Reading Economic Data: NFP, CPI, and Rate Decisions

At 8:29 a.m. New York time on the first Friday of most months, the world’s largest financial market goes strangely quiet. Quotes widen. Order books thin out. Traders who have spent the morning arguing about the number stop arguing and watch the clock. Then, at exactly 8:30, the US Bureau of Labor Statistics publishes a single estimate — how many jobs the American economy added last month — and within microseconds, before any human has read a single word, currencies are already moving.

That monthly ritual is the clearest window into how economic data and foreign exchange (FX) actually interact. Currencies respond to a country’s economic health, but not in the vague way that phrase suggests. They respond to specific, scheduled, forecastable releases — and they respond less to the numbers themselves than to how the numbers compare with what everyone expected. This guide walks through the calendar that organizes those releases, the three families of data that matter most — employment, inflation, and rate decisions — and the two-step chain that turns a statistical surprise into a currency move.

The Calendar Is the Market’s Shared Diary

Most market-moving news arrives unannounced — a headline, a crisis, a surprise policy shift. Economic data is the exception. Statistical agencies and central banks publish their release schedules months or even a year in advance, to the minute. The US jobs report lands at 8:30 a.m. Eastern, usually — though not always — on the first Friday of the month; CPI arrives at 8:30 a.m. around the middle of the month; the Federal Reserve announces its rate decisions at 2:00 p.m. on eight pre-announced days a year. (The exact dates live on the agencies’ official calendars, which is where professionals check them — the “first Friday” rule is a habit, not a law, and a few months each year break it.) None of this is secret. Everyone in the market is looking at the same diary.

An economic calendar is simply that diary in usable form. A typical row shows the release time, the country or currency it affects, an importance rating, and three numbers: previous (the last reading), consensus (what forecasters expect this time), and — once published — actual (what came out). The consensus is usually the median of surveyed economists’ forecasts, compiled by the major news and data providers.

Reading a calendar well means internalizing two habits. First, filter aggressively: dozens of releases come out every week, and most of them barely move anything. The importance rating exists because a handful of releases — payrolls, inflation, rate decisions, and a few others such as GDP and the major business surveys — do most of the work. Second, always read the actual against the consensus, never in isolation. A report showing 150,000 new jobs is neither good news nor bad news until you know whether the market expected 100,000 or 250,000.

Everything in this article, incidentally, is designed to be used with a calendar open in front of you. The Currency Stack maintains a live one at thecurrencystack.com/economic-calendar — this guide is, in effect, its user manual.

That second habit is the foundation for everything that follows, so it is worth pausing on.

Consensus: The Number Behind the Number

Markets are forward-looking: today’s exchange rate already reflects what participants collectively expect the economy to do. We covered this idea — expectations beat reality — in What Moves Currency Prices?, and data releases are where it shows up most vividly. If the consensus says the economy added 150,000 jobs and the report says 150,000, almost nothing new has been learned, and the market’s reaction is usually modest. The information content of a release is the surprise: the gap between actual and consensus.

This explains a pattern that baffles newcomers every single month. A “strong” report comes out and the currency falls; a “weak” one comes out and the currency barely blinks. In the first case, the market had positioned for something even stronger — the whisper around the desks ran hotter than the published consensus — so a merely good number was, in market terms, a disappointment. In the second, the weakness was already priced in. The direction of the surprise, not the direction of the number, is what carries the punch.

One practical note: the consensus is a private survey product, not an official statistic — there is no government-published forecast to check the number against. The figures on your calendar come from polls of economists run by the major data providers. Public benchmarks do exist for the curious reader without a terminal: the Philadelphia Fed has run a quarterly Survey of Professional Forecasters since 1968, and the Cleveland Fed publishes a daily inflation nowcast that has a respectable track record against the professionals.

Magnitude matters too. Forecasters are usually clustered fairly tightly, so a release only slightly away from consensus is routine noise. The violent reactions come when the actual lands far outside the range of forecasts — the genuine shocks. And one refinement worth knowing: the market often reacts to the composition of a report, not just its headline. A jobs report with a solid headline but soft wage growth can end up trading as a weak report once the details are digested — which is one reason the first move after a release is not always the lasting one.

Nonfarm Payrolls: The Report That Owns Friday Morning

If the calendar has a main character, it is nonfarm payrolls (NFP) — the headline figure of the US Employment Situation report, published monthly by the Bureau of Labor Statistics (BLS) at 8:30 a.m. Eastern, usually on the first Friday. It estimates the net change in US jobs over the prior month, excluding farm work and a few other categories such as the self-employed. Because the United States sits on one side of the overwhelming majority of currency trades, a scheduled reading on the health of the US economy is, in practice, a scheduled event for nearly every currency pair at once.

The report is really two surveys stapled together. The establishment survey asks a large sample of businesses and government agencies — around 119,000 employers covering some 622,000 worksites — how many people were on their payrolls in the period containing the 12th of the month. That is where the headline jobs number, hours, and average hourly earnings come from. The household survey asks about 60,000 households about their own employment, and produces the unemployment rate and the participation rate. The two surveys count different things — the establishment survey counts jobs (a person with two jobs appears twice), the household survey counts people, including the self-employed the other survey misses — so they can, and regularly do, tell slightly different stories in any given month.

Traders read the report in layers. The headline payrolls number gets the first reaction. The unemployment rate and average hourly earnings get the second look — earnings especially, because wage growth feeds directly into the inflation outlook, which feeds into interest rate expectations. And the revisions get the third look: each month’s report also restates the two previous months, and those restatements are sometimes large enough to change the picture on their own.

Now for the discipline that separates professional readers from headline readers: respect the error bars. The BLS itself publishes them. The 90% confidence interval on the monthly payrolls change is roughly ±122,000 — meaning a reported gain of 50,000 is statistically consistent with anything from a loss of about 70,000 to a gain of about 170,000. A one-tenth move in the unemployment rate is likewise well inside sampling noise. And the industry detail is noisier still, with confidence bands often wider than the changes being reported. The market will still react to a 30,000-job miss as if it were real information — that is how the game works — but you should know that much of what gets traded at 8:30:01 is, statistically speaking, static.

Revisions deserve the same respect. Each monthly figure is preliminary when published, revised twice as more survey responses arrive, and only then final; once a year, the whole level is reset against near-complete employment records from unemployment-insurance filings — the benchmark revision, which lands as a preliminary estimate each late August. Those resets can be humbling: the March 2025 benchmark ultimately cut the reported level of US employment by 898,000 jobs — first signaled in September 2025, when the preliminary estimate of roughly 911,000 became the largest such adjustment on record. The lesson is not that the numbers are useless; it is that any single print is a first draft, and the trend across several months is far more trustworthy than any one Friday’s headline.

A live illustration of reading in layers, from July 2026: the headline showed a loss of 23,000 jobs — superficially recessionary. The layers said otherwise. Fifty thousand of the weakness came from one administrative category (local government education); the private sector actually added 30,000; the whole print sat comfortably inside that ±122,000 band; and hours and wage growth were steady — not the pattern of a genuine collapse in labor demand. The qualifier cut the other way too: the two prior months were revised down by a combined 103,000, so the trend was weaker than previously believed even as the single month was noisier than it looked. Headline readers saw a disaster; layer readers saw a soft trend and a noisy print. That distinction is this article’s whole argument in one report.

CPI: The Number Central Banks Are Paid to Control

The Consumer Price Index (CPI) measures the prices US households pay for a fixed basket of goods and services — food, energy, rent, cars, haircuts, insurance, and hundreds of other items. The BLS publishes it monthly, at 8:30 a.m. Eastern around the middle of the month, covering the month before. Its percentage change over twelve months is what news reports mean by “the inflation rate.”

Markets dissect the release along two axes. Headline versus core: the headline index includes everything, while core CPI strips out food and energy — not because groceries and gasoline don’t matter to households, but because their prices swing so much month to month that they can drown out the underlying trend central banks care about. Energy is only about 6% of the basket by weight, yet it regularly dominates the headline’s swings — in June 2026, headline CPI actually fell 0.4% on a gasoline slide while core was flat, and anyone trading the currency off that soft headline was really trading the oil price, not the inflation trend. Month-over-month versus year-over-year: the annual figure makes the headlines, but the monthly change, reported to one decimal place, is the fresher signal — and around it, tenths of a percentage point are moving markets. A monthly core reading of 0.4% when 0.3% was expected sounds trivial; compounded over a year, that gap is the difference between inflation converging to target and inflation stuck above it.

Then there is the detail that surprises even experienced readers: shelter. Housing is the single biggest component of the index — around 36% of the whole basket, most of it an estimate of what homeowners would pay to rent their own homes. And it is measured slowly by design: the CPI’s rent sample is surveyed only every six months, in rotating panels, with each observed change spread across the months between surveys. The result is that the largest slice of the US inflation number describes the rental market as it was several months to a year ago. When actual market rents turn, official shelter inflation follows with a long, mechanical lag — which is why analysts watching for inflation turning points often discount shelter and look at the timelier categories around it, and why “core CPI is sticky” sometimes means “the measurement is slow,” not “inflation is stubborn.”

Why does inflation data hit currencies so hard? Because inflation is the variable central banks are explicitly mandated to control — the Federal Reserve, for instance, defines its longer-run goal as 2% inflation. A hot CPI print raises the odds that the central bank holds rates higher for longer; a cool one raises the odds of cuts. (A wrinkle for the detail-minded: the Fed’s official 2% target is actually defined on a different index, the personal consumption expenditures price index, or PCE. CPI still dominates market reaction because it arrives about two weeks earlier each month and usually tells a similar story.)

One more habit separates careful readers from headline readers here: inflation data is a rate of change, so the base it is measured against matters. A year-over-year figure can fall simply because a big price spike twelve months ago dropped out of the calculation — the market calls these base effects — without anything improving in the current month. It is another reason the month-over-month sequence, not the annual headline, is where professionals look first. (A final wrinkle for completeness: unlike payrolls, the CPI’s published index levels are final when issued — but the seasonally adjusted monthly path is recalculated every February, revising up to five years of the month-over-month history you may have watched markets trade.)

Rate Decisions: The Main Event

Employment and inflation data are, in the end, inputs. The output is monetary policy — and the scheduled moments when central banks announce it are the heaviest events on the calendar.

In the United States, interest rate policy is set by the Federal Open Market Committee (FOMC), which holds eight scheduled meetings a year. Each ends with a written statement at 2:00 p.m. Eastern announcing the target range for the federal funds rate — the overnight rate that anchors borrowing costs across the economy — followed by the Chair’s press conference, with the meeting’s minutes arriving three weeks later. Four times a year the committee also publishes its Summary of Economic Projections, including the famous dot plot: each policymaker’s anonymous projection of where rates should go, plotted one dot per person. Rate moves come almost exclusively in multiples of a quarter of a percentage point — 25 basis points, a basis point being one hundredth of a percent.

The other major central banks run the same essential play with revealing differences in staging. The Bank of England’s nine-member committee announces at noon London time, eight times a year — and publishes its minutes with the decision, including exactly who voted for what, by name, so the whole information set lands at once. The European Central Bank announces at 2:15 p.m. central European time with a press conference half an hour later — and publishes no vote count at all; the closest thing to minutes, the “account,” follows about four weeks later. Same instrument, three different disclosure philosophies — which is worth knowing, because what to read first differs by central bank: the dots in Washington, the vote split in London, the press conference in Frankfurt.

Here is the part that surprises most people: the rate decision itself is usually the least interesting part of the event. By decision day, interest rate futures markets have typically priced the outcome with high confidence — you will see it reported as “markets price a 95% probability of a quarter-point cut,” figures usually derived from fed funds futures (CME’s FedWatch tool is the standard public reference) — and central banks, which prize predictability, rarely spring surprises they haven’t carefully telegraphed. What moves currencies is everything around the decision: the wording of the statement compared with the last one, the tone of the press conference, the shift in the dots, and the vote split. A central bank that cuts rates but signals the cut may be the last can see its currency strengthen on the day. Traders compress all of this into two words worth defining precisely — hawkish, leaning toward tighter policy and higher rates, usually to fight inflation, and dovish, leaning toward easier policy and lower rates to support growth. The market’s real question on decision day is never “what did they do?” It is “what did they tell us about what they’ll do next?”

Even a meeting where nothing happens can move markets, for exactly this reason. In July 2026, both the Fed and the Bank of England held rates — and at each, three members dissented in favor of a hike. The level did not change; the distribution of views did, and that is information about the future path. A hold with hawkish dissents is a different event from a unanimous hold, even though the two produce identical headlines. Professionals read the decision the way they read the jobs report: layers, not headlines.

The Two-Step: How a Data Surprise Becomes a Currency Move

Pull the threads together and a single mechanism emerges. Economic data moves currencies mostly through interest rate expectations, in two steps.

Step one: the surprise updates the policy outlook. A much stronger-than-expected jobs report, or a hotter-than-expected CPI, shifts the market’s view of what the central bank will do — fewer cuts, later cuts, maybe hikes. A weak surprise shifts it the other way. This repricing happens in seconds, in the interest rate futures markets, before most people have read past the headline.

Step two: the new rate path moves the currency. As we covered in What Moves Currency Prices?, expected interest rates are the reward for holding a currency, and capital flows toward the better reward. So the classic chain reads: hot data → higher expected rates → stronger currency, and cold data → lower expected rates → weaker currency.

The research on how releases actually move exchange rates broadly backs this framing — with wrinkles that deserve their own telling. The jobs report turns out to move currencies far more reliably than inflation data does, and the size of any reaction depends heavily on how the market happens to be positioned when the number lands. What actually unfolds in the seconds around a release — who is quoting, who has stepped back, and why the same surprise can produce wildly different moves — is the subject of this article’s companion piece, When the Number Drops (Members).

The chain is reliable enough to be the default assumption on any data day — and unreliable enough that it should never be treated as a law. When markets are gripped by fear, the safe-haven bid can swamp the rate logic entirely: the US dollar has strengthened on catastrophically bad news often enough that “bad news, strong dollar” is a recognized crisis pattern. And there are stretches when equity markets cheer weak data because it promises rate cuts — the “bad news is good news” regime — which can scramble the usual FX correlations too. The two-step is the right starting framework; the regime you are in decides how literally to apply it.

It is also not only a currency story. The same repricing of rate expectations that moves the dollar moves gold — which trades off real interest rates, as we explored in What Drives the Gold Price? — and, increasingly, crypto. One release, one mechanism, echoing across every asset class this site covers.

What Release Time Looks Like — Briefly

The mechanics around the moment of release are their own education, and they get their own article — but the outline belongs here.

In the last minute before a major print, market makers quietly step back: quoted spreads widen and the size available at the best prices shrinks, not because anyone knows the number, but because nobody wants to be the one showing a stale price a microsecond after it lands. Then the release hits, and the first responders are machines — pricing engines that parse the numbers and react within microseconds, millionths of a second, repricing long before any human has read a word. The price adjustment to the surprise is essentially complete within a few minutes; on payrolls days, trading volume runs several times its normal level for about two hours. And the first move is not always the lasting one — sharp initial reactions regularly reverse in part as human readers get past the headline to the revisions and the composition. It is why even experienced professionals treat the seconds around a release as close to untradeable, and why this site’s aim is to help you understand these moments rather than encourage you to trade them.

Who steps back and who stays in, why different clients of the same bank see different prices at 8:30, and what the research says about why the same surprise can move a currency violently one year and not at all the next — that is the anatomy we dissect in the companion piece, When the Number Drops (Members), which picks up exactly where this section stops.

The Bottom Line

Economic data moves currencies on a schedule everyone can read. The economic calendar tells you when; the consensus tells you what is already priced; and the big three release families — jobs, inflation, and rate decisions — supply most of the market-moving surprises. When a number lands, the market runs one calculation above all others: what does this change about the central bank’s next move? The answer travels from the data, through interest rate expectations, into the exchange rate — fast, first-draft, and frequently revised.

Learn to read releases that way — actual against consensus, details behind headlines, single prints inside trends — and the Friday-morning chaos resolves into something orderly: a market doing, at high speed and in public, exactly what this site keeps describing. Not reacting to news, but repricing expectations.

The Currency Stack provides educational and research content only. Nothing here is financial, investment, or trading advice, or a recommendation to buy or sell any asset. Markets carry risk; do your own research and consider professional advice before acting.

Further reading: This article pairs with When the Number Drops: How Markets Really Trade Economic Data (Members), which dissects what actually happens in the seconds around a release. For the framework underneath it all, read What Moves Currency Prices? Rates, Risk, and Macro and the pillar guide What Is the FX Market and Why Does It Matter?; the same rate-expectations machinery applied to gold is in What Drives the Gold Price? Keep the live economic calendar open as you read. Terms in bold are defined in our glossary. Sources: US Bureau of Labor Statistics release schedules, the Employment Situation Technical Note and CES benchmark documentation; Federal Reserve FOMC calendars and statements; Bank of England and European Central Bank policy pages; and Federal Reserve and BIS research on market reactions to macroeconomic announcements.