Did Sister Wood Make a Mistake? A Dispute Over the Definitions of Beta and Alpha

Recently, Bitcoin has been showing a weaker trend, and market sentiment has been swinging accordingly, which also brings to mind a core view held by Ark Invest founder and CEO Cathie Wood: Bitcoin can outperform in a bull market and serve as a hedge in a bear market.

It sounds like “both offense and defense” is indeed like a fairy tale, but if you look at the longer term, you’ll find that Bitcoin has repeatedly demonstrated a different rhythm from traditional assets across several cycles: sometimes highly correlated, sometimes decoupled, sometimes like a risk asset, and sometimes like a new independent narrative. Ark Invest and Cathie Wood often cite a set of statistics: since 2020, the correlation (or beta) of Bitcoin with gold, bonds, and the S&P 500 has been approximately 0.14, 0.06, and 0.28 respectively, meaning it can provide non-correlated risk in a portfolio.

My view is simpler: Bitcoin is not what Ark calls “digital gold”; it is more like an equity asset — during periods of liquidity contraction and declining risk appetite, it often shows higher synchronization (high beta). So why has it historically often taken its own path? The answer may not lie in beta but in alpha: Bitcoin possesses unique commercialization pathways and network effects that other risk assets do not have — from institutional entry, custody and compliance infrastructure improvements, to on-chain financial and payment scenarios evolution, and the supply shock expectations driven by the halving mechanism every four years. These factors are closer to sources of exclusive incremental returns rather than linear exposure to a market factor in the traditional sense.

This leads to an often overlooked fact: whether alpha or beta, linear regression is fundamentally a tool in a mathematician’s toolbox — inherently intangible, not an objective law verifiable in the market. The same data, when selected over different intervals, frequencies (weekly/daily), or with different explanatory variables (S&P, NASDAQ, real interest rates, dollar index), can produce completely different beta values; even with the same data, different linear regression equations can be derived (see the diagram below). More importantly, low correlation does not automatically mean better. Investment products can also achieve low beta by significantly lagging the market, which is statistically valid but clearly not what investors want.

Therefore, I do not think Wood was wrong; she simply overlooked an investment constraint: a low beta is just a description of risk structure, not a promise of returns. What truly determines the long-term experience is what you consider Bitcoin to be — a defensive asset, a risk asset, or an optional asset with strong narrative volatility. When the market enters a macro shock phase (policy expectations, changes in Federal Reserve personnel and interest rate paths, rising geopolitical uncertainties), Bitcoin tends to behave more like a risk asset; but during phases driven by technology and adoption, it may also release significant idiosyncratic excess alpha.

But I agree with one point Wood makes: Bitcoin’s risk-return distribution is indeed unique — whether you categorize it as beta or alpha, incorporating it into a portfolio proportionally to its risk will inevitably improve the efficient frontier. She often mentions that even a “5%” allocation can, without significantly increasing the risk related to traditional assets, enhance the portfolio’s upside resilience.

Diversification is the cornerstone of modern portfolio theory and will be the theme of our next article: when correlations rise under stress, how can investors diversify in a way that is more aligned with reality?

Stay tuned.