The market reacts to the gap from 'expectations'

STEP 1Consensus and surprise

In Lesson 19 we saw a day when jobs grew by nearly three times the forecast and yet the stock market fell. Who decides that 'forecast', and how? Let's start there.

The forecast is a collection of experts' numbers

Before major indicators such as CPI or the jobs report come out, financial information firms like Bloomberg, Reuters and Dow Jones ask economists at banks, brokerages and research institutes, "What do you expect this month's number to be?" When the dozens of answers are gathered and the middle value (median) or average is taken, that is the consensus, the market forecast. The box labeled 'forecast' in an economic calendar is exactly this number.

Consensus is not a number meant to be right; it is a number that shows 'what the market currently believes'. Investors have already bought and sold stocks and bonds based on this forecast. So what the market checks at the moment of release is not "is the number high or low?" but "how different is it from what we believed?"

Is 'above forecast' good or bad?

Here is something easy to mix up. 'Above forecast' only means the number is bigger than forecast, not that it is good. Inflation above forecast is bad news (worry about rate hikes), and jobs above forecast are good for the economy but raise rate worries when inflation is high. So for each indicator you have to think separately about 'which direction of surprise is a burden for the market'.

STEP 2Why only the gap matters

In STEP 1 we calculated the surprise. Now let's build up 'why' the market reacts only to the gap rather than the number itself, and why the reaction differs even for the same gap.

The principle of Lesson 14 applies directly

In Lesson 14 you learned that market interest rates price in the future policy rate ahead of time. For example, if the expectation that CPI will be in the 4% range is widespread, the 2-year Treasury yield already contains the expectation that 'the Fed will raise rates further'. If the actual number matches the forecast, that is merely confirmation that the expectation was right. If it is higher than forecast, the expectation must be revised to 'they'll raise more', and if lower, to 'they'll raise less', and rates and stock prices move by the size of that revision. In the words of Lesson 1, a surprise is an event that instantly resets the multiple knob.

How different must it be to count as a 'surprise'?

The size of the gap that matters differs by indicator. For US CPI, forecasts cluster very tightly, so a gap of just 0.1–0.2 percentage points can shake the market for a whole day. Job gains, by contrast, are inherently hard to predict and expert forecasts are widely scattered, so a difference of tens of thousands is common and it takes a miss of over 100,000 to count as a big surprise. How widely the forecasts are scattered, that is, 'how unsure even the experts were', decides the weight of a surprise.

The details and 'today's worry' change the reaction

It's easy to be wrong if you judge the reaction from one headline alone. With CPI, even if the headline is lower than forecast, the market can be disappointed if core inflation is higher than forecast. With jobs, even if the number of jobs added beat forecast, the reaction can go the other way if the previous two months were revised sharply down or wages rose less than forecast. And as you saw in Lesson 19, the same direction of surprise is read differently depending on whether the market is worried about inflation or about the economy.

See it in a chart: the surprise gauge

Pick an indicator and move the forecast and actual. For jobs, also change 'what the market is worried about now'. You can see the reaction change even though the number is the same.

The first screen is today's question. The actual 3.0% is higher than the 2% target but 0.2 percentage points below the 3.2% forecast, so the gauge points to 'below forecast' and market rates tend to fall. Now pick jobs, raise the actual number well above the forecast, and switch 'current worry' back and forth between inflation and the economy.

STEP 3Five real release-day scenes

Now that you know the principle, let's go to real release days. In all five scenes it was not the 'number' but the 'gap from expectations' that moved the market.

From forecast to actual: the direction of the arrow

The gray dot on each line is the forecast and the tip of the arrow is the actual release. Tap a line to see how the market reacted that day.

November 10, 2022 is the day that best shows the power of a surprise. US October CPI was 7.7%, still nearly four times the 2% target. But just because it was 0.2 percentage points lower than the 7.9% forecast, 2-year and 10-year yields fell about 0.15 percentage points early in the session, and the S&P 500 rose 5.5% in one day, its biggest one-day gain in about two years. A number that was 'high but less high than forecast' created the expectation that the pace of rate hikes would slow.

The jobs cases reconfirm the lesson of Lesson 19. In February 2023 and June 2026, when inflation was high, stronger-than-forecast jobs pushed up rates and weighed on stock prices, while in August 2024, when inflation had come down, weaker-than-forecast jobs stirred recession worries. On October 2, 2026, September jobs came in at just 29,000, far short of the 90,000 forecast, and the rate-hike expectations for October that had been building faded, pushing the 10-year yield down about 0.07 percentage points on the day.

Company earnings work on the same principle

This principle is not limited to economic indicators. When a company reports earnings, there is also a consensus gathered from brokerage analysts' forecasts, which in Korea is compiled by information firms such as FnGuide. When operating profit far exceeds consensus it is called an earnings surprise, and when it falls far short it is called an earnings shock. Even if profit grew 50% from last year, if the market expected an 80% increase it becomes an earnings shock and the stock price can fall. In the end, the earnings knob from Lesson 1 also moves 'relative to expectations'.

The order for reading release-day numbers

1. Write down the forecast before the release. Check the date and the forecast in the economic indicator calendar.
2. Look at the direction and size of the headline surprise. Its weight differs: 0.1 percentage points for CPI, tens of thousands for jobs.
3. Look at the details. Check whether core inflation, revisions to the previous month and wages point the same way as the headline.
4. Match it against what the market is worried about now. The same number is read differently when inflation is the worry and when the economy is the worry.

For stocks, it goes like this

Most surprises first move interest rates and the dollar, and then spread to stock prices through the multiple knob of Lesson 1. Growth stocks and long-term bonds, which rely on far-future earnings, are the most sensitive, and sectors such as banks, whose profits can rise when rates rise, sometimes move the opposite way. To see which sectors each indicator affects, tap events such as 'CPI above forecast' or 'jobs surprise' in the causal chain of the Macro Map.

For Korean investors, most US indicators come out at night Korean time. So the Korean stock market moves the next morning on the result of the US market having already absorbed the surprise. Making it a habit to first check "what differed from expectations last night" in the morning is the shortcut to understanding that day's market.

It doesn't always move that way. On many days the first reaction right after the release reverses within hours, because the market changes its interpretation after reading all the details. Please read a surprise as a clue to the 'first direction'.