Traditional “value“ investing (book-to-market, HML) has struggled in modern factor models. In the Fama-French five-factor framework, the classic value factor can become redundant once profitability and investment are included. So what should investors and researchers look at instead?
This post explains an alternative that ties stock prices to economic value creation: the Value-Creation Pricing Factor, built from value creation-to-market (V/M) and implemented as a factor-mimicking portfolio called CME (Cheap Minus Expensive). For background, see full Master’s thesis.
If you want the academic foundations, start with Fama & French’s five-factor model paper (Journal of Financial Economics) and the official factor documentation at the Kenneth R. French Data Library.
Table of contents
- The problem with “value” in modern factor models
- The big idea: measure economic value creation
- What is V/M (value creation-to-market)?
- CME explained: Cheap Minus Expensive
- What the data says (1963–2018): a large, significant premium
- Why this works: linking corporate finance to asset pricing
- How practitioners can use the concept (without overfitting)
- FAQ
- Further reading
The problem with “value” in modern factor models
In classic factor investing, “value” often means buying stocks with high book-to-market (cheap relative to accounting book value) and shorting low book-to-market (expensive). In the Fama-French ecosystem, that’s the HML factor.
But in their five-factor model, Fama & French show that profitability and investment can absorb much of what HML used to explain, making HML look redundant in some settings. (If you haven’t read it, the JFE version is here: A five-factor asset pricing model.)
That creates a practical dilemma:
- Investors still want a “cheap vs expensive” signal.
- But book-to-market is an imperfect proxy for what we actually mean by cheap.
- Profitability helps – but profitability alone can reward firms that grow profits by burning capital.
What if “cheap” should mean: a firm creates lots of economic value relative to what the market is pricing in?
The big idea: measure economic value creation
Accounting profits don’t ask the most important question in corporate finance:
Did the company earn more than its cost of capital?
Economic value creation is built on a simple logic: a firm creates value when the return on invested capital exceeds the weighted average cost of capital.
Value creation (conceptual formula)
Value Creation = (ROIC − WACC) × Invested Capital
This helps avoid a common trap: some firms can look “profitable” while destroying value if they require huge capital to generate those profits. (That’s the difference between accounting performance and economic performance.)
This thinking also connects nicely to the broader profitability literature – e.g., Novy-Marx’s work on profitability as a return predictor (JFE link; working paper PDF: OSoV.pdf).
What is V/M (value creation-to-market)?
Once you have a measure of value creation (V), you can scale it by the market value of equity (M) to create something that behaves like a “valuation” signal – except it’s anchored in economic value creation:
V/M = Value Creation ÷ Market Equity
Intuitively, V/M asks:
- Is the market paying a lot for each unit of economic value creation? (Low V/M → “expensive”)
- Or is the market paying relatively little for strong value creation? (High V/M → “cheap”)
One helpful mental model is that V/M acts like a discount-rate / expectations lens: it blends fundamentals (value creation) with how optimistic or skeptical the market is (market equity).
CME explained: Cheap Minus Expensive
To turn V/M into a tradeable factor, you form portfolios based on how “cheap” or “expensive” stocks are relative to economic value creation – then take the return spread:
CME (Cheap Minus Expensive) = returns of high V/M (“cheap”) stocks minus returns of low V/M (“expensive”) stocks.
This is conceptually similar to how classic factors are built (HML, RMW, CMA, etc.) and fits naturally into the same modeling toolkit. If you want to see how Fama-French build and describe their factor sets, the official descriptions are here: Fama/French factor definitions.
Plain-English translation:
CME tries to buy companies that create a lot of economic value relative to their price, and short companies priced richly despite weak economic value creation.
What the data says (1963–2018): a large, significant premium
When stocks are sorted by V/M and the high-minus-low spread is formed as a factor (CME), the research finds a strong and statistically significant premium over the US sample spanning 1963–2018.
Headline result (high level):
Sorting on V/M produces a large return spread; “cheap” (high V/M) outperforms “expensive” (low V/M) by an annualized premium of about 6.29%, with very strong statistical significance.
Importantly, the effect is not just a tiny microcap story. The premium is pronounced for smaller stocks but remains meaningful for larger stocks as well.
Why this matters for factor models
When you evaluate models on diversified test portfolios (the standard asset pricing approach), a key question is: Does adding a factor reduce unexplained return patterns?
In model-comparison terms, adding CME can improve the description of returns especially for portfolios where the classic models struggle – such as portfolios formed on size + investment, size + V/M, and multi-sorted portfolios involving profitability and investment.
These are typically evaluated using time-series regressions and joint tests on intercepts (alphas). One well-known joint test framework is the GRS test (Gibbons, Ross & Shanken). If you want the original reference, you can find a PDF here: GRS (1989) Econometrica PDF.
Why this works: linking corporate finance to asset pricing
V/M is powerful because it compresses several important economic ideas into one lens:
- Profitability (but with an economic focus)
- Capital discipline (profits relative to capital employed)
- Cost of capital (capital isn’t free)
- Market expectations (the market price embeds beliefs about future opportunities)
This helps answer two questions that traditional “value” can struggle with:
- When is a productive company actually expensive? (High value creation but even higher market expectations.)
- When is an unproductive company still cheap? (Low value creation but an even more depressed price.)
In other words, it’s a “value” signal that tries to be honest about what investors should care about: economic surplus relative to price.
How practitioners can use the concept (without overfitting)
You don’t need to rebuild a full academic factor library to benefit from the idea. Here are practical, low-drama ways to apply it:
1) Upgrade your definition of “cheap”
Instead of ranking stocks by book-to-market alone, consider ranking by an economic spread like (ROIC − WACC), then relating it to price/market cap. The goal is not precision to the fourth decimal – it’s reducing category errors (e.g., “cheap” companies that are actually value destroyers).
2) Combine with quality screens thoughtfully
V/M is already tied to economic profitability, but in real portfolios you may want simple guardrails: avoid extreme leverage, watch for one-off accounting spikes, and sanity-check that the “value creation” is repeatable.
3) Keep the process boring
- Use long lookback windows.
- Prefer robust rebalancing (e.g., quarterly or annual).
- Avoid constant tweaking. If you change your recipe every time the market changes, you’re probably just fitting noise.
Important: This post is educational and not investment advice. Real-world implementation involves data cleaning, survivorship-bias controls, transaction costs, constraints, and risk management.
FAQ
Is CME just profitability or investment in disguise?
It’s related – but not identical. The point of V/M is that it integrates profitability with the cost of capital and the market’s expectations. In spanning-regression language, the goal is that CME adds something not fully replicated by the other factors.
Does this replace value investing?
Think of it as a better definition of value: not cheap versus book value, but cheap versus economic value creation.
Where can I get factor data to learn the basics?
The easiest starting point is the Kenneth R. French Data Library. It provides the classic factor returns and many portfolio sorts used in academic research.
Further reading
- Fama, Eugene F. & French, Kenneth R. “A five-factor asset pricing model” (JFE). ScienceDirect page.
- Kenneth R. French Data Library (factors, sorted portfolios, documentation). Data library.
- Novy-Marx, Robert. “The other side of value: The gross profitability premium” (JFE). ScienceDirect page; PDF: OSoV.pdf.
- Gibbons, Ross & Shanken (1989) joint test of intercepts (GRS). PDF: Econometrica paper PDF.
Bottom line: If you believe markets price more than accounting ratios, a value signal grounded in economic surplus – value creation relative to market equity – can be a cleaner way to define “cheap” and a compelling complement (or alternative) to traditional value factors.
