Limits to Arbitrage: Why Mispricing Can Last

Why smart investors cannot always correct wrong prices, including noise trader risk, costs and the danger that markets stay irrational longer than you can stay solvent.

If a share is overpriced, why don’t smart investors sell it until the price falls to fair value? In theory, arbitrage should quickly correct mispricing. In practice, there are important limits to arbitrage.

Noise trader risk

In 1990, economists Bradford De Long, Andrei Shleifer, Lawrence Summers and Robert Waldmann described noise traders: investors who trade on emotion, rumours or trends. Their behaviour can push prices further from fair value before they return. An arbitrageur betting on correction can lose money in the meantime.

Short-term pressure

In 1997, Andrei Shleifer and Robert Vishny explained that professional investors manage other people’s money. If their bets lose money in the short run, clients may withdraw funds, forcing them to close positions at a loss, just when the opportunity is greatest.

A saying often attributed to Keynes captures this: markets can stay irrational longer than you can stay solvent.

Costs and constraints

Short selling can be costly or restricted.
Transaction costs and borrowing costs.
Fundamental risk: the investor might be wrong about the true value.
No perfect substitutes: it may be impossible to hedge a bet exactly.

Famous examples

Twin shares: Royal Dutch and Shell were two shares with claims on the same company’s profits in a fixed ratio, yet their prices diverged for years from that ratio.
The dot-com bubble: some fund managers who bet against overpriced tech stocks in the late 1990s lost clients before the crash proved them right.
LTCM, a large hedge fund, collapsed in 1998 when its arbitrage bets moved against it and it could not hold on.

Why it matters

Limits to arbitrage explain why bubbles can grow, and why “the market is wrong” doesn’t mean you can easily profit from it.

The early sceptic

In 1998, a fund manager decides internet stocks are overvalued and avoids them. Her fund lags the market for two years as tech soars, and clients leave. By the time the bubble bursts in 2000, she manages far less money. She was right, but too early.

Thinking mispricing is always quickly corrected

Costs, risks and short-term pressures mean mispricing can persist for a long time.

Key takeaways

Arbitrage should correct mispricing, but it has limits.
Noise traders can push prices further from fair value.
Short-term pressure can force professionals out of correct bets.
Twin shares, the dot-com bubble and LTCM illustrate these limits.