Financial Theory in Practice: Gaining Insight from Models with Marlena Lee (EP.79)

Episode Summary

Executive Summary: Marlena Lee of Dimensional explains how evidence-based investing uses imperfect models to identify drivers of expected returns, why factor premiums are real but noisy, and how implementation matters as much as theory. She discusses value, size, profitability, momentum, bonds, alternatives, and market timing, emphasizing diversification, discipline, and realistic expectations over chasing headlines or simplistic rules.

Main Topics: Models and their limits in investing: Lee argues that models are simplified tools, not perfect descriptions of reality. Investors should choose the right model for the right job and avoid over-interpreting any single framework or dataset. What factors are and how Dimensional uses them: She defines factors as systematic return drivers and explains that many purported factors are overlapping variations on a few themes. Dimensional focuses on understanding mechanisms, timing, geography, and interaction effects before using them in portfolios. Why value, size, and profitability matter: Lee explains Dimensional’s preference for book-to-market, size, and profitability as robust, overlapping signals of expected returns, and why alternative valuation metrics often add little once those are included. Persistence, noise, and investor discipline: She emphasizes that premiums can persist even when they underperform for long periods. The key is setting expectations and staying invested through noisy stretches rather than trying to time reversals. Dividend stocks, fixed income, and implementation trade-offs: Lee argues dividend strategies often just package value and profitability exposure with reduced diversification, while fixed income should be analyzed using observable yields and forward rates rather than noisy proxies. Alternatives and market timing skepticism: She is skeptical that liquid alternatives add diversification, is cautious on private equity due to fees and manager dispersion, and says yield curve inversions and similar signals are not reliable timing tools. Human capital, bonds, and life-cycle allocation: For young investors, she notes that human capital is a major bond-like asset, so equity-heavy portfolios can be sensible if they can be tolerated; leverage is theoretically possible but usually impractical.

Key Arguments: Financial models are inherently incomplete, but can still be useful if matched to the problem being solved. Factor investing should be based on understanding drivers of expected returns, not just counting variables or fitting data. Many ‘different’ factors are variations on a few common themes, so Dimensional tries to avoid redundant signals. Value, size, and profitability provide complementary information; earnings-to-price and cash-flow metrics often overlap with those exposures. Higher expected returns are not guaranteed; investors must tolerate tracking error and long stretches of underperformance. Theoretical rationale for the value premium is evergreen: lower prices should imply higher expected returns. Dividend-oriented portfolios mainly repurpose value/profitability exposure and can reduce diversification; yield preference is mostly behavioral. Fixed income is different from equities because yields are directly observable, so investors should use the observable rate rather than noisy proxies. Liquid alternatives usually do not expand the investment universe enough to justify calling them diversifying. Private equity may expand the universe, but fees, idiosyncratic manager risk, and selection difficulty make its value uncertain. Market timing based on yield curves or macro signals is not supported by evidence and requires two correct decisions to succeed. Young investors often already hold a large bond-like asset in the form of human capital, allowing more equity risk if they can stay disciplined.

Data Points: Dimensional global AUM: almost $800 billion Canadian globally - Describing Marlena Lee’s role and the scale of Dimensional Fund Advisors Podcast episode number: 79 - Rational Reminder’s first episode of 2020 Podcast milestone target: 100th episode in July - Hosts discuss a possible 100-episode celebration Dimensional tenure: 11 years - Lee has worked at Dimensional for 11 years Momentum turnover in academic literature: 200% to 300% - Lee contrasts academic momentum portfolios with live implementations Momentum turnover in live funds: 75% to over 100% - Live fund turnover is materially lower than academic examples Value premium assumption in Fama-French paper: 29 basis points per month - Used to illustrate how even a positive premium can still underperform over long horizons Value premium annualized assumption: roughly 3.5% annually - Converted from the monthly premium in the volatility discussion Chance of value underperforming the market over 10 years: more than 5% - From the Fama-French volatility illustration U.S. period of equities underperforming T-bills: 17 years - Example showing even the broad equity premium can be negative for long stretches SP 500 Dividend Aristocrats constituent count: just over 50 - Used to show dividend strategies can be less diversified Endowments studied: over 800 U.S. endowments - Referenced in a study on alternatives allocation Endowment AUM studied: over $600 billion USD - Scale of the endowment dataset Alternative allocation in endowment aggregate portfolio: half of portfolio - Includes private equity, hedge funds, private real estate, and natural resources 10-year endowment return: 5.8% annualized - Average endowment return ending June 2018 Global 60/40 return: 6.8% annualized - Benchmarked against average endowment performance Private equity manager dispersion: over 40% cross-sectional standard deviation - Illustrates huge variation across private equity managers

Pivotal Quotes: "We think of models as something that's useful to simplify reality, but hopefully they're still capturing some salient features of how something works." — Marlena Lee: Opening explanation of why all investment models are incomplete but useful "If you're hoping to add value by always being faster than the rest of the market, that seems to be a potentially more tenuous proposition." — Marlena Lee: Discussing machine-learning-style factors and the limits of speed-based advantages "If an asset allocation is not one that an investor can stick with and stay invested during a period of disappointing performance, then it wasn't the right asset allocation." — Marlena Lee: On long-term discipline, volatility, and choosing an implementable portfolio

Implications: Listeners should expect factor premiums to be real but noisy, and build portfolios they can actually hold through underperformance. The episode reinforces evidence-based implementation over trend-chasing, skepticism toward timing and alternatives, and the importance of matching strategy to behavior and goals.