Undervalued Assets: How to Identify Them, and Why the Market Disagrees

Most guides to undervalued assets stop at the screen: low P/E, low price-to-book, buy. That finds cheap assets. It does not find undervalued ones. Cheap is a ratio you can read off a data terminal. Undervalued is a claim that the price sits below what the asset will actually earn, and it needs two more things before it means anything: a reason the market has priced it wrong, and an event that would show it. Without those, a low multiple is usually the market’s correct verdict on a business in trouble.

Below: the four multiples and the trap in each, what a century of US data says about buying cheap stocks (including a drought that began in 2007 and has not fully reversed), and how to tell an undervalued sector from a permanently cheap one. The long-run figures are our own, calculated from the Kenneth R. French Data Library with data to July 2026.

The four multiples, and the trap in each

Every valuation screen uses some version of these four. Each compares a price with one line of the accounts, and each fails in a predictable way.

Metric	What it compares	A low reading suggests	The trap	The check that catches it
P/E	Share price ÷ earnings per share	You pay little for each dollar of profit	Cyclical companies look cheapest at the peak of their earnings cycle, just before profits fall. One-off gains also flatter the “E”.	Use earnings averaged over a full cycle; strip out one-off items
EV/EBITDA	Market value plus net debt ÷ operating earnings before depreciation	The whole business, debt included, is priced low	EBITDA ignores capital spending. A business that must reinvest most of its cash to stand still looks cheap and is not.	Compare EBITDA with capital expenditure; look at EV/EBIT or free cash flow instead
P/B	Market value ÷ book value of equity	You pay close to what the assets are recorded at	Book value records what was paid, not what the assets can earn. Obsolete plant, doubtful loans and goodwill all inflate it. Brands, software and research barely appear in it.	Ask whether the assets earn their cost of capital; adjust for write-downs and intangibles
FCF yield	Free cash flow ÷ market value	The business produces a lot of cash relative to its price	One year of free cash flow can be flattered by cutting investment or squeezing suppliers. A high yield on shrinking cash flow is a melting ice cube.	Look at several years; ask whether investment is below depreciation

No single ratio is right. P/B works tolerably for banks and insurers, whose assets are mostly financial, and badly for software companies. EV/EBITDA suits capital-heavy businesses with different debt levels. FCF yield is closest to what an owner receives. The practical consequence: use at least two, and treat any disagreement between them as the first question to answer.

What “undervalued” actually means

An undervalued asset is one whose price is below a defensible estimate of its value. That estimate is the present value of the cash the asset will produce, discounted at the return you require. Both inputs are assumptions, so “undervalued” is always conditional on a model, never a property of the price alone.

Take a stock at $50 expected to generate $4.00 of free cash flow per share next year, an 8.0% FCF yield. If that cash flow grows at a steady rate forever, its value is next year’s cash flow divided by required return minus growth.

Required return	Growth 0%	Growth 1%	Growth 3%
8%	$50.00 (0%)	$57.14 (+14%)	$80.00 (+60%)
9%	$44.44 (−11%)	$50.00 (0%)	$66.67 (+33%)
10%	$40.00 (−20%)	$44.44 (−11%)	$57.14 (+14%)

The same stock is worth 60% more than its price or 20% less, depending on two inputs nobody knows to within two percentage points. So run the model backwards. At $50 and a 9% required return, the price implies cash flow growing about 1% a year forever. The question becomes “is 1% too pessimistic, and why?”, and that one can be checked against the business.

This is also what separates an undervalued asset from an underperforming one. A stock that has fallen 40% is underperforming. It is undervalued only if the fall exceeded the damage to its future cash flow; often the price fell because the cash flow did. That is a fundamental question: in the split between technical and fundamental analysis, price trends tell you what other participants are doing, not what the business is worth.

A century of buying cheap stocks

The systematic version of this idea is the value premium: stocks with high book-to-market ratios (value) have historically earned more than stocks with low ones (growth). Eugene Fama and Kenneth French documented it in 1992, and French’s data library publishes the portfolios monthly. The table uses the value-weighted portfolios of the cheapest 30% and most expensive 30% of US stocks by book-to-market, with NYSE breakpoints, rebalanced each June.

Period	Value (cheapest 30%)	Growth (dearest 30%)	Difference	US market
1930s	−3.2%	1.6%	−4.8	−0.3%
1940s	17.0%	6.9%	+10.1	9.5%
1950s	22.0%	17.7%	+4.4	18.3%
1960s	12.1%	7.9%	+4.2	8.3%
1970s	13.7%	3.8%	+10.0	6.1%
1980s	20.6%	15.0%	+5.6	16.9%
1990s	16.4%	19.2%	−2.8	18.0%
2000s	6.0%	−1.2%	+7.2	−0.4%
2010s	11.2%	15.1%	−4.0	13.6%
Jan 2020–Jul 2026	18.2%	16.7%	+1.5	15.0%
Jul 1926–Jul 2026	13.2%	10.2%	+3.0	10.4%
Jul 1926–Dec 2006	14.0%	9.5%	+4.5	10.2%
Jan 2007–Dec 2020	5.3%	12.8%	−7.5	10.1%
Jan 2021–Jul 2026	22.4%	13.4%	+8.9	13.5%

Over the full century the cheap stocks compounded about 3 points a year faster than the expensive ones, and value won in six of the nine complete decades from the 1930s to the 2010s. That is the case for looking for undervalued assets at all. The rest of the table is the case for humility.

The drought: what 2007–2020 did to cheap stocks

The cleaner measure is HML (“high minus low”), Fama and French’s paper portfolio that buys value and sells growth among both small and large stocks, before costs.

Compounding HML month by month, it reached a peak in December 2006. From there it fell 57.8% to a trough in September 2020, 165 months later. Measured from the previous peak, it is the worst drawdown in the hundred-year record. HML was positive in only 36.9% of those months. In long-only terms, $1 in the value portfolio at the start of 2007 became $2.06 by the end of 2020; $1 in growth became $5.38, and $1 in the whole market $3.83.

Value has recovered part of the ground since. HML rose 66.8% from its 2020 trough to July 2026. It is still 29.6% below its December 2006 peak, 235 months later: almost twenty years without a new high. The only earlier drawdowns deeper than 30% were four in the Depression and war years (the deepest, −43.7%, took until March 1937 to recover) and one of −39.8% during the technology bubble, recovered by February 2001.

Rolling ten-year windows show how unusual this was. Of the 847 overlapping ten-year periods (one ending each month) that ended before 2007, value trailed growth in 10.3%. Of the 235 ending in 2007 or later, it trailed in 73.2%, with the worst ten years ending in August 2020 at −8.9 points a year. The ten years to July 2026 are roughly level, at −0.5 points a year.

Why it happened, and why it is not settled

Fama and French themselves, looking at July 1963 to June 2019, found value premiums much lower in the second half than in the first, but monthly returns too volatile to say reliably whether the expected premium had fallen. On the current data, HML averaged 5.0% a year in the first half (July 1963 to June 1991) and 2.5% in the second (July 1991 to June 2019), and the second figure is not statistically distinguishable from zero (t = 1.28).

Arnott, Harvey, Kalesnik and Linnainmaa (2021) attribute the drought to two things: book value misses the intangible assets that now make up much of what companies own, and value stocks got cheaper relative to growth stocks. In their decomposition, that change in the valuation spread explains the entire drawdown. Their measure of value’s relative valuation went from the top quartile of its history at the start of 2007 to the bottom percentile in June 2020.

The French files show the same revaluation. Divide the aggregate book-to-market of the value portfolio by that of the growth portfolio: the higher the number, the cheaper value is. At December 2006 prices it was 3.5, lower than in all but 7% of the years since 1927: value went into the drought unusually expensive. At December 2019 prices it was 11.0, the 96th percentile. At December 2025 prices it was still 9.9, the 94th percentile, against a 1927–2026 median of 4.9.

The practical consequence: an undervalued asset can become more undervalued for a very long time, and being right about the value does not pay until the price agrees. Any approach built on buying cheap has to survive a decade of being wrong, which is a question of position sizing before it is a question of analysis.

The hard question: why does the market disagree with you?

A value portfolio wins on average because a minority of its stocks do very well. In Piotroski’s study of US high book-to-market stocks (1976–1996), fewer than 44% earned a positive market-adjusted return in the two years after they were bought. The typical cheap stock trails the market; the portfolio average is carried by the few that recover. That is why the individual decision needs a reason the market is wrong, not just a low ratio.

So list why the price is low, and be honest about which reasons are mispricing and which are information.

Why it looks cheap	Mispricing or information?	What would prove you right
Earnings are at a cyclical peak	Usually information: the market is pricing the downturn	Profits holding up through the next downturn, measured against the prior cycle
The industry is in structural decline	Usually information: book value overstates what the assets can earn	Cash returned to shareholders exceeding what the price implies, even as revenue shrinks
Heavy debt or refinancing risk	Mixed: the equity is an option on survival	Refinancing completed, debt falling, interest cover rising
Accounting or governance doubts	Usually information until disproven	Clean audit, cash flow matching reported earnings, insiders buying
Forced or indifferent selling: index deletion, spin-off, small size, no analyst coverage	Can be genuine mispricing: the seller is not trading on value	Selling pressure ending; results reported without the old parent or index
Recent bad news extrapolated too far	Possible mispricing: the market over-weights the recent trend	Earnings stabilising while the price still assumes continued decline

The quality of the cheap business matters as much as the cheapness. Piotroski’s nine-point score uses simple accounting signals: positive and rising profitability, cash flow above reported earnings, falling leverage, no new shares issued. Selecting the financially strong high book-to-market firms raised the mean return by at least 7.5 points a year in his sample, and buying the strong while shorting the weak earned 23% a year from 1976 to 1996. That sample is thirty years old and the method is widely published, so read it as evidence that the quality filter matters, not as a return to expect.

Write down the catalyst and the exit before you buy

A catalyst is the event that would make the market see what you see: a cycle turning, a refinancing, a spin-off, an asset sale, a change in capital allocation. Without one, you are waiting for the market to change its mind unprompted, which is what value investors did from 2007 to 2020.

Write three things down at purchase: the thesis in one sentence, the catalyst and a rough date, and the evidence that would prove you wrong. The third matters most: the expensive habit with cheap stocks is holding a loser because selling would admit the mistake. That is the disposition effect, one of the biases with a measured cost, and a written exit condition is the cheapest defence against it. Averaging down because it is “even cheaper now” belongs on the list of stock market mistakes.

Undervalued sectors: cheap against whom?

Sectors add one trap. Some industries are cheap almost all the time because their assets are heavy, regulated or financial; comparing a bank’s price-to-book with a software company’s tells you about accounting, not mispricing.

We tested two definitions of a cheap sector on the 49 US industry portfolios in the Ken French library. Each July from 1963 to 2026, industries were ranked by aggregate book-to-market, using only data available at the time, and the cheapest fifth and dearest fifth were held for twelve months, industries weighted equally.

Definition of “cheap”	Cheapest fifth	Dearest fifth	Gap, full period	Gap 1963–2006	Gap 2007–2020	Gap 2021–Jul 2026
Against other industries (book-to-market)	12.2%	10.3%	+1.9	+2.9	−5.1	+12.3
Against its own history (book-to-market ÷ its 10-year median)	12.5%	9.8%	+2.6	+4.2	−2.3	+3.4

Two findings. First, cheapness against other industries is highly persistent: on average 54.8% of the industries in the cheapest fifth were still there five years later, against 20% if rankings were random. Steel and textiles were in the cheapest fifth in 50 of the 64 years, banks in 41, utilities in 38. A sector that is always cheap is not undervalued; it is priced for what it is. Second, measuring each industry against its own history gave a larger gap over the whole period and lost less in the 2007–2020 drought. The exception is 2021–2026, when the industries that were cheapest outright rallied hardest and the first definition did better.

This is the closest the data gets to how to identify undervalued sectors early. The criteria that follow from it:

• Compare the sector with its own history, not with other sectors. A valuation well below its own ten-year median is a signal; a valuation below the market average usually is not.
• Normalise earnings over the cycle. Cyclical sectors look cheapest on current earnings near the top. Use average margins over a full cycle.
• Watch supply, not just demand. Falling capital spending, closures and consolidation are what restore returns in a depressed industry. Rising investment at low prices usually means the pain is not over.
• Check the balance sheets. A cheap sector where the weakest firms can refinance is a different bet from one where they cannot.
• Accept that “early” is expensive. Finding a sector before the crowd means holding it while it gets cheaper. The 2007–2020 column is what early looked like for fourteen years.

None of this names a sector to buy. A published list of cheap sectors is a list of low prices the market has already seen.

How to use this without betting the account

Size for being early. A single undervalued stock can lag for years before its catalyst arrives. Each idea should be small enough that a long spell of being wrong is a setback, not an ending; the arithmetic is in position sizing and risk of ruin.

Judge results after risk. A cheap stock that doubled while swinging twice as hard as the market has not necessarily done better; the Sharpe ratio is the standard adjustment, with its limits.

Be honest about the alternative. Professionals who hunt for undervalued assets full time mostly trail their index after fees, value managers included over long horizons; see twenty-five years of SPIVA scorecards. If you pick undervalued assets yourself, ask what you know that they do not.

What this does not say

Not that value investing is dead: value has beaten growth by 8.9 points a year since January 2021 and remains unusually cheap on book-to-market. Not that it is back either. Book-to-market is a crude measure, the figures are before costs, and the sector test is one backtest in one country. None of it predicts the next decade.

Frequently asked questions

What does undervalued assets mean?
An undervalued asset is one whose market price is below a reasoned estimate of its value, usually the present value of the cash it will generate. Because that estimate depends on assumed growth and required return, “undervalued” is always a claim about a model, not a fact about the price.

How do you identify an undervalued asset?
Screen with at least two multiples (P/E, EV/EBITDA, price-to-book, free cash flow yield), then work out what growth the price implies. It is a candidate only if you can explain why that growth is too pessimistic and name an event that would make the market see it.

How do you identify undervalued sectors early?
Compare each sector with its own history rather than with other sectors, normalise earnings over a full cycle, and watch for falling capital spending and consolidation. In US industry data for 1963–2026, cheapness against an industry’s own ten-year median gave the larger gap and lost less in the 2007–2020 drought.

What is the difference between undervalued and underperforming assets?
An underperforming asset has fallen in price or lagged its benchmark. It is undervalued only if the price fell further than its future cash flows did. Many underperformers are correctly priced because the business deteriorated.

What is a value trap?
A value trap is a stock that looks cheap on its multiples but stays cheap or gets cheaper because the low price reflected real problems: peak earnings, structural decline, debt or poor governance. In Piotroski’s 1976–1996 sample, fewer than 44% of high book-to-market stocks beat the market over the following two years.

Does value investing still work after the 2007–2020 drought?
The US value factor (value minus growth) fell 57.8% from December 2006 to September 2020 and was still 29.6% below that peak in July 2026. Value has done better since 2021, but whether the long-run premium has shrunk is unresolved, even for the researchers who documented it.

Which ratio is best for finding undervalued stocks?
None works everywhere. Price-to-book suits banks and insurers, EV/EBITDA capital-heavy companies, and free cash flow yield is closest to what an owner receives. Use two or more and investigate where they disagree.

Sources

• Kenneth R. French Data Library, Tuck School of Business, Dartmouth: “Fama/French 3 Factors”, “Portfolios Formed on Book-to-Market” and “49 Industry Portfolios”, monthly returns July 1926–July 2026 and annual book-to-market ratios, files created from the CRSP database of July 2026, downloaded 25 September 2026. mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html
• Fama, E. F. and French, K. R., “The Cross-Section of Expected Stock Returns”, Journal of Finance, vol. 47, no. 2, June 1992, pp. 427–465. doi.org/10.1111/j.1540-6261.1992.tb04398.x
• Fama, E. F. and French, K. R., “The Value Premium”, Review of Asset Pricing Studies, vol. 11, no. 1, March 2021, pp. 105–121, sample July 1963–June 2019. academic.oup.com/raps/article/11/1/105/5961926
• Arnott, R. D., Harvey, C. R., Kalesnik, V. and Linnainmaa, J. T., “Reports of Value’s Death May Be Greatly Exaggerated”, Financial Analysts Journal, vol. 77, no. 1, 2021, pp. 44–67, published online 8 January 2021. Author PDF
• Piotroski, J. D., “Value Investing: The Use of Historical Financial Statement Information to Separate Winners from Losers”, Journal of Accounting Research, vol. 38 (supplement), 2000, pp. 1–41, sample 1976–1996. University of Chicago Selected Paper 84 reprint (PDF)
• Lakonishok, J., Shleifer, A. and Vishny, R. W., “Contrarian Investment, Extrapolation, and Risk”, Journal of Finance, vol. 49, no. 5, December 1994, pp. 1541–1578. doi.org/10.1111/j.1540-6261.1994.tb04772.x
• Annualised returns, drawdowns, rolling windows, the valuation spread and the industry backtest are our own calculations from the French data. The drawdown is measured from the previous peak of HML compounded monthly. Returns are gross of trading costs and taxes.

Related reading

• Technical vs fundamental analysis: what each one can tell you.
• Index funds vs active management: how often professional stock pickers beat the index.
• Position sizing and risk of ruin: sizing an idea that may take years to work.
• Behavioral finance: six biases with a measured cost: including the one that keeps value traps in portfolios.
• Stock market mistakes, ranked by what they cost: concentration and averaging down among them.
• The Sharpe ratio: comparing returns after adjusting for volatility.

This article is general information, not personalised investment advice. It does not take into account the financial situation, objectives or risk tolerance of any individual reader, and it does not recommend any specific security or sector. All figures describe specific historical periods and portfolios, are gross of costs and taxes, and do not predict future results. Capital is at risk and past performance does not indicate future results.