1. Surprise versus level: the price moves on the difference

Level 4: What really moves the price

The Macro course dropped an idea in passing and moved on: the market does not react to the number, it reacts to the difference between the number and what was already priced in. There it applied to a CPI print. Here it applies to one specific company, and it stings rather more, because results arrive wrapped in a headline that appears to pass judgment all by itself: beats forecasts, disappoints, record high. That headline is, almost always, irrelevant to the price.

The level is already paid for

When you buy a share at 40 euros you are not buying the company's profits. You are buying the difference between what the company will do and what that price already assumes it will do. The level of results — how much it sells, how much it earns, how fast it grows — is built into the 40. You have already paid for it. The only thing that can move the price for you from tomorrow onwards is the surprise: the part of reality that did not fit inside that assumption.

That is where the uncomfortable rule of this level comes from. An excellent company and a mediocre one can deliver exactly the same return, and a company on the brink of insolvency can hand you more than the best name in the index. Not because the market is stupid, but because excellence was demanded of the excellent one and disaster was demanded of the other. What you get paid for is not quality: it is the distance between reality and expectation.

Beating forecasts is the expected news

Two things get confused daily, and they are worth pulling apart:

• The analysts' forecast is a published number, soft and fairly easy to manage. Companies guide cautiously and beat by a whisker, quarter after quarter.
• The expectation built into the price is published by nobody. It is the full path of growth, margin, investment and duration that makes 40 euros be 40 euros and not 25.

Clearing the first by two cents says nothing about the second. That is why, in a normal quarter, the large majority of the companies in an index like the S&P 500 beat the earnings-per-share forecast — the usual proportion is around three quarters — and yet nothing like that same proportion of shares rises on the day of the announcement. If beating were good news, almost all of them would rise. Beating is the floor, not the ceiling.

What gets revised is not the quarter, it is the path

Here is the heart of the matter, and it shows up best in round numbers. Picture a company selling 1,000 million a year whose price builds in growth of 20% a year for eight years. At that rate, eight years out it would be selling around 4,300 million. That is what you are paying for.

The quarter arrives. Sales come in 2% above forecast: on roughly 250 million for the quarter, that is 5 million of surprise. Green headline. But on the same call management explains that, with the core market saturating, sustainable growth looks more like 12%. At 12% a year, eight years out it would be selling around 2,480 million, not 4,300.

Compare the two figures. The good news is worth 5 million and happens once. The revision knocks around 1,800 million off annual sales at the end of the period, and every year in between sinks with it. There is nothing contradictory about the share falling 20% on the day it beats forecasts: what has been revised is not the quarter, it is the slope of the line running out of it.

The same mechanism works in reverse. A company the price has written off, with sales shrinking 5% a year, is on 1,000 million today and around 663 million eight years out on that path. If it reports a weak quarter but stops shrinking — zero growth, not a euro more — the path becomes 1,000 million eight years out. It has disappointed on the level and handed over more than 300 million of future annual sales on the path. It rises.

How to tell them apart in practice

Three questions, in order, that work for reading any odd reaction:

• Does the news affect one period or every period after it? A bad quarter caused by a fire at a factory affects one. A price cut by a competitor affects all of them.
• Has the company changed its own forward guidance, or has it only reported the past? Guidance is the path; the past was priced in weeks ago.
• What moved: the magnitude or its rate of change? What the price revises is almost always the rate, not the magnitude.

There is also a phenomenon documented for decades in the academic literature on reactions to earnings announcements: after a large surprise, the price does not finish adjusting on the day, but keeps drifting in the same direction for weeks. Put another way, the market takes the surprise on board quickly but not all at once. It is a useful reminder that an expectation is not a switch, but something revised in stages.

A real case: the same company, two opposite reactions

Meta's results in February 2022 are the textbook example. The company had closed a year with revenue clearly up, but it guided revenue for the following quarter below expectations and acknowledged, for the first time, a fall in daily active users. The share dropped around 26% in a single session, the largest single-day destruction of value recorded up to then at a US company. The level was good; the path, worse.

A year later, in February 2023, the same company reported annual profit far below the previous year, down by a high double-digit percentage. Objectively, a bad year. But it announced cost discipline, a cut in investment and a share buyback, and the stock rose sharply, on the order of 20% in the following session. It disappointed on the level and surprised on the path.

The exact figures are best taken as orders of magnitude rather than as exam answers. What is not approximate is the lesson: you get paid for the revision, not for the result.