Representativeness heuristic

The representativeness heuristic is the tendency to judge the probability that something belongs to a category based on how similar it is to your image of a typical member of that category. If a company fits your mental template of a “growth stock,” you judge it as more likely to grow, even if the base rate of growth stocks that actually deliver is low. You neglect the relevant statistical baseline in favor of the similarity judgment.

The classic example

Tversky and Kahneman asked subjects: “Linda is 31 years old, single, outspoken, and concerned with issues of discrimination and social justice. Is Linda more likely to be (a) a bank teller or (b) a bank teller and active in the feminist movement?”

Most subjects chose (b), even though (b) is a subset of (a) and mathematically cannot be more likely. Linda’s description is representative of a feminist activist, so subjects judged that category as more likely, ignoring that any category that includes another category must be at least as probable.

In investing, the equivalent might be: “TechCo has brilliant engineers, a charismatic founder, a huge market opportunity, and venture funding. Is it more likely to become (a) a successful company or (b) a successful company and achieve a billion-dollar valuation?” Investors often choose (b) because TechCo is representative of a “unicorn” stereotype, even though (b) is a smaller set and less likely.

How representativeness distorts judgment

Representativeness operates by similarity matching. You form a mental image of what a “typical growth stock” or “typical failing company” looks like, and then you judge new cases by how well they match that image. The stronger the match, the higher the probability you assign — regardless of the actual base rate.

If a company looks like it fits the template of a successful biotech (promising science, experienced team, large addressable market), investors assign a high probability of success. But the base rate for biotech companies that go public — the actual fraction that deliver substantial returns — is much lower than the similarity-based probability would suggest.

The problem is that similarity is not the same as probability. A company can be very representative of the “successful startup” category and still be unlikely to succeed, simply because most startups fail.

Representativeness and price-to-earnings ratios

Representativeness helps explain why growth stocks often trade at high valuations. A company with high revenue growth, market leadership, and a charismatic CEO looks like the “template” of a great investment, so investors pay a premium multiple. But the base rate of companies that deliver returns commensurate with those multiples is low. Many high-growth stocks turn out to be overvalued.

Conversely, a cheap, slow-growing, dividend-paying industrial company may not look like the “template” of a good investment, so it trades at a low multiple — even though cheap, overlooked companies historically deliver better returns than expensive, popular ones.

Representativeness and the conjunction fallacy

Representativeness directly causes the conjunction fallacy: the belief that a conjunction (two events happening together) is more likely than one of the individual events. This is mathematically impossible but psychologically very common when the conjunction matches a stereotype better than the individual event does.

In investing: “Fund Manager X has beaten the market for 10 years. Which is more likely: (a) she will beat the market next year or (b) she will beat the market next year and the market will be a bull market?” Many investors choose (b) because the conjunction fits the “she succeeds in strong markets” stereotype, even though (b) is a subset of (a).

Representativeness and sector-level bias

At the sector level, representativeness can cause entire groups of investors to systematize their mistakes. During the dot-com bubble, any company with “.com” in its name seemed representative of the “internet success” category, so stocks were valued as if success were certain. Representativeness caused the market to mis-price an entire sector.

Similarly, after a sector outperforms (tech, crypto, energy), it becomes representative of “hot sector,” and money floods in. When it underperforms, it becomes representative of “dead sector,” and money floods out — often at the worst times.

Distinguishing representativeness from base-rate neglect

Representativeness is about similarity to a stereotype. Base-rate neglect is about ignoring the actual statistical frequency of a category. They often occur together: you judge something as representative of “successful company” (similarity judgment) and therefore assign it a high success probability, ignoring that the actual base rate of successful companies is low.

Representativeness is also related to anchoring bias in that both distort judgment, but anchoring is about dependence on a starting number, while representativeness is about similarity to a mental template.

Defenses against representativeness bias

• Learn base rates. Before investing in a category (growth stocks, startups, emerging markets), find out: what fraction of companies in this category actually succeed? What is their median return? Anchor your judgment to that base rate, not to the similarity of individual cases.
• Separate “good story” from “good investment.” A company can have a wonderful, representative-of-success story (great product, talented team, big market) and still be a poor investment if it is already priced for perfection.
• Use a decision framework based on valuation, not on narrative fit. Does the price-to-earnings ratio match historical returns for similar companies? That question is much less susceptible to representativeness than “does this story sound like a winner?”
• Diversify within categories. Rather than choosing the most representative growth stock, hold a basket of growth stocks at varying valuations. This forces you to confront actual base rates, not stereotypes.

See also