How Central Banks Communicate

A central bank decision arrives as three separate documents released over about an hour, and the market trades a different part of each. Knowing which part is which is most of what separates a trader watching a decision from a trader guessing at one.

The statement and what changed in it

The statement is a short document — often under a page — published at the moment of the decision. It is written with extraordinary care, revised word by word, and deliberately similar to the previous one.

That similarity is the point. The market reads the statement as a diff. A phrase that survived unchanged says policy is unchanged; a phrase that was softened, dropped or added is the signal, and the removal of a single word can move a currency more than the rate decision it accompanies.

Reading one in practice

1. Open the previous statement beside the new one. Most central banks publish an archive, and several news outlets publish a marked-up comparison within minutes.
2. Read the changes first, then the whole thing. The unchanged parts carry no new information by construction.
3. Find the vote. A committee split 5–4 has far less conviction than one that was unanimous, and a split can precede a turn.
4. Note anything about the balance sheet. Quantitative tightening or easing is policy too, and it is often relegated to a paragraph nobody reads.

Projections and dot plots

Several times a year, some central banks publish forecasts alongside the decision: growth, unemployment, inflation and — most importantly — where the committee expects rates to be.

The Federal Reserve's version is the dot plot: each participant marks where they think the policy rate should be at the end of each of the next few years, anonymously, one dot each.

Two warnings that get forgotten every cycle. The dots are not a commitment — they are individual views at one moment, and they have been badly wrong. And they are not a forecast of the market's path, which is priced separately and often differs substantially; when they diverge, that gap is itself the story.

The rest of the projection summary

Growth, unemployment and inflation forecasts matter mainly for their internal consistency. A bank forecasting inflation back at target without raising rates is telling you something about how it sees the economy; a bank forecasting a rise in unemployment while staying hawkish is telling you it has decided to accept that cost.

The press conference

Usually thirty to sixty minutes after the statement, the governor or chair reads a prepared opening and then takes questions. This is routinely where the day's real move happens.

• The prepared remarks are as carefully written as the statement and rarely surprise.
• The questions are not. Journalists ask what the statement avoided, and the answers are unscripted.
• A single phrase can reverse the entire post-decision move, which is why a currency can jump on the decision and be trading the other way an hour later.

The exchanges that move markets are the ones about what would change the bank's mind: what would make you cut?, is a pause the same as being done?, how much weight are you giving to one bad month of data? Those answers describe the reaction function — the rule connecting future data to future policy — which is worth far more than any single decision.

Decision versus guidance

The distinction the whole lesson has been circling: the decision is what they did; the guidance is what they signalled about next time. The second is nearly always worth more.

Rows 2 and 4 look paradoxical and are the normal case. A currency falling on a rate rise is not the market being irrational; it is the market repricing a path that just got shorter, exactly as lesson 1 described.

A workable routine

1. Before: know what is priced. If a hike is fully priced, a hike is not news.
2. At the decision: compare with what was priced, not with last month's rate.
3. In the statement: read the changes.
4. At the conference: listen for the reaction function, not the tone.
5. After: wait. The first move is frequently reversed, and the settled direction an hour or two later is the more informative one.

Step 5 costs you the initial spike and saves you from being on the wrong side of the reversal. For anyone not trading these events as a specialty, that is a good trade.