BLS CPI and BEA GDP Releases: What They Measure, When They Land and How to Read Them

What to take away

Two releases, two agencies, one morning

The two numbers that dominate American economic coverage come from separate agencies with separate survey systems. Reporters often file them as one story. They are not.

The Bureau of Labor Statistics, part of the U.S. Department of Labor, runs the Consumer Price Index. CPI measures the average change over time in prices urban consumers pay for a market basket of goods and services: food, housing, transportation, medical care and more.

The Bureau of Economic Analysis, part of the U.S. Department of Commerce, produces gross domestic product. GDP measures the value of final goods and services produced in the United States in a given period, and it is the broadest single gauge of national output.

Both agencies release at 8:30 a.m. Eastern and post the full text online the moment the lock lifts. That shared habit is why one morning can carry inflation and growth news at once.

CPI describes prices. GDP describes output. A month can bring hot inflation and weak growth, or the reverse, and coverage that blurs the two misleads readers about their own budgets.

A useful frame: CPI answers what things cost, GDP answers how much the country produced. Household effects flow from both, through different channels. Our economic release guide walks those channels for reporters who need to explain them fast.

The agencies also revise differently. BLS revises CPI lightly, mostly for seasonal factors. BEA revises GDP repeatedly as source data mature. That asymmetry sets how careful you should be with each first print.

The CPI schedule and what lands when

CPI comes out monthly, typically in the second or third week, covering the previous month. BLS posts the full schedule in advance, and the release is a complete document rather than a data dump.

Each release carries the headline all-items index, the core index excluding food and energy, and detailed indexes for dozens of categories. It also carries the unadjusted 12-month change, the figure most often written into wage and benefit contracts.

BLS publishes CPI for two populations: All Urban Consumers, or CPI-U, and Urban Wage Earners and Clerical Workers, or CPI-W. CPI-U covers roughly 93 percent of the U.S. population and is the series newsrooms cite. CPI-W feeds Social Security cost-of-living adjustments.

Regional detail matters for local coverage. BLS publishes indexes for the Northeast, Midwest, South and West. It also covers larger metros such as New York, Los Angeles, Chicago and Atlanta. Not every metro gets a monthly index; some publish bimonthly or semiannually.

The release also includes real earnings, which pair price change with average hourly earnings. That pairing is where CPI stops being an abstraction and starts describing purchasing power. Our piece on inflation household context shows how to turn those tables into reader-facing numbers.

Timing rules are strict. BLS releases CPI at 8:30 a.m. Eastern. Credentialed media get an embargoed lockup, but the public web release goes live at the same moment. Nobody outside the lockup sees it early.

The CPI Home page explains how the index is built, which series the agency publishes and how each release is assembled. BLS posts the full CPI news release as the lock lifts, so readers can go straight to the tables.

How BEA publishes GDP

GDP does not arrive once. It arrives three times per quarter, and each print carries a different name and a different level of confidence.

The advance estimate comes roughly four weeks after the quarter ends. The second estimate follows about a month later, the third about a month after that. Each incorporates more source data than the last.

BEA describes GDP as the value of final goods and services produced in the United States. Its headline figure is annualized: the quarterly change scaled as if it continued for a full year. That convention surprises readers expecting a plain quarterly rate.

The release breaks out personal consumption, business investment, inventories, government spending and net exports. Those components explain why the headline moved, which is usually the more useful story.

BEA also publishes gross domestic income, or GDI, alongside GDP. In theory the two match, since one measures output and the other the income that output generates. In practice they diverge, and the gap is worth reporting.

Everything lands on the BEA gross domestic product page at the scheduled time, with tables, charts and a full news release. Our comparison of economic indicators compared puts GDP next to CPI, PCE prices, GDI, retail sales and real earnings in one frame.

Seasonal adjustment, plainly

Prices swing with the calendar. Gasoline rises before summer driving season. Airfares spike around holidays. Lettuce gets expensive in winter. Seasonal adjustment removes those recurring patterns so adjacent months can be compared.

Both agencies adjust their headline series. BLS recalculates seasonal factors annually and revises the prior five years of adjusted data each February. That is why January CPI coverage sometimes notes that recent history changed.

Seasonal adjustment is a statistical model, not a judgment about what matters. It estimates what a normal January looks like, then reports how this January differs. When a winter storm or a pandemic breaks the pattern, the model can misfire.

BLS publishes both adjusted and unadjusted series, and the choice matters. The 12-month unadjusted change shows up in contracts and benefit formulas. The adjusted monthly change shows up in market coverage.

GDP gets the same treatment. BEA adjusts for weather, holidays and other recurring factors, and publishes the raw data too. Quarter-to-quarter comparisons usually rely on the adjusted, annualized figure.

A practical rule: say which series you are citing. "Prices rose 0.4 percent in August" is ambiguous. "Seasonally adjusted prices rose 0.4 percent in August" is a fact a reader can check.

Real versus nominal

Nominal figures are stated in current dollars. Real figures are adjusted to remove price change. The gap between them is inflation, and it compounds over time.

If nominal GDP grows 5 percent while prices rise 3 percent, real output grew about 2 percent. Report the nominal number as growth and you overstate what the economy produced. That error appears every quarter.

CPI splits the same way. A nominal wage increase of 4 percent against 3 percent inflation leaves a real gain of roughly 1 percent. Readers care about the real number, because that is what they can spend.

BEA publishes both. Its headline GDP figure is real, chained to a base year and adjusted for price change. The nominal figure, sometimes called current-dollar GDP, appears in the same release.

Deflators do the work. The GDP price index and the PCE price index are the broadest inflation measures in the national accounts. CPI tracks a fixed urban basket; the deflators track what the economy actually bought.

That difference explains why CPI and PCE inflation rarely match. Coverage that treats them as interchangeable muddles what the Federal Reserve is watching. The Fed targets PCE inflation, not CPI.

Revision patterns that trip up early coverage

First prints are provisional. BEA revises GDP as quarterly and annual source data arrive, and those revisions can be large enough to change a story a newsroom already published.

Advance-to-second-estimate revisions are routine and usually modest, but not always small. Inventory and trade data arrive late and can move the headline by several tenths of a percentage point.

Annual revisions go deeper. BEA updates benchmarks, folds in new source data and can revise several years of history at once. A quarter once reported as contraction can later read as growth.

CPI revisions are smaller but real. BLS revises seasonal factors each February and corrects errors when they occur. Unadjusted indexes are not revised, which is one reason contracts lean on them.

The reporting lesson is to timestamp your figures. "GDP grew at a 2.1 percent annual rate in the second quarter, according to the advance estimate" tells a reader exactly what they are looking at. A bare number does not.

Our economic reporting problems piece catalogs the common errors, from mixing series to ignoring base effects, and pairs each with a fix. The economic release checklist covers the mechanics of a release morning.

How the numbers reach readers

Most Americans never open the BLS or BEA releases. They meet the numbers through wire copy, network segments, push alerts and social posts, usually within minutes of 8:30 a.m. Eastern.

Wire services prepare stories in advance with the numbers blanked out, then fill them in at the lock. That is why coverage appears almost instantly, and why an early error propagates fast.

Television and radio pick up the wire copy. Digital outlets rewrite it, often adding market reaction. By mid-morning the same two or three figures sit on hundreds of pages, frequently stripped of the caveats in the original release.

Compression is where meaning gets lost. A seasonally adjusted monthly change becomes "prices rose." An annualized quarterly rate becomes "the economy grew." Both drop the qualifiers that make the numbers interpretable.

Readers can skip the middle layer. The release pages from both agencies are free and need no registration.

Here is how a single morning plays out.

The GDP sequence stretches longer. An advance estimate in late July is followed by a second estimate in late August and a third in late September. Each can move the headline, and annual revisions can move it again.

The discipline for this beat: treat the first print as a draft, cite the estimate by name, and check the calendar before you publish.

Where to check the schedule

Schedules are published well in advance and updated when needed. Two pages matter most.

BLS maintains a schedule of selected releases covering employment, CPI, producer prices and real earnings. It lists dates and times months ahead and flags changes. BEA maintains its own release schedule covering GDP, personal income, trade and industry accounts.

Bookmark both, and confirm the date the morning of the release. Schedules occasionally shift for federal holidays or operational reasons, and a story citing the wrong release date undermines everything else in it.

When a number surprises, resist explaining it before checking the components. Most surprises live in one or two categories, and the release usually names them.

Correct promptly. Revisions are normal, and a newsroom that updates when the second estimate lands builds more trust than one that quietly leaves the first print standing.

Common questions

What time does the CPI release come out?

BLS releases CPI at 8:30 a.m. Eastern on the scheduled date, with the full document posted online at the same moment. Credentialed media work from an embargoed lockup, but the public release goes live simultaneously.

How often is GDP released?

GDP gets three estimates per quarter, plus annual revisions. The advance estimate comes about four weeks after the quarter ends, the second about a month later, and the third about a month after that.

What is the difference between real and nominal GDP?

Real GDP is adjusted for price change; nominal GDP is not. Growth coverage uses the real figure, while the nominal figure appears in the same release for level comparisons.

Why do CPI and GDP numbers get revised?

More source data arrives over time. BEA revises GDP repeatedly as quarterly and annual data mature, while BLS recalculates seasonal factors each February and corrects errors when they occur.