Using Survey Expectations to Measure the Output Gap

ABSTRACT Research Question/Issue How can survey‐based measures of expected output be used to construct a reliable, timely measure of the output gap for the Australian economy? Research Findings/Insights A Vector‐Autoregressive model of actual and expected output (VAR‐E), estimated using Consensus Economics survey forecasts alongside ABS GDP data, yields a Beveridge‐Nelson trend and associated output gap with desirable statistical properties, including parsimony and the ability to capture complex output dynamics without the excess volatility typical of univariate trend estimates. Decomposing shocks into short‐lived and longer‐lived components produces two gap measures — an unadjusted gap and a news‐adjusted gap that abstracts from very short‐lived shocks — which diverge by more than 1 percentage point in 10 of 71 quarters between 2008 and 2026. Practitioner/Policy Implications The news‐adjusted gap is particularly relevant where policy actions take time to affect the economy, offering monetary authorities and forecasters a benchmark measure less distorted by transient noise than conventional real‐time gap estimates. Methods Used A six‐variable VAR‐E is estimated on quarterly Australian GDP growth (2008q1–2026q1) combined with nowcasts and one‐ to four‐quarter‐ahead survey forecasts. The Beveridge‐Nelson trend and a Blanchard‐Quah approach to identifying short‐lived versus longer‐lived transitory shocks are then used to derive the respective gap measures.