What Predicts Long-Term Outperformance? An Evidence-Based Summary

The investment industry produces an endless stream of metrics, ratios, and scoring systems, each claiming to predict future returns. But when you examine the academic evidence rigorously, only a handful of factors have genuinely demonstrated persistent predictive power over 5-10 year horizons. Many popular metrics are noise dressed up as signal. This post summarizes what decades of financial research actually tells us about what works — and what does not — in predicting long-term stock returns.

The evidence is surprisingly clear: a small set of well-understood factors explains a large portion of cross-sectional return differences, and they have held up across geographies, time periods, and market regimes.

Quality and Profitability: The Strongest Long-Term Signal

If you could only use one factor to predict long-term stock performance, the evidence points to profitability and quality. Companies with high returns on invested capital (ROIC), high gross margins, low accruals, and stable earnings significantly outperform over 5-10 year periods. The research by Novy-Marx on gross profitability, and later by Asness, Frazzini, and Pedersen on quality-minus-junk, shows that high-quality companies — those with high profitability, stable growth, and conservative balance sheets — outperform low-quality companies by 3-5% annually. Critically, this factor is persistent even after controlling for valuation — you do not need to buy cheap quality stocks to benefit. Quality works because the market systematically underestimates the durability of high returns on capital, and overestimates mean reversion.

Value: Still Works, but Slowly

The value premium — buying cheap stocks and avoiding expensive ones — is the most studied factor in finance. Measured by metrics like price-to-book, price-to-earnings, EV/EBITDA, and free cash flow yield, value has delivered 2-4% annual outperformance over very long periods. However, the value factor is notoriously cyclical. It can underperform for years at a stretch (as it did from 2017-2020) before delivering concentrated bursts of outperformance. For long-term investors, the evidence still supports value — but only with the patience to endure multi-year drawdowns and the discipline to avoid abandoning the approach at the worst possible time. Combining value with quality significantly improves results, filtering out the value traps that drag down pure cheap-stock strategies.

What Doesn't Predict Well

Several popular metrics have surprisingly weak predictive power over long horizons. Analyst ratings and price targets show almost no correlation with future returns. Dividend yield, while popular with income investors, is a poor predictor of total returns. Revenue growth in isolation does not predict returns — it must be combined with profitability metrics to be useful. Even momentum, which is one of the most robust short-term factors, decays significantly beyond 12 months and has limited value for buy-and-hold investors. Short interest, insider selling, and most technical indicators also show weak or inconsistent evidence over 5+ year horizons.

Building a Multi-Factor Approach

The strongest evidence supports combining factors rather than relying on any single metric. A portfolio that selects stocks ranking highly on quality (ROIC, margin stability), value (earnings yield, FCF yield), and moderate momentum has historically delivered the most consistent long-term outperformance with the lowest drawdowns. Factor diversification works because quality, value, and momentum tend to have low correlations with each other — when one factor is out of favor, the others often compensate. The key insight is that no single metric is a magic bullet, but a disciplined multi-factor approach harnesses the most reliable patterns in equity markets.

Start building a quality-focused portfolio by screening for companies with high returns on capital using our High ROIC screen — targeting the profitability factor that the evidence shows is the single strongest predictor of long-term outperformance.