What Is a Recession, and Who Decides When One Has Started?

Almost everyone repeats the same definition: two consecutive quarters of falling GDP. It is a useful shorthand and it is not the official definition anywhere.

Understanding how recessions are actually identified explains why the announcement always arrives late, and why markets so often move in what looks like the wrong direction.

What a recession actually is

The working definition used by economists is broader: a significant decline in economic activity, spread across the economy, lasting more than a few months.

Three elements matter, and all three must hold:

• Depth — the decline is significant, not a rounding error.
• Diffusion — it is spread across sectors, not confined to one industry.
• Duration — it persists rather than being a one-month blip.

A sharp fall concentrated in a single sector is not a recession. A shallow, broad, sustained decline can be.

Who decides

In the United States, the National Bureau of Economic Research — a private non-profit, not a government body — maintains the official chronology through a business cycle dating committee.

The committee examines several measures rather than GDP alone:

• Employment
• Real personal income excluding transfers
• Industrial production
• Real consumer spending
• Wholesale and retail sales

It waits deliberately for data revisions before deciding. The consequence is that recessions are announced long after they begin, sometimes after they have already ended. This is not incompetence; it is a preference for being right over being fast, and initial economic data is revised substantially.

Other countries use their national statistical agencies or similar committees, and several do lean on the two-quarters convention more heavily than the US does.

Why the two-quarters rule misleads

It fails in both directions.

It can miss real recessions. GDP can be propped up by one component — government spending, or inventory building — while employment falls and activity contracts broadly. The NBER has declared recessions without two consecutive negative quarters.

It can flag non-recessions. Two mildly negative quarters driven by a technical factor such as a trade or inventory swing, with employment still growing, is not the broad contraction the concept describes.

The rule survives because it is simple and checkable from a single published number. That is its only advantage.

Indicators that move earlier than GDP

GDP is comprehensive and slow. These arrive sooner:

Indicator	Why it leads
Weekly jobless claims	Published weekly; firms cut hours and staff early
Purchasing managers indices	Survey-based, monthly, forward-looking; 50 divides expansion from contraction
Credit spreads	Corporate borrowing costs widen before equity markets react
Yield curve inversion	Signals expectations of rate cuts, usually implying expected weakness
Consumer confidence	Spending intentions shift before spending does
Building permits	Construction responds early to financing conditions

On the yield curve specifically: it has preceded most US recessions since the 1960s, which is why it gets so much attention. It is worth holding that claim carefully. The lag between inversion and recession has ranged from a few months to over two years, it has signalled falsely, and the total number of recessions in the sample is small. It is a signal worth knowing, not a timing tool.

What happens during one

Unemployment rises, and it is a lagging indicator — firms cut staff after conditions deteriorate and rehire well after recovery begins.

Central banks cut interest rates to support demand, which lowers bond yields and raises bond prices.

Corporate earnings fall, and companies with heavy debt or cyclical demand suffer most.

Credit tightens. Banks lend less freely precisely when borrowers need it, which amplifies the downturn.

Government borrowing rises, as tax revenue falls and support spending increases.

Why markets recover before the news does

This is the most practically important point, and the most counterintuitive.

Share prices reflect expectations about the future, typically six to eighteen months out. So markets tend to fall before a recession is visible in the data, and to bottom while the news is still bad — often while unemployment is still rising.

The result: waiting for confirmation that things have improved means buying after the recovery has already happened. Investors who exit during the worst headlines and re-enter once conditions feel safe have systematically bought back higher.

This is not an argument for market timing. It is an argument against it — the signals arrive in the wrong order for it to work.

What to actually watch

• Jobless claims, weekly, as the fastest read on labour conditions.
• PMIs for both manufacturing and services, monthly.
• Credit spreads between corporate and government bonds.
• The yield curve, understood as context rather than a countdown.
• Revisions to previous data, which frequently change the picture more than the new number does.

The bottom line

A recession is a significant, broad, sustained decline in activity — not an arithmetic result from two GDP prints. It is dated by committees using multiple indicators, and announced long after it starts.

For anyone following markets, the useful implication is that the official confirmation is the least actionable piece of information in the sequence. By the time it arrives, markets have usually moved on.

This article is educational and is not financial advice.