Strategies / Quality

Quality

5 min read · Strategy guide

Buy companies that are profitable, cash-generative, and lightly levered. Boring, slow, and remarkably effective in drawdowns.

What it is

A quality strategy ranks the universe by a composite of fundamentals — typically high return on equity (ROE), high free cash flow yield, low debt-to-equity, and stable earnings growth — and overweights the top quintile. The thesis is that markets systematically underprice the durability of high-quality businesses; you earn the spread between what fundamentals say they're worth and what short-horizon flow says they're worth.

The canonical academic reference is Asness, Frazzini & Pedersen (2013) — “Quality Minus Junk”, which formalised the long-quality / short-junk portfolio and showed it earned a positive risk premium across multiple markets and decades.

The core formula

# Composite z-score across four pillars
profitability = z(ROE) + z(gross_margin)
cashflow      = z(FCF / enterprise_value)
safety        = -z(debt_to_equity) - z(earnings_volatility)
growth        = z(5y_earnings_growth)

quality_score = (profitability + cashflow + safety + growth) / 4

# Long top quintile, rebalance quarterly

The pillars matter more than the exact composition. Different practitioner implementations (MSCI Quality, GMO, AQR) weight them differently, but a strategy that ranks on at least one profitability + one safety metric will look broadly similar.

When it works

Late-cycle and drawdown regimes. When credit tightens or growth slows, the market stops paying for narrative and starts paying for cash flow. Quality outperformed sharply in 2008, 2018-Q4, March 2020, and 2022 — every regime where the cost of capital reset and weak balance sheets got punished first.

Quality also tends to win on a risk-adjusted basis even when it loses on raw return — the strategy gives up some upside in rallies but loses far less in drawdowns. The Sharpe is competitive with momentum but the Calmar is usually better.

When it fails

Risk-on rallies, especially the speculative tail. 2020 (post-March) and 2021 were murder for quality: ARKK-style profitless growth, meme stocks, and beaten-down junk all ran 5–10x while solid franchise businesses compounded at 15%. Quality underperformed the SPY by something like 800 bps in 2020 and 600 bps in 2021.

The mechanism is intuitive: when liquidity is loose and risk appetite is high, the bid migrates to whatever has the most operational leverage to a recovery — which is exactly what a quality screen filters out.

What to watch in the result card

MaxDD vs SPY. Quality should drawdown less than the broad market in stress. If your backtest's MaxDD is similar to or larger than the benchmark, the composite is doing the wrong thing — likely overweighting one factor (e.g. growth) that's anti-correlated with the others.
IC decay slow. Quality factors are persistent. If your IC decay ratio drops below 0.7 by horizon t+1, the “quality” you're measuring is actually a momentum proxy in disguise. Real quality scores have ICs that stay positive 6+ months out.
Turnover low. Fundamentals don't change quarter-to-quarter for most companies. Healthy quality strategies run 50–100%/yr turnover. Anything above 200% is mis-classified or you're chasing earnings revisions, not quality.
Beta < 1. Quality is naturally defensive. Expect 0.7–0.9 beta to the broad market. A quality backtest with beta > 1.0 is probably tilting toward growth-quality, not safety-quality.

Common mistakes