RMW Explained: The Profitability Factor, Out-of-Sample Evidence, and Avantis vs Dimensional

Factor premia are historical averages with wide return distributions and meaningful out-of-sample uncertainty.

RMW (Robust Minus Weak) is the profitability factor in Fama-French’s five-factor model. The simplest reading: profitable companies tend to earn higher subsequent returns than otherwise similar unprofitable companies, and that information is not fully captured by market beta, size, or value alone. The richer reading is what this guide is about: where the evidence is strong, where it is weaker than the marketing implies, why “quality at the operating-business level” is not the same thing as “quality vibes,” and how AVUV, AVDV, DFSV, and DFSVX translate the academic factor into long-only retail portfolios.

The most useful one-line reframe is Robert Novy-Marx’s title: profitability is the other side of value. Value asks how much business you get per dollar invested. Profitability asks how productive the business you’re buying actually is. Owned together, the two are complements; owned in isolation, neither tells you the full story.

What RMW Means: The Academic Version

RMW is a factor-mimicking portfolio: long robust-profitability stocks and short weak-profitability stocks, controlling for size. The construction is mechanical and diversified, which makes the “buy good companies” reading misleading. Kenneth French’s data library defines it precisely: “the average return on the two robust operating profitability portfolios minus the average return on the two weak operating profitability portfolios”, constructed as 1/2 (Small Robust + Big Robust) - 1/2 (Small Weak + Big Weak). Kenneth French data library.

The accounting definition uses operating profitability: revenues minus cost of goods sold, minus interest expense, minus selling/general/administrative expense, divided by book equity. NYSE percentile breakpoints split the universe into robust (top 30%), weak (bottom 30%), and neutral (middle 40%). The construction is mechanical, diversified, and rebalanced annually. That makes RMW useful for explaining returns. It is not identical to what a retail investor buys through a “quality” ETF.

Why operating profitability and not, say, net income or ROE? Because operating profitability sits closer to the actual productive economics of the business and farther from accounting items more easily manipulated by leverage, taxes, or one-time charges. It is the variable Fama and French landed on after a decade of work testing alternatives.

The Intuition: Profitability and Value Are Complements

Novy-Marx’s 2013 paper The Other Side of Value: The Gross Profitability Premium in the Journal of Financial Economics (108(1), pp. 1-28) made this case rigorously: gross profits-to-assets had roughly the same power as book-to-market in explaining the cross-section of average returns, and gross profitability was complementary to book-to-market, contributing economically significant information above what valuation alone provides. The paper won the 2013 Fama-DFA Prize for the best capital markets paper in the JFE.

The economic intuition comes from valuation math: holding valuation and investment policy constant, a company expected to produce more profits from the same capital base should have higher expected returns. Both value and profitability are designed to acquire productive capacity cheaply, but through different mechanisms. Value finances the purchase of inexpensive assets by selling expensive ones. Profitability finances the purchase of productive assets by avoiding unproductive ones. The two are arithmetic complements, not competitors.

What the Historical Evidence Shows

Fama and French’s 2015 paper A Five-Factor Asset Pricing Model formalized RMW as part of an extended model that added profitability and investment to the 1992 three-factor model. Fama & French (2015). Adding profitability and investment improved the model’s ability to describe average stock returns. It also exposed a notable failure case: small firms that invest aggressively despite weak profitability had especially low returns the five-factor model still struggled to explain.

Internationally, Fama and French (2017) International Tests of a Five-Factor Asset Pricing Model in the JFE (123(3), pp. 441-463) tested the model in North America, Europe, Japan, and Asia Pacific. Average returns increased with profitability and decreased with investment in three of the four regions. Japan was the exception: “the relation between average returns and B/M is strong, but average returns show little relation to profitability or investment.” That regional inconsistency matters. A factor that works in most developed markets but not Japan is not a universal law of finance; it is a pattern with regime dependence.

Out-of-Sample Evidence: The Credibility Section

This is where most RMW content stops. The answer is that RMW has stronger support than the average factor-zoo anomaly, but the broader literature on cross-sectional predictors should temper any claim that profitability is a guaranteed premium.

Sunil Wahal’s 2019 The Profitability and Investment Premium: Pre-1963 Evidence in JFE (131(2), pp. 362-377) hand-collected accounting data from Moody’s Manuals for 1940-1963 and tested the factors in pre-Compustat data. The result was partly supportive: the profitability premium was robust controlling for value, while the investment premium was much weaker. In spanning tests, RMW and HML improved the tangency portfolio; CMA did not. That is genuinely encouraging for RMW specifically.

Other work cuts the other direction. Linnainmaa and Roberts (2018) The History of the Cross-Section of Stock Returns in the Review of Financial Studies (31(7), pp. 2606-2649) examined 36 anomalies including value, profitability, and investment over 1926-2015. Their finding: most accounting-based return anomalies are at least partly spurious, with average returns and Sharpe ratios decreasing when examined out of sample. Profitability survives somewhat better than the median anomaly in their work, but the broad finding is sobering for any factor-based pitch.

Harvey, Liu, and Zhu (2016) ...and the Cross-Section of Expected Returns in the RFS (29(1), pp. 5-68) made the multiple-testing case: because hundreds of factors have been tested over the decades, a new factor should clear a much higher statistical hurdle than the traditional t-statistic of 2.0. They suggest a t-statistic above 3.0 as the credible threshold. RMW clears that bar in the original Fama-French samples; many of its factor-zoo cousins do not.

McLean and Pontiff (2016) Does Academic Research Destroy Stock Return Predictability? in the Journal of Finance (71(1), pp. 5-32) found that cross-sectional predictors declined out of sample by about 26% on average and declined by about 58% post-publication, consistent with some combination of data mining and arbitrage. The post- publication decay is not a fatal blow to factor investing, but it argues for humility about premium magnitude going forward.

The fair summary: profitability is one of the more durable factors in the literature, but factor research is noisy, publication-sensitive, and vulnerable to implementation details. Treat any specific premium estimate as a distribution, not a point.

How to Access RMW Exposure

There is no clean “RMW ETF” that perfectly owns the academic long-short factor. Individual investors get profitability exposure in three imperfect ways.

• Integrated multifactor funds. Avantis (AVUV, AVDV, AVGV) and Dimensional (DFSV, DFSVX) build long-only small-cap value portfolios that overweight higher-profitability companies and underweight cheap companies with weak operating economics. These are the academically cleanest retail implementations and are covered in detail in the next section.
• “Quality” ETFs. Funds like QUAL target profitable, low-leverage, stable-earnings companies. Useful as a convenience product but not the academic factor; many quality ETFs load on large-cap growth more than on RMW. The marketing word “quality” covers a wider range of strategies than the academic measure.
• DIY screens. Building a personal profitability screen is usually a bad fit. The academic factor is diversified, rebalanced, long-short, and implementation-aware. A homemade screen tends to become concentrated, tax-inefficient, expensive to trade, and accidentally loaded on sector or growth risk.

Whichever fund you choose, the empirical question worth answering is whether it actually has positive RMW loading. Many funds marketed as “quality” load on large-cap growth more than profitability. Summitward’s factor regression tool runs Fama-French and Carhart regressions against historical fund returns and reports the loadings on each factor, including RMW. That is the empirical way to verify whether the marketing matches the math.

Avantis vs Dimensional: How They Translate RMW

AVUV, AVDV, DFSV, and DFSVX do not buy RMW directly. They build long-only small-cap value portfolios that overweight or favor companies with low relative prices and higher profitability, and underweight or exclude some lower-profitability or higher-price stocks. That gives them positive profitability/RMW loading without shorting weak names. The two firms diverge on how to measure profitability, and the disagreement is the most interesting practitioner content in this space.

Avantis: cash-based operating profitability + joint ranking

Avantis prospectus language defines high-profitability companies as those with higher cash-based operating profitability. In Avantis’s public framework, profitability is not a standalone screen; it is combined with valuation into a joint metric that weights stocks from highest to lowest expected return. The ranking determines how much each stock is overweighted or underweighted relative to market-cap weight.

Avantis cites Ball and others showing that cash-based operating profitability may better identify persistent profitability and avoid accounting noise (especially around accruals). The argument is that cash flow from operations less working-capital changes captures business productivity more reliably than reported operating income.

Dimensional: operating profitability is sufficient

Dimensional defines profitability as operating income before depreciation and amortization minus interest expense, scaled by book equity. Dimensional has published research arguing that cash-based profitability does not contain more information about future profitability and returns than operating profitability once other expected-return drivers are accounted for. Dimensional: Do Accruals Adjustments Help Capture the Profitability Premium? Their argument is that what looks like a cash-profitability advantage may actually be an indirect underweight to high-investment firms. If a strategy already incorporates investment characteristics, the cash-vs-operating distinction adds little.

The comparison

Both firms are systematic active rather than passive index funds, and both are credible research-driven implementations. The choice between them comes down to fee, accessibility (Avantis ETFs trade openly; Dimensional ETFs do too, while DFSVX requires advisor access), and which firm’s philosophy the investor is willing to hold for decades. The cash-vs-operating profitability disagreement explains a small fraction of return differences, not the whole choice.

Try It: The Profitability Tilt + Tracking-Error Calculator

The calculator below estimates the expected uplift of a profitability tilt and the realistic tracking-error window you would face. The verdict tile is a behavioral filter, not an alpha forecast. Most investors who say “I would hold through 5 years of underperformance” later say otherwise in real time.

Who RMW Exposure Is For

• Investors with a diversified core portfolio who already believe in evidence-based factor investing.
• Households that can tolerate years of tracking error without abandoning the tilt at the bottom.
• Investors who want a quality-control filter inside value and small-cap value strategies (the cleanest RMW use case).
• Long-horizon accumulators with stable income and broad household-level diversification.

Who It Is Not For

• Investors who need short-term benchmark-like performance.
• Anyone likely to abandon a factor after three or five bad years (which is most people in real time).
• DIY stock-pickers expecting profitability to act as a single-stock alpha screen. The factor is diversified, rebalanced, long-short. A concentrated profitability bet is not the same exposure.
• Anyone using profitability tilts as a substitute for emergency funds, asset allocation, tax planning, or household risk management.

Frequently Asked Questions

Is RMW the same as buying high-quality stocks?

Not exactly. The academic factor is a long-short portfolio sorted on operating profitability. Retail “quality” ETFs use varying definitions and often load more on large-cap growth than on the academic RMW factor. Verify empirically with a factor regression rather than trusting the label.

Why does Japan break the model?

Fama and French (2017) found that Japanese average returns relate strongly to book-to-market but show little relation to profitability or investment. The reasons are debated; candidates include Japan’s 1990-2015 equity-premium anomaly, regulatory and corporate-governance differences, and sample-period idiosyncrasies. The honest takeaway is that RMW is a regularity that holds in most developed markets, not a universal law.

How big is the profitability premium?

Long-run academic estimates run from roughly 1.5% to 4% per year for the long-short academic factor, with wide standard errors. Retail long-only fund implementations capture a loading of roughly 0.20 to 0.40 on that factor. The realized excess return for an investor depends on premium realization, loading, fees, taxes, and behavior. The calculator above lets you stress-test your own assumptions.

Does RMW make sense in a tax-advantaged account vs taxable?

Slight preference for tax-advantaged. Profitability-aware small-cap value funds often have higher turnover and distribute realized capital gains more frequently than broad-market index funds. A Roth IRA, traditional 401(k), or HSA shields that turnover. In taxable, it is workable but requires watching distribution profiles. See You Have One Household Portfolio for asset-location framing.

Should I use Avantis or Dimensional?

Both are credible. Avantis ETFs are typically cheaper and more accessible for DIY investors. Dimensional has a longer live-strategy history and a deeper integrated-process pedigree but is partly advisor-gated for some products. Don’t agonize over the cash-vs-operating profitability disagreement; it explains a small fraction of expected-return differences. The bigger decision is whether you want a persistent small-value/profitability tilt at all.

How do I check whether a fund actually has RMW loading?

Run a factor regression. The dependent variable is the fund’s monthly excess return; the regressors are the Fama-French factors (market, SMB, HML, RMW, CMA) and optionally a momentum factor. Statistically significant positive coefficient on RMW means the fund has profitability exposure; a coefficient near zero means the fund is marketing “quality” without delivering it. Summitward’s factor regression tool does this for any fund.

Related Guides

• Fama-French Factors covers the broader five-factor model (market, size, value, profitability, investment) and the Carhart momentum extension.
• AVUV, AVDV, and AVGV covers the broader case for Avantis small-cap value funds beyond the profitability dimension.
• Small-Cap Value covers the historical evidence for the small-cap value premium and the integrated case for tilting.
• Modern Portfolio Theory for Real Life covers the marginal-fund test for evaluating any new tilt against your existing portfolio.
• You Have One Household Portfolio covers asset-location decisions for tilts: typically place higher-turnover factor funds in tax-advantaged accounts.

Key Takeaways

• RMW is “quality at the operating-business level,” not “quality vibes.” The academic factor is mechanical, diversified, and built on operating profitability divided by book equity.
• Profitability is the other side of value, not a substitute. Novy-Marx’s framing: both target acquiring productive capacity cheaply, through different mechanisms. Owned together they help avoid value traps and overpriced quality.
• The evidence is stronger than average but not bulletproof. Wahal’s pre-1963 work supports profitability; Linnainmaa-Roberts shows accounting anomalies fade out of sample on average; Harvey-Liu-Zhu argue for t>3 hurdles; McLean-Pontiff show post-publication decay. Treat the premium as a distribution, not a point.
• AVUV, AVDV, DFSV, and DFSVX translate RMW into long-only portfolios. Avantis uses cash-based operating profitability and a joint value/profitability ranking. Dimensional uses operating profitability and argues cash-based adds little once investment characteristics are accounted for. Both are credible.
• Verify factor loadings empirically. Marketing language for “quality” covers a wide range of strategies. A factor regression on monthly returns tells you whether a fund actually has RMW exposure. Summitward’s factor regression tool handles this.