factors in Cryptocurrency: Factor Investing Beyond Equities

Factor Investing Comes to Crypto

For decades, factor investing in equities followed a well-established playbook. Fama and French identified size and value. Carhart added momentum. Novy-Marx contributed profitability. By the 2010s, hundreds of factors had been documented, tested, and (in many cases) exploited by systematic managers. But as cryptocurrency markets matured from a niche experiment into a $2 trillion asset class, a fundamental question emerged: do digital assets have their own risk factors, distinct from the equity factors that traditional finance spent half a century cataloguing?

Liu, Tsyvinski, and Wu (2022) addressed this question directly in their Journal of Finance study "Common Risk Factors in Cryptocurrency." Their answer is striking; and carries significant implications for anyone allocating capital to digital assets.

The Core Finding

The paper's central result is that cryptocurrency returns are driven by three crypto-specific factors that are distinct from traditional equity factors. These three factors (a crypto market factor, a crypto size factor, and a crypto momentum factor) collectively explain over 80% of the cross-sectional variation in cryptocurrency returns. Meanwhile, the standard Fama-French equity factors (market, size, value, profitability, investment) have essentially zero explanatory power for crypto returns.

This is not a minor technical distinction. It means that an investor who understands equity factor exposures thoroughly (who can decompose a stock portfolio into its value, momentum, quality, and size tilts) cannot simply transfer that framework to cryptocurrency. Crypto operates under its own factor structure.

How the Factors Are Constructed

Liu, Tsyvinski, and Wu follow the methodology established by Fama and French for equities, adapted to the unique characteristics of cryptocurrency markets.

CMKT (Crypto Market Factor): The value-weighted return of all cryptocurrencies above a minimum market capitalization threshold. This is analogous to the equity market factor (MKT-RF), representing the broad return to holding crypto exposure. The crypto market factor exhibits substantially higher volatility than its equity counterpart; annualized volatility exceeding 80% compared to roughly 15-20% for equities.

CSMB (Crypto Size Factor): Constructed by sorting cryptocurrencies by market capitalization and computing the return spread between small-cap and large-cap crypto portfolios. Smaller cryptocurrencies tend to outperform larger ones, echoing the size effect documented in equities by Banz (1981) and Fama and French (1993). However, the magnitude of the crypto size premium is considerably larger (on the order of 3-5% per month during the sample period) and likely reflects the higher information asymmetry and liquidity constraints in small-cap crypto markets.

CMOM (Crypto Momentum Factor): Built by sorting cryptocurrencies on past one-week to four-week returns and computing the spread between winners and losers. Crypto momentum operates on much shorter horizons than equity momentum (which typically uses 12-month formation periods with a 1-month skip). This reflects the faster information diffusion and higher turnover in crypto markets.

Why Equity Factors Fail in Crypto

One of the paper's most important contributions is demonstrating why traditional equity factors cannot explain crypto returns. The authors test the standard Fama-French five-factor model (market, size, value, profitability, investment) and the Carhart four-factor model (adding momentum) against cryptocurrency returns. The results are unambiguous: none of these equity factors carry significant factor loadings when applied to crypto portfolios.

The explanation is intuitive when you consider the fundamental nature of each asset class. Equity factors like value (HML) are rooted in accounting fundamentals (book equity, earnings, cash flows. Cryptocurrencies have no book value, no earnings, and no cash flows in the traditional sense. The profitability factor (RMW) requires revenue and cost data that simply does not exist for most tokens. Even the equity size factor (SMB), while conceptually similar to the crypto size factor, operates through different economic mechanisms) small-cap equity outperformance relates to information asymmetry and illiquidity in fundamentally driven markets, while small-cap crypto outperformance likely reflects network adoption dynamics and speculative attention flows.

This independence has a portfolio construction implication: adding crypto to a diversified equity factor portfolio does not merely add more of the same risk exposures. It introduces genuinely different risk factors, which is the theoretical basis for diversification benefits.

The Network Effect: What Makes Crypto Size Different

The crypto size factor deserves particular attention because its economic mechanism differs fundamentally from the equity size factor. In equities, small firms outperform (when they do) partly because they are riskier, less liquid, and less followed by analysts. In crypto, the size effect operates through network adoption dynamics.

Small cryptocurrencies that gain user adoption and network activity tend to experience disproportionate price appreciation as they move from obscurity to recognition. This is a power law phenomenon: a protocol moving from 100 to 10,000 users represents a 100x increase in network value under Metcalfe's Law, while a large protocol moving from 10 million to 10.1 million users represents marginal growth. The crypto size factor effectively captures this adoption-driven return, which has no analog in equity markets.

However, this mechanism also carries a warning. The crypto size premium is concentrated in the subset of small coins that survive and gain traction. Many small cryptocurrencies fail entirely, going to zero. The size factor's average return masks a highly skewed distribution: a small number of enormous winners offset a large number of total losses.

Momentum in Crypto: Faster and More Fragile

Crypto momentum operates on compressed timescales compared to equity momentum. Where Jegadeesh and Titman (1993) documented the equity momentum effect using 3-12 month formation periods, crypto momentum is strongest at 1-4 week horizons. Several mechanisms explain this acceleration.

First, crypto markets trade 24/7 across global exchanges with no circuit breakers. Information (and sentiment) is incorporated into prices continuously, compressing the underreaction-to-overreaction cycle that drives momentum in equities.

Second, the retail-dominated nature of crypto markets amplifies attention cascades. When a token begins trending on social media, retail capital flows in rapidly, creating short-lived but intense momentum effects that resemble the behavioral patterns documented by Barber and Odean (2000) in equity markets, but on an accelerated timeline.

Third, and critically, crypto momentum is more fragile than equity momentum. The same compressed timescales that create momentum also produce sharper reversals. Daniel and Moskowitz (2016) documented "momentum crashes" in equities; sudden, severe reversals following extended bear markets. In crypto, these reversals occur more frequently and with less warning, making crypto momentum strategies harder to implement in practice.

Practical Implications for Portfolio Construction

For investors allocating to cryptocurrency, the three-factor framework offers several concrete insights.

Factor-aware allocation. Rather than simply buying Bitcoin or the top 10 coins by market cap, investors can think about their crypto exposure in factor terms. A Bitcoin-heavy portfolio is essentially a CMKT bet with minimal size or momentum exposure. Adding smaller altcoins introduces CSMB exposure, while active rebalancing based on recent performance introduces CMOM exposure. Understanding which factors you are exposed to helps set realistic return expectations and risk budgets.

Diversification benefit assessment. The finding that crypto factors are independent of equity factors provides a quantitative basis for including crypto in a multi-asset portfolio. However, this independence is conditional: during extreme risk-off events, correlations between crypto and equity markets have spiked, temporarily reducing the diversification benefit. The factor independence documented in the paper reflects average conditions, not crisis conditions.

Rebalancing frequency. The shorter timescales of crypto factors (particularly momentum) suggest that crypto portfolios may benefit from more frequent rebalancing than equity portfolios. Monthly rebalancing, standard in equity factor strategies, may be too slow for crypto. Weekly rebalancing captures more of the momentum premium but also incurs higher transaction costs.

Limitations and Caveats

Several important limitations should temper the practical application of these findings.

Survivorship bias. Cryptocurrency markets have experienced thousands of coin delistings, rug pulls, and project failures. Any factor study that uses only surviving coins will overstate factor premia. Liu, Tsyvinski, and Wu mitigate this by using a comprehensive database, but the issue is more severe in crypto than in equities, where exchange-listed firms face regulatory scrutiny before listing and during delistings.

Transaction costs. The crypto size and momentum factors require trading in smaller, less liquid coins. Bid-ask spreads, slippage, and exchange fees in small-cap crypto can be substantial; often 50 basis points or more per trade. The gross factor premia reported in the paper may not survive realistic transaction costs, particularly for the size factor and for frequent momentum rebalancing.

Regulatory risk. The crypto factor structure could change significantly if major regulatory actions alter market composition. Token delistings, exchange restrictions, or classification changes can affect entire segments of the market simultaneously.

Short sample period. Cryptocurrency markets have existed for roughly 15 years. The sample period for academic studies is inherently shorter than the decades of data available for equity factors. Whether the factor structure documented in 2022 will persist is an open question. McLean and Pontiff (2016) showed that published equity anomalies decay after publication; the same effect may apply to crypto factors as more capital targets them.

Market maturation. As crypto markets institutionalize (with the arrival of spot ETFs, regulated futures, and institutional custody) the factor structure may evolve. Institutional participation tends to reduce the very inefficiencies that generate factor premia. The size premium may narrow as liquidity improves in mid-cap tokens, and momentum may weaken as algorithmic traders compress the signal horizon.

Where the Evidence Stands

Liu, Tsyvinski, and Wu's framework represents the first rigorous application of asset pricing methodology to cryptocurrency markets, published in the field's top journal. Their finding of three distinct crypto factors (market, size, and momentum) that are independent of equity factors provides a structured way to think about crypto exposure within a multi-asset portfolio.

The key insight is not that crypto offers "alpha" in the traditional sense, but that it represents exposure to a genuinely different set of systematic risk factors. This distinction matters: alpha decays as it is discovered and exploited, while factor premia (if they reflect genuine economic mechanisms like network adoption and information asymmetry) can persist.

For practitioners, the message is nuanced. The crypto factor structure exists and is statistically robust, but implementing it faces challenges that equity factor investing solved decades ago: reliable data, reasonable transaction costs, survivable drawdowns, and a long enough track record to separate signal from noise. Crypto factor investing is where equity factor investing was in the early 1990s; academically validated but practically immature.