Trading and Exchanges Chapter 17: How Arbitrageurs Keep Markets Honest (Part 1)
Chapter 17 is about arbitrageurs, and it is one of those chapters that changes how you think about markets. Arbitrageurs are the people who keep prices consistent across different markets and different instruments. Without them, you could have oil priced at 80 dollars in New York and 70 dollars in London, and nobody would fix it.
What Exactly Is Arbitrage?
Arbitrageurs trade instruments whose prices are correlated. They buy the one that looks cheap and sell the one that looks expensive. They profit when prices converge back to their normal relationship.
Harris introduces useful terminology. The portfolio an arbitrageur builds is called a hedge portfolio. Its positions are legs. The price difference between instruments is the basis. The arbitrage spread is the gap between the actual basis and the fair value. When that gap gets big enough, arbitrageurs step in.
Here is the clever framing: think of the hedge portfolio as its own instrument. When arbitrageurs “buy” it, they are buying its long positions and selling its short positions simultaneously. Arbitrage is just regular trading in a synthetic instrument.
Pure Arbitrage vs Speculative (Risk) Arbitrage
This is the big distinction in the chapter.
Pure arbitrages involve instruments where the basis is mean-reverting. It always tends to come back to some average value. Like rolling dice: after you roll a 12, the next roll will probably be lower. After a 2, probably higher. Average is always 7. Pure arbitrages are safer because some mechanism guarantees prices eventually converge.
Harris breaks pure arbitrages into three types:
Shipping arbitrages are the simplest. Same instrument, two markets, different prices. Buy where it is cheap, sell where it is expensive. Ship if needed. Crude oil between New York and London is the classic example. Oil shippers literally reroute tankers from the Persian Gulf based on which port pays more. The arbitrage bounds are set by shipping costs.
Delivery arbitrages involve futures contracts and their underlying cash instruments. The delivery mechanism forces convergence at expiration. If wheat futures are expensive relative to cash wheat, you buy cash wheat and sell futures. If the basis closes before delivery, you unwind and profit. If not, you deliver the wheat. Either way you win (minus carrying costs).
Conversion arbitrages involve converting risk from one form to another. Options market makers do this constantly. They sell option contracts and hedge with the underlying stock, essentially manufacturing options. The soybean crush is a beautiful example. Millers crush soybeans into soy oil and meal. Arbitrageurs on the Chicago Board of Trade do the same thing financially: buy soybean futures, sell oil and meal futures. Unlike physical millers, they can “reconstitute” beans by reversing the trade. Try that with actual soybeans.
Speculative arbitrages (also called risk arbitrages) are the opposite. The basis is nonstationary, meaning it wanders around without returning to any particular value. Think of a random walk. Arbitrageurs here are betting that short-term mean reversion will dominate before the long-term random drift causes damage.
Pairs Trading
Pairs trading is probably the most relatable speculative arbitrage. You find two stocks that normally move together, notice one has risen while the other has not, and bet on convergence.
Harris gives two examples that perfectly illustrate success and failure.
Success: Bernie notices Washington Mutual rises on interest rate news, but FirstFed Financial, a similar savings and loan, does not move. He buys FirstFed and shorts Washington Mutual. Next day, buyers show up for FirstFed. He closes both positions for 2.8% net. With leverage, his return on actual capital was 5.2%.
Failure: Gerry sees Ford rising relative to GM with no obvious reason. He buys GM and shorts Ford. Then Ford announces the Excursion SUV is stealing sales from GM’s Suburban. The spread never closes. Gerry eats a loss.
The lesson: if the price divergence is because of uninformed trading noise, the pairs trade wins. If it is because of real fundamental news, you lose. And you often cannot tell the difference in real time.
Statistical Arbitrage and Risk Arbitrage
Harris extends pairs trading into two more sophisticated strategies.
Statistical arbitrageurs use factor models to scale pairs trading across many instruments at once. Instead of tracking two stocks, they model returns as weighted sums of common factors (interest rates, inflation, market levels) plus stock-specific noise. When a price is inconsistent with the model, they trade. Heavy quant work.
Risk arbitrage in the narrow sense means trading around mergers. When company A announces it will buy company B, the target’s stock should converge to the offer price. Risk arbitrageurs buy the target and short the acquirer. Looks like arbitrage, but it is really a bet that the deal closes. If it falls apart, you get destroyed. Harris is clear: more speculation than arbitrage, despite the name.
Why This Matters
Arbitrageurs are not just making money for themselves. They enforce the law of one price across markets, supply liquidity, and move it from where there is excess to where it is needed. They even manufacture financial products by converting risk from one form to another. When regulators restrict arbitrage (which has happened after crashes), the costs hit everyone through wider spreads and less efficient prices.
Part 2 will cover the specific risks arbitrageurs face, how they manage those risks, and the important relationship between arbitrageurs and dealers.
Previous: Chapter 16: Value Traders