Potential output is an estimate of the real output an economy can sustain when labour and capital are used at rates consistent with stable inflation. For a quant studying macroeconomic regimes, it provides the benchmark against which observed output can indicate spare capacity or demand pressure. The benchmark is unobserved and must be estimated. RBA, sections 4.1 and 4.4.
The word potential does not mean every machine operates continuously or every person works. Maintenance, normal working hours and sustainable resource use still apply. Actual output can temporarily exceed potential, so potential is not an absolute physical ceiling. Nor is it a forecast of what actual GDP will be next year. A forecast must also account for demand and the economy's adjustment towards or away from capacity. CBO's February 2026 outlook accordingly distinguishes growth of real GDP from growth of real potential GDP.
A growing benchmark can coexist with slack
Let denote estimated potential real output and actual real output in the same period, units and price basis. Potential growth is
Consider a fictional economy with annual real-output indices on a common scale. Its estimated potential is 100 in year 0 and 102 in year 1:
Year 1 actual output is 98. The output gap, using potential as the denominator, is
The benchmark grew by 2%, while actual output is about 3.92% below the year 1 benchmark. These statements answer different questions. Neither tells us actual GDP growth, because year 0 actual output was not supplied. The four-index-point shortfall is also different from a four-percent shortfall: the denominator is 102.
Fictional annual indices on one common real-output scale. Left: growth of the estimated benchmark differs from actual output relative to that benchmark. Right: changing the year 1 estimate changes the measured gap without changing actual output.
What moves potential, and how is it estimated?
Potential can rise through a larger sustainable labour input, capital accumulation, or greater productivity. CBO's US framework considers labour-force trends, noncyclical unemployment, capital and productivity. These are inputs to an estimate, rather than directly observed measurements of potential itself. CBO, Growth of Real Potential GDP.
Several approaches address the estimation problem:
| Approach | What it uses | Main interpretive limit |
|---|---|---|
| Statistical trend | Separates persistent movements in real output from shorter fluctuations | A smooth trend need not equal an inflation-consistent capacity benchmark; the endpoint is sensitive to new data |
| Production function | Combines estimates of sustainable labour, capital and productivity | The inputs and their trends must themselves be estimated |
| Multivariate model | Uses output with information such as unemployment and inflation | Economic relationships and assumptions affect how it separates supply from demand |
The IMF's measurement discussion explains statistical and production-function approaches. The OECD's 2019 methodology combines a production function with filtering of underlying components and cyclical adjustments intended to reduce endpoint problems. The RBA uses several models, including multivariate and production-function approaches. Different models can embody different horizons and concepts of sustainable supply; agreement is not guaranteed.
Treat the estimate as dated information
The most recent estimate is especially fragile: a filter has no future observations to help distinguish a temporary dip from a persistent change. New data, revised GDP and changed model assumptions can all alter the estimated historical path. This is why an old pre-shock trend should not automatically be treated as today's capacity. IMF, Hard to measure.
The fictional example makes the consequence concrete. Holding actual output at 98, a revised year 1 potential estimate of 100 would imply . The measured shortfall narrows even though no additional production has occurred. This comparison changes only the benchmark; real releases may revise both quantities.
For a historical signal or backtest, preserve the provider, model, release date and data vintage available at each decision time. Using today's revised potential path to represent yesterday's information introduces look-ahead bias. Compare matching real-output definitions, frequencies and seasonal-adjustment conventions before calculating a gap.
Potential also informs how much economic growth may be sustained without increasing demand pressure. Weak observed growth alone does not establish abundant slack: the supply benchmark may also be growing slowly. The RBA's August 2026 assessment, sections 2.5–2.6, illustrates this reasoning by considering productivity and capacity indicators together. A gap estimate is one input to an inflation assessment, alongside supply shocks and other information; it is not a mechanical forecast of inflation or interest rates.
Sources
- RBA — In Depth: Potential Output, May 2024: sustainable capacity, model approaches and uncertainty.
- CBO — The Budget and Economic Outlook: 2026 to 2036: US real potential GDP and its inputs.
- OECD — The OECD potential output estimation methodology, 2019: production-function and filtering framework.
- IMF — The Output Gap: Veering from Potential: estimating unobserved potential and endpoint limitations.
- RBA — Economic Conditions, August 2026: a dated application of potential supply and capacity assessment.