Evidence Based Factor Investing | Matt Zenz

Episode Summary

Executive Summary: Matt argues that evidence-based investing should be systematic, transparent, and grounded in long-horizon data, not short-term product innovation or factor-chasing. He defends value, profitability, size, and momentum as implementable signals, emphasizes that factor premia come from risk and migration over time, and explains Longview’s “passive-aggressive” approach: start with market-cap weights, then modestly tilt toward higher expected returns when valuation spreads justify it.

Main Topics: What evidence-based investing really means (Priority: 5/5): Matt defines evidence-based investing as rules-based, transparent, and built on a strong expectation that the process will work across full market cycles. He pushes back on managers who use lots of data but still behave like active stock pickers. Factor definitions and the factor zoo (Priority: 5/5): He explains that a factor must have a sensible economic story, empirical support over long horizons, and be implementable in the real world. He warns that many of the roughly 200 documented factors are redundant or just repackaged versions of the same idea. Value, growth, and long-run mean reversion (Priority: 5/5): Matt repeatedly argues that value underperformance does not invalidate the premium; it reflects expected variation around a long-run relationship. He uses coin-flip and migration analogies to show why decades of evidence matter more than recent cycles. Portfolio construction and factor interaction (Priority: 4/5): Longview uses factors in combination, especially value plus profitability/quality, and also incorporates momentum and size. He says combining factors improves expected-return targeting and avoids unnecessary turnover or noise. Passive-aggressive investing and valuation spreads (Priority: 4/5): The firm’s 'passive-aggressive' framework begins at market-cap weights and then modestly tilts toward value when valuation spreads are unusually wide. He frames this as a controlled way to increase expected return without pretending factor timing can be precise. Implementation, fund size, and trading efficiency (Priority: 4/5): Matt argues that implementation alpha matters more than new research. Smaller, nimble funds can trade faster, reduce opportunity costs, and better harvest short-lived signals like momentum or issuance-related effects, especially in less liquid small-cap names. AI, uncorrelated assets, and investor discipline (Priority: 3/5): He sees AI as a tool for speeding up research and portfolio tooling, not for discovering magical new premia. He also rejects the idea that 'uncorrelated' alone is a good reason to own an asset, and closes by urging investors to stay low-cost and diversified.

Key Arguments: Evidence-based investing should be systematic, transparent, and based on long-term statistical evidence rather than compelling stories or recent performance. A valid factor needs three things: an economic rationale, long-horizon empirical support, and practical implementability. Most factor 'innovation' is redundant; many of the 200 documented factors are just different views of the same underlying information. Value’s recent weakness does not imply it is broken; with noisy markets, decades of data are needed before revising conclusions. Large-cap growth stocks can have lower expected returns because they became large after already delivering strong past performance, and their risk profile is typically lower. Value should be measured simply and robustly; Longview uses price-to-book because it isolates value better and avoids mixing in quality effects and extra turnover. Combining value with profitability/quality creates a cleaner expected-return signal than using either factor alone. Momentum is useful, but mainly as a trading/implementation filter because it acts over months while value/quality act over years. Valuation spreads can justify modest factor tilts ('sin a little'), but factor timing should be small, slow, and grounded in expected return, not forecasts. Small, nimble managers can better capture short-lived anomalies and avoid the trading lags and conflicts that large fund families face. AI can accelerate tooling and analysis, but overfitting makes it dangerous as a source of purported new alpha. The most durable edge in investing is implementation efficiency, especially tax efficiency and minimizing cash drag, rather than discovering a brand-new factor. A plain-vanilla low-cost diversified portfolio remains extremely hard to beat, and investors should be skeptical of products that promise too much.

Data Points: Weighted coin example: 60% heads vs. 50/50 - Used to illustrate how long-run evidence can reveal a true underlying premium despite short-term noise. Long-run evidence horizon: 50 to 100 years - Matt says meaningful confidence in factor research requires decades, not a few years. Number of documented academic factors: About 200 - He cites the 'factor zoo' and says many documented factors are redundant. Meaningful factor count: About 5 - Matt argues only a handful of factors have real incremental value after accounting for overlap. Value-stock turnover / migration: About 25% turnover a year - He uses this to describe the slow migration of value and profitability exposures over time. Momentum turnover: 200-300% - Matt contrasts momentum’s rapid turnover with slower-moving value and quality signals. Value exposure tilt: A couple percentage points - Longview only modestly increases or decreases value exposure when spreads are wide or tight. Bottom-end exclusion: Bottom 5% of the market - The firm excludes the worst-looking companies, often lottery-like small caps with no profits. Implementation frequency: Every day / roughly quarterly review - They monitor spreads continuously but typically reassess tilts on a slower cadence to manage trading costs. Value premium timing context: Above one standard deviation from historical average - When valuation spreads are in the widest third, the firm leans in more to value. Small-cap value opportunity set: About 3,000 U.S. stocks in the investable universe - He references the broad U.S. equity opportunity set when discussing concentration and liquidity.

Pivotal Quotes: "You can't eat value exposure, right? But you can eat returns." — Host / opening framing: Sets up the episode’s central distinction between having factor exposure and actually earning realized performance. "If something worked all the time, you wouldn't be paid to invest in it." — Matt: Explains why factor premia must be intermittent and why underperformance does not invalidate a long-run edge. "The traditional 60-40 portfolio is a fantastic portfolio. It is unbelievably diversified. It is low cost, liquid, tax efficient." — Matt: Used in his closing critique of chasing uncorrelated alternatives without a compelling improvement to the portfolio.

Implications: Listeners should expect factor investing to be a long-game, implementation-sensitive discipline, not a short-term signal chase. For the industry, the edge increasingly lies in execution, tax efficiency, and disciplined tilting rather than flashy new models or AI-generated factor discovery.