In Praise of Dumb Money: Fischer Black's 'Noise'

In 1986, Fischer Black stood up to give his presidential address to the American Finance Association. This is normally an occasion for a dignified survey of one's field. Black, who co-invented the Black-Scholes option pricing formula and was as rigorous as anyone alive, instead delivered a short, strange, almost conversational essay with hardly any equations, arguing that markets run on people trading for no good reason, that prices are typically wrong by a lot, and that this is fine.

It is the most quotable paper in finance, and almost nobody's model has fully caught up with it forty years later.

The problem: why does anyone trade at all?

Start with a puzzle that should keep you up at night.

Suppose every trader in the market is rational and trades only on information. You want to buy 10,000 shares from me. I am rational. I ask myself: why does this person want to buy from me at this price? The only answer is that they know something I don't. In which case I should not sell.

And you, being rational, ask the same question about me. Neither of us should trade. Volume should be roughly zero. This is not a joke, it is a real theoretical result, and it is deeply awkward, because the actual market trades billions of shares a day.

So the puzzle is: what are all these people doing?

The key idea via analogy: static on the radio

Black's answer is one word. Noise.

He defines noise as the opposite of information. It is everything that gets into prices and volume that isn't real news: hunches, chart patterns that mean nothing, tips from a brother-in-law, a hot newsletter, boredom, a need to feel busy, a story that sounds compelling and is false.

The crucial part, the thing that separates Black from a scold complaining about retail investors, is this: noise traders are not a bug. They are the power supply.

Think of a radio. Static is annoying and it corrupts the signal. You would rather not have it. But now consider a market where there is no static at all: every single participant is perfectly informed and perfectly rational. In that market, as we just saw, nobody trades. There is no liquidity, no volume, no prices to observe. The market does not exist.

Noise traders are people who think they are trading on information but are actually trading on noise. And crucially, they must think this, or they wouldn't show up. Black's line is that noise trading is what makes markets possible, and he means it structurally, not sarcastically. Here is the chain:

• Noise traders show up in size, convinced they have an edge, and are willing to trade.
• Their presence gives the genuinely informed trader someone to trade against, and a crowd to hide in. If you have real information you can now put it on without instantly revealing yourself, because the person on the other side might just be a noise trader, and the market maker cannot tell which you are.
• That is the only reason information gets into prices at all. Noise is the fog that lets information move. (Compare Grossman and Stiglitz in 1980, who needed exactly this fog for their model to work. Black is telling the same story in English instead of algebra.)
• Noise traders collectively lose money to the informed. That is their fee for the entertainment, and it is the informed trader's paycheck.

Then Black delivers the line that ruined a lot of people's afternoons. He suggests we call a market efficient if price is within a factor of two of value, and guesses that markets are efficient by that standard most of the time.

Read that again. He is not saying prices are approximately right. He is saying a stock worth 50 dollars could reasonably trade at 25 or at 100, and he would still call that market efficient. The high priest of quantitative finance is telling you that the price is usually in the right postcode, not the right house.

And then he closes the trap. Noise makes prices wrong. But noise also makes it impossible to know how wrong, because you can never cleanly separate signal from static. So the very thing that creates the mispricing is the thing that prevents you from confidently exploiting it. Noise "both causes markets to be inefficient and prevents us from taking advantage of the inefficiencies."

Why it mattered

• It legitimized noise traders as a modeling ingredient. Four years later, De Long, Shleifer, Summers and Waldmann built the formal model showing that noise traders don't just add liquidity, they create a risk that scares away arbitrageurs and lets mispricing persist. Black is the essay, that is the equation.
• It reframed liquidity. After Black, the market maker's spread is understandable as a wager: you profit from noise traders and lose to informed ones, and you set your spread wide enough that the first exceeds the second. Glosten and Milgrom, and Kyle, formalized this. If you have ever thought about adverse selection, you are thinking Black's thoughts.
• It gave everyone permission to be humble about "fair value". If prices are routinely off by tens of percent and you cannot tell when, then your beautiful discounted cash flow model with its 0.5 percent precision is theatre. Position sizing, drawdown tolerance and survival matter more than being right.
• It is honest about how hard alpha is. Black is not saying "markets are irrational, go get rich". He is saying "markets are irrational, and good luck telling which of your ideas is information and which is noise, because from the inside they feel identical." That is a far more useful warning than any efficient markets lecture.

The honest limitations

• It is an essay, not a model. There is barely a formula in it. Black asserts, gestures, and provokes. The paper's influence comes from being right and vivid, not from being derived. A referee at a modern journal would reject it on sight, which says something uncomfortable about modern journals.
• The "factor of two" is a guess. Black says so himself. It is not estimated from anything. It has become one of the most cited numbers in finance despite being, by the author's own admission, an intuition.
• Noise is defined negatively and is unobservable. Noise is "whatever isn't information". You cannot measure it directly, you can only infer it from residuals, and any inference depends on your model of what information is. This is a real methodological hole, and it is why the noise trader literature has always had to proxy for sentiment with awkward stand-ins like closed-end fund discounts.
• It does not tell you how to make money. In fact it goes out of its way to tell you that you probably can't, reliably. If you came to behavioral finance hoping the mispricings would be easy to harvest, Black is the person waiting at the door to explain that the same fog hiding the treasure is hiding the rocks.

The one-line takeaway

Fischer Black argued that markets are powered by people trading on noise while sincerely believing it is information, that this is what gives markets liquidity and lets real information get into prices, and that as a consequence prices are usually wrong by a wide margin and you will rarely be able to prove which way.