ART ProduçõesThe CAPM (Capital Asset Pricing Model), formulated by William Sharpe in 1964, predicts that a stock's expected return grows linearly with its beta — its sensitivity to the market. Riskier stocks (high beta) should deliver higher expected returns; less risky ones (low beta), lower returns. It's elegant, intuitive, and was the foundation for a generation of portfolio theory.
The problem: the data, for fifty years now, shows the opposite. Low-volatility portfolios deliver risk-adjusted returns — Sharpe ratio — systematically superior to high-volatility ones. In some periods and markets, even absolute returns are higher. This is the low-volatility anomaly, or low-volatility anomaly, perhaps the most embarrassing empirical finding for classical finance.
The empirical evidence
The first to notice the problem was Black, Jensen and Scholes (1972) — yes, the same Fischer Black from Black-Scholes — who documented that the observed relationship between beta and return was flatter than predicted by CAPM. Robert Haugen, in several papers throughout the 90s, showed that minimum-variance portfolios in the S&P 500 historically delivered returns comparable to the index with 25-30% lower volatility.
Frazzini and Pedersen (2014) formalized the anomaly in a framework called Betting Against Beta: buy leveraged low-beta stocks, sell high-beta ones. The result in global data was extraordinarily robust alpha in virtually all tested markets, including emerging ones.
At B3, Lobão, Pacheco and Pereira (2017) and Mussa et al. (2012) replicated the phenomenon. Low-vol portfolios constructed with IBrX-100 delivered, over 10+ year windows, Sharpe ratios 40-60% higher than IBOV.
Low-vol score of a stock at time t: volatility of daily returns over the last 252 trading sessions (≈12 months), inverted (lower volatility → higher score). One can also use variance or semi-annual standard deviation.
Why it exists — three explanations
1. Leverage aversion (Frazzini-Pedersen)
Investors who want returns above the market have two options: (a) leverage a low-beta portfolio; (b) buy high-beta stocks without leverage. Most cannot or will not leverage — pension funds have restrictions, individuals don't have easy access to cheap leverage. Result: excessive demand for high-beta stocks, which become overpriced; insufficient demand for low-beta ones, which become cheap. Returns reflect this distortion.
2. Lottery preference
Behavioral studies (Kumar, 2009; Bali et al., 2011) show that individual investors have a systematic preference for stocks that behave like lottery tickets — high volatility, possibility of enormous gains, low expected return. This pushes prices of volatile stocks up and creates the low-volatility premium on the other end.
3. Institutional mandate constraints
Fund managers are evaluated against benchmarks. Buying a portfolio very different from the benchmark — even if statistically better — generates high tracking error, and high tracking error is punished by investment committees. The result is that nobody buys the optimal low-beta portfolio, even though it empirically exists.
Why low-vol is not arbitraged away
The natural question: if this anomaly has been known since the 70s, why does it still exist? The answer has three components:
- Institutional constraints persist. Pensions, insurers and endowments have statutory rules that limit how far they can deviate from benchmarks. This doesn't change.
- Behavioral bias persists too. Individuals continue to prefer "stocks that can double" to "boring stocks that rise 12% a year with low volatility".
- The factor has carrying costs. In sharp bull markets (like 2020-2021), low-vol underperforms the index for years. Those who enter expecting immediate outperformance exit before the premium materializes.
Practical factor construction
There are two schools. Total volatility uses the standard deviation of daily returns, plain and simple. Idiosyncratic volatility first removes the systematic component (regressing against IBOV) and uses the residual — it's theoretically more "pure" but empirically captures the same phenomenon in most windows.
VORTEX QSP uses total volatility over the last 252 trading sessions, normalized cross-sectionally. The choice for simplicity is deliberate: complexity doesn't compensate for robustness.
Low-vol concentrates exposure in defensive sectors — utilities, staples, telecom. Without sector restriction, a purely low-vol portfolio becomes an "Energy + Telecom portfolio". VORTEX QSP applies an anti-concentration sector limit to avoid this, maintaining the low-volatility premium without becoming hostage to a sector macro-thesis.
Combination with other factors
Low-vol alone has two problems: (a) it underperforms in vertical bull markets; (b) it concentrates in sectors. The classic solution is to combine it with momentum (which captures trends) and quality (high ROE, low leverage) — the intersection of the three typically beats each individual factor in Sharpe and absolute returns in most tested windows.
This is exactly the logic of the composite score described in Technology: equal-weight across five orthogonal factors (momentum, low volatility, quality, value, low beta), with risk constraints applied at the portfolio level.
In closing
Low volatility is not a "conservative defensive" factor — it's a factor that delivers a premium independent of regime, risk-adjusted, and that historically beats the market in metrics relevant to the long-term investor. If your goal is to maximize Sharpe ratio (and not beat IBOV in every bull market), low-vol should be a central component of your portfolio.
And yes, that's exactly what VORTEX QSP does. Not in isolation — together with momentum, quality and the others — but with significant weight.