The Value Factor: Price-to-Book versus Composites, the 2007–2020 Drawdown and Value Traps

What you will learn

• How value is measured: price-to-book alone versus a composite of several cheapness ratios.
• Why intangible assets weaken book value as a measure, and one way to adjust for it.
• What the 2007–2020 value drawdown looked like and what the research says caused it.
• How a value trap differs from the value premium and which filters reduce trap risk.

Explanation

The value factor is the historical tendency of shares that are cheap relative to a fundamental anchor to earn higher average returns than shares that are expensive relative to the same anchor. The anchor is the key design choice. Fama and French used book-to-market, the accounting book value of equity divided by market capitalisation; it is the inverse of the price-to-book ratio that you already know from the stock analysis course. Book value was chosen because it is stable and widely available, not because it is the best measure of worth.

Book value has a weakness that has grown over time: under accounting rules, spending on research, software, brands and customer acquisition is mostly expensed, not recorded as an asset. A company built on intangible investment therefore shows a small book value and looks "expensive" on price-to-book even if it is cheap relative to earnings or cash flow. Several researchers, including Arnott, Harvey, Kalesnik and Linnainmaa in a 2021 Financial Analysts Journal article, show that capitalising intangibles (treating past research spending as an asset that depreciates over several years) produces a value measure that classified stocks more sensibly over recent decades.

Because of this, most index providers and quantitative managers use a composite value score: they rank companies on several ratios, such as book-to-price, earnings-to-price, cash-flow-to-price, sales-to-price and EBITDA-to-enterprise-value, and average the ranks or standardised scores. A composite is more robust because any single ratio can be distorted by a one-off item, a loss year or an accounting quirk.

Now the drawdown. Using the standard book-to-price HML series, the same 2021 article measured a peak-to-trough fall of about 55 percent from 2007 to June 2020, the largest since the series starts in 1963. Their decomposition attributes the whole loss to a change in the valuation spread: value stocks became relatively cheaper and cheaper compared with growth stocks, moving from the top quartile of their historical relative valuation in early 2007 to the bottom percentile by June 2020. In plain words, value underperformed mainly because the market kept paying more for growth and less for value, not because value companies did worse as businesses. A spread can widen for a long time, and nothing forces it to close on a schedule.

Finally, the value trap: an individual share that looks cheap because its business is deteriorating, so the denominator (earnings, book value) is about to fall. The value premium is a statistical result across hundreds of cheap stocks, some of which are traps. A diversified value portfolio accepts traps as part of the cost; a concentrated stock picker cannot. Two filters reduce trap exposure: requiring positive, stable profitability (lesson 4) and avoiding shares in a strong downtrend (lesson 3).

Worked example

Hypothetical example: four companies, all figures invented and rounded. We rank each on three cheapness measures, where rank 1 is cheapest, and average the ranks.

Ranks on B/P (higher is cheaper): A 1, D 2, B 3, C 4. Ranks on E/P: B 1, D 2, C 3, A 4. Ranks on S/P: D 1, A 2, B 3, C 4.

• A: (1 + 4 + 2) ÷ 3 = 2.33
• B: (3 + 1 + 3) ÷ 3 = 2.33
• C: (4 + 3 + 4) ÷ 3 = 3.67
• D: (2 + 2 + 1) ÷ 3 = 1.67

Price-to-book alone would put A first. The composite puts D first and demotes A, because A's earnings yield of 2.5% signals that its low book multiple may reflect weak profits, a classic trap warning. C is expensive on every measure. The ranking involves no fees, taxes or judgement about business quality.

Common mistake

Practice activity

A hypothetical software company has a market capitalisation of €1,200 million and a reported book value of €400 million. It has expensed €60 million of research per year for five years. Capitalise the research with straight-line depreciation over five years (the latest year counts 100%, then 80%, 60%, 40%, 20%) and calculate the reported and adjusted price-to-book ratios.

Key takeaways

• Value needs an anchor; book value is stable but understates intangible-heavy businesses.
• Composite scores across several ratios are more robust than any single multiple.
• The book-to-price HML series fell about 55% from 2007 to June 2020, driven mainly by a widening valuation spread.
• Value traps are individual failures inside a statistical premium; profitability and trend filters reduce them.

Next step

The next lesson turns to momentum: how the 12-minus-1 signal is built, why it crashed in 2009 and why it is expensive to hold.