Intuitive and Reliable Estimates of the Output Gap from a Beveridge-Nelson Filter

Abstract

BIS Working Papers are written by members of the Monetary and Economic Department of the Bank for International Settlements, and from time to time by other economists, and are published by the Bank. The papers are on subjects of topical interest and are technical in character. The views expressed in them are those of their authors and not necessarily the views of the BIS. This publication is available on the BIS website (www.bis.org).

Estimating and accounting for the output gap with large Bayesian vector autoregressions

“…In particular, Nelson (2008) argues that an AR(1) model of output growth is a sensible choice for performing a BN decomposition for aggregate output because, while extremely parsimonious, it produces comparatively good out‐of‐sample forecasts. Thus we view a competitive out‐of‐sample forecast for output growth relative to a univariate AR(1) model as crucial to address Nelson's critique of standard approaches to trend‐cycle decomposition such as the Hodrick‐Prescott (HP) filter that have implicit out‐of‐sample forecasts of output growth that perform much worse (e.g., see Kamber et al., ).…”

“…Figure plots the estimated output gap for our 23‐variable benchmark model, along with 90% credible sets calculated using the approach for the BN decomposition developed in Kamber et al. (). The output gap is positively associated with NBER reference cycles and tends to be most significantly different from zero during recessions and the ends of expansions, with the degree of uncertainty consistent with findings for other measures of the output gap in Garratt et al.…”

“…The dotted lines indicated the bounds of the 90% credible set, calculated using the approach detailed by Kamber et al. (). Shaded bars correspond to NBER recession dates [Colour figure can be viewed at wileyonlinelibrary.com]…”

“…First, recall that Figure in the previous section suggests that the estimated output gap is not completely robust to the size of the system during the Great Recession, although it is, perhaps surprisingly, always smaller in magnitude than in the 1981–82 recession. In a univariate setting, accounting for a break in long‐run output growth around the Great Recession has been shown to significantly alter inference about the output gap based on long‐run forecasts (e.g., Eo & Morley, ; Kamber et al., ). Thus we check if this is also the case in a multivariate setting, with the estimates across different‐sized systems possibly influenced to differing degrees by a failure to account for structural change.…”

Growth Trends and Systematic Patterns of Booms and Busts‐Testing 200 Years of Business Cycle Dynamics

“…The ‘output‐gap’ definition defines the business cycle as ‘transitory fluctuations of the economy away from a long‐run trend’ (Morley and Piger, , p. 209), which is not necessarily independent of long‐run growth. Morley and Piger () and Kamber, Morley and Wong () take this argument further by emphasizing the importance of allowing for correlation between permanent and transitory movements. Thus, trend and cycles interact.…”