Ep. 380: Alex Edmans on Market Bubbles, Investor Biases…

PodcastMacro Hive ConversationsAlex EdmansOct 16

Edmans examines how investor biases can distort prices, from overreaction and herding to confirmation bias and anchoring. Examples ranging from Cisco's growth expectations and the 3Com–Palm spin-off to Amazon's cloud business show how markets can misprice both trends and information. He also discusses why professional investors can struggle to exploit these errors and how disciplined analysis may matter more than confidence or selectively observed returns. On AI, he warns that it could weaken critical-thinking development and distort peer review, while arguing that human skills such as negotiation and relationship-building may remain valuable.

Market Valuations Investor Sentiment Equities

Preview · 22 more points in the app

Read the Full Summary

Key Takeaways

Edmans warns that AI could favor conventional research over innovative work and that relying on it to write essays may hinder critical-thinking development. He recommends building skills such as critical thinking, presentation, negotiation, and relationship-building, while using AI to polish work rather than replace it.

Edmans argues that bubbles can form when investors correctly identify a trend but overestimate its pace: Cisco’s valuation implied sixfold growth in five years, though it took 25 years, while EV adoption could face infrastructure and battery constraints and AI progress could be slowed by regulation. He also says investors in early 2021 valued EV companies as if each would be the ultimate winner, despite the possibility that the industry could support only three survivors.

Edmans says market overreaction can create opportunities, as a strategy buying three-year losers and selling three-year winners found that the pattern later reversed. But he cautions that overvaluation can worsen before a reversal—as in Julian Robertson’s tech short—and that managers with flighty capital may be unable to wait it out.

Edmans links individual-investor underperformance to behavioral biases: anchoring to a purchase price can lead investors to sell winners and hold losers, while a study of 78,000 brokerage accounts found stocks bought in round-trip trades underperformed those sold by about 3% annually. He also warns that familiarity with a field, such as medicine, does not necessarily provide an investing edge without knowledge of costs, competitors, marketing, and regulation.

Edmans gives examples of prices diverging from underlying business value: Allbirds’ rebrand to Newbird AI was followed by a 582% stock-price rise before the stock returned to its pre-rebrand level, and the 3Com–Palm spin-off briefly valued Palm above 3Com as a whole. He says flashy changes that do not alter dividends, profits, or discount rates can still move prices, while the 3Com discrepancy meant investors could effectively buy Palm through 3Com for less than Palm’s standalone market value.

TECHNOLOGY

AI Peer Review — Edmans warns that AI-assisted peer review could favor conventional work and reject genuinely innovative research that AI cannot understand.

“But if we're just outsourcing it to AI, then actually we might just have value regression to the mean type of responses…”—

AI And Thinking — Edmans advises students not to rely on AI to write essays, because doing so could hinder their critical-thinking development. He suggests writing the essay themselves, then using AI to polish it if they wish.

“So, yes, you could use AI to write an essay, but then you're just not developing the critical thinking skills…”—

MARKETS

Professional Biases — Edmans says market mistakes are not limited to amateurs: professionals managing large sums can also be overconfident, overreact, or underreact. That matters to people who entrust managers with their pensions.

“So you might think, okay, it might only be amateurs who make mistakes in the market, but even professionals who are paid millions to invest billions, they can suffer from…”—

Markets As Laboratory — Edmans describes markets as a laboratory for studying decisions because they provide extensive data on investor and fund-manager performance. Evidence of overreaction and underreaction may suggest that people misread information in other fields, too.

“But the market is such a great laboratory because you have loads of data. You can see the investment performance of an investor or a fund manager.”—

Bubbles And Timelines — Edmans argues that bubbles can form around genuine trends when investors get the direction right but overestimate how quickly the trend will develop. Cisco's 190 price-earnings multiple implied sixfold growth within five years, but the company took 25 years to grow sixfold. EV adoption could face infrastructure and battery constraints, while regulation could slow AI progress.

“So maybe there's a tulip bubble a couple of centuries ago, the tech bubble in the late nineteen nineties. People suggest there might be an AI bubble right now…”—

MacroHive Markets Update

The host announces MacroHive Markets, a new regulated business offering institutional clients execution services alongside MacroHive strategy and quantitative insights, then introduces Alex Edmans.

Cisco Bubble And Timing

Edmans argues that bubbles often reflect a correct view of a trend but an exaggerated timeline, illustrating how Cisco’s growth took decades rather than the five years its valuation implied.

Contrarian Investing Limits

Past losers can outperform past winners as prices reverse, but the Tiger Management example shows that short-term losses and impatient outside investors can keep professionals from surviving until a bubble breaks.

World Cup Market Mood

Edmans describes research finding that national stock markets fall after World Cup elimination, arguing that personal emotions can influence professional judgments while wins produce no comparable measurable boost.

Allbirds Rebrand Reaction

The Allbirds-to-Newbird AI rebrand and dot-com-era name changes show how investors can overreact to salient labels, even when the underlying business has not fundamentally changed.

Herding And Selective Evidence

Herding can turn individual mistakes into shared market errors, especially when investors copy friends’ visible wins while overlooking losses, though relevant expertise can make shared information useful.

Amazon Cloud Underreaction

Confirmation bias delayed recognition of Amazon Web Services, while less-visible assets such as employee satisfaction and research spending can take years to be reflected in stock prices.

CEO Signals And Governance

Edmans explains how AGM locations, corporate-jet use, and vocal cues can reveal executives’ information or character, urging investors to assess people as well as financial figures.

3Com Palm Mispricing

The 3Com–Palm spin-off illustrates narrow framing and faulty valuation arithmetic, while Edmans cautions that investors should distinguish genuinely new product categories from inflated sector-wide expectations.

Anchoring And Trading Bias

Investors often anchor on purchase prices, sell winners too early, and hold losers too long; a useful corrective is to imagine starting with cash and reconsider each holding from scratch.

Overconfidence And Investor Edge

Brokerage-account evidence suggests active retail trades often underperform, so Edmans advises investors to identify a real competitive edge and recognize that patience may matter more than stock-picking confidence.

AI, Learning, And Research

Edmans warns that AI may overwhelm peer review and weaken students’ critical thinking when used as a shortcut, while emphasizing human skills such as presenting, negotiation, and relationship-building.