The output gap is the difference between actual real GDP and estimated potential GDP, usually expressed as a percentage of potential GDP. It asks whether production is above or below the economy's sustainable capacity. For a quant researcher, it is a possible measure of economic slack, but its estimate and publication date belong in the dataset alongside its value.
Potential output is an estimate, not a physical ceiling. It describes sustainable production with labour and capital employed at sustainable rates. An economy can temporarily produce above it by using resources unusually intensively. It is neither the highest output ever recorded nor the output achievable by making everybody work continuously. See the CBO's definition of real potential GDP.
The denominator is potential output
Let be actual real output and estimated potential real output for the same period. The proportional gap is
Report that ratio in percent. The OECD's measurement convention uses the difference in volume output relative to potential GDP.
Consider two independent fictional annual scenarios. Both use a common real-output index scale on which estimated potential output is 100:
- Actual output of 98 gives : a negative gap, indicating slack relative to the estimate.
- Actual output of 103 gives : a positive gap, indicating production above the estimated sustainable level.
Fictional annual scenarios, not observations or a time series. The dashed benchmark is estimated potential output; signed gaps use potential output as the denominator.
The inputs must cover the same economy, period and output definition, with compatible price references, seasonal adjustment and annualisation conventions. Do not divide a nominal GDP level by a real potential GDP level, or independently rebase each series so that meaningful level differences disappear.
A gap is also different from economic growth. Growth compares output across dates; the gap compares actual and potential output at one date. Production can grow while remaining below potential, and weak growth can coexist with a positive gap if potential output is also weak.
Why different estimates disagree
Potential GDP cannot be directly observed. A production-function approach estimates capacity using capital, sustainable labour input and productivity; the OECD uses this general approach. Statistical filters separate a trend from shorter fluctuations. Multivariate models also use economic relationships, including signals from inflation and unemployment. These methods need not produce the same benchmark. The RBA explains its suite of approaches rather than selecting one model as definitive.
Uncertainty is especially relevant at the end of a sample: later observations can change a filter's estimate of the earlier trend. This is the end-point problem described by RBA researchers. Revised GDP, revised inputs and changed model assumptions can also change a historical gap. A precise-looking published number is still an estimate. The IMF's discussion of real-time output-gap revisions shows why an assessment made with later information can differ from the one available to policymakers at the time.
For a backtest, retain the release vintage and the date the estimate became available. A full-sample trend fitted today can use future observations when estimating yesterday's gap. Treating that estimate as yesterday's signal introduces look-ahead bias. Compare models using the information each forecast origin actually allowed.
Capacity pressure is one input
A persistent positive gap can contribute to upward pressure on inflation; a negative one can indicate spare capacity. Neither sign fixes the next inflation print or the next monetary policy decision. Supply shocks, expectations and input costs also matter, and the RBA assesses output gaps alongside other indicators. The gap alone supplies neither an automatic trading rule nor a return forecast.
A revision check
Fictional annual real GDP is 204. Potential GDP is initially estimated at 200, then revised to 208 for that same year. Actual GDP is unchanged. Does the output gap keep its sign?
Answer. Initially, . After the revision, . The estimated gap changes from positive to negative without any change in actual production. The benchmark changed. Using the later estimate in an earlier decision would change the information set.
Sources
-
RBA — Output Gaps in Real Time, RDP 2002-06, abstract: end-point uncertainty and real-time estimation.
-
OECD — Sources and methods of the Economic Outlook, “Output gaps”: volume measurement, denominator and production-function approach.
-
CBO — The Budget and Economic Outlook: 2026 to 2036, “Potential Output”: sustainable-resource definition and supply-side inputs; no forecast numbers are used here.
-
RBA — In Depth: Potential Output, May 2024: alternative models, estimation uncertainty and inflation interpretation.
-
IMF — Making Monetary Policy Decisions in the Dark, August 2015: revisions and the distinction between real-time and later assessments.