An economic expansion is the rising phase of the business cycle, when broad real economic activity increases from a trough toward the next peak. For a quant researcher, the distinction matters: positive output growth, recovery of an old output level and a historically dated expansion are different pieces of information.
In the United States, the NBER chronology places expansions between troughs and peaks. Its monthly counting convention includes the peak month in the expansion and the trough month in the preceding recession. That jurisdiction-specific chronology is broader than the latest quarterly GDP reading.
Growing again can still mean producing less
Consider a fictional, seasonally adjusted quarterly real-output index. It was 110 at an earlier local peak, fell to 100, then rose to 105 in the next quarter. These are index levels, with the low quarter set to 100; they are not observations from an actual economy.
The quarter-to-quarter growth rate is
Output increased, but 105 is still 5 index points below the earlier peak of 110. The growth rate answers how fast output changed relative to the preceding quarter. The level answers how much was produced relative to the index's reference level. Neither question substitutes for the other.
Fictional seasonally adjusted quarterly real-output index; low quarter = 100. Lines connect observations. The 5% rebound is not annualised; this single series does not date an economy-wide expansion.
Keep the rate convention attached to the number. US BEA headline quarterly GDP changes are generally reported at annual rates, which compound the quarterly change over four quarters. The fictional 5% quarter-to-quarter figure above uses a different convention. See the BEA's explanation before comparing growth headlines.
Recovery is part of an expansion
One useful convention calls the early part of an expansion a recovery until activity regains its previous peak. The Richmond Fed describes this subdivision and notes that the NBER does not separately date it. Under that convention, the rise from 100 to 105 is a partial recovery of this fictional output series.
An expansion can therefore begin while output remains depressed. It does not require every household, sector or indicator to improve together. The NBER FAQ describes a range of real-activity and employment measures, without fixed weights. An isolated rebound in one series supplies evidence to assess; it does not establish the broad phase by itself.
Real also matters. Higher nominal sales may reflect inflation. The RBA's growth explainer distinguishes price changes from changes in the volume produced. Even real aggregate growth says little on its own about per-person output or how income gains are distributed.
The phase and the signal have different dates
The NBER dates turning points retrospectively, after enough evidence has arrived. A trough's observation month and its announcement date are different. In a backtest, an eventual expansion label can describe historical outcomes; using it to trigger a trade before its announcement introduces look-ahead bias.
Revisions create a second timing problem. A modern download may contain a different history from the one a researcher could have seen at the decision time. The Philadelphia Fed's real-time dataset preserves successive vintages, or snapshots, for this purpose.
For an implementable expansion signal, specify the jurisdiction, indicators, real or nominal basis, frequency, seasonal adjustment and rate convention. Preserve publication times and the vintage available at each decision. Evaluate that real-time estimate separately from the eventual cycle chronology.
The advanced MV real-time macro forecasting capstone develops release-date and vintage discipline for US industrial-production forecasts. Membership and sign-in are required. It helps with the information problem here; its forecast target is not an official expansion classification.
Sources
- NBER — Business Cycle Dating: chronology, monthly phase convention and retrospective dating.
- NBER — Business Cycle Dating Procedure: Frequently Asked Questions: expansion versus previous peak and breadth of indicators.
- Federal Reserve Bank of Richmond — Business Cycles: recovery as a conceptual subdivision of expansion.
- US Bureau of Economic Analysis — Why publish quarterly percent changes at annual rates?: quarterly versus compounded annual rates.
- Reserve Bank of Australia — Economic Growth: real versus nominal output and limits of aggregate GDP.
- Federal Reserve Bank of Philadelphia — Real-Time Data Set for Macroeconomists: historical vintage snapshots.