Core Inflation and Trend Inflation

Abstract

We thank Marco del Negro and Giorgio Primiceri for discussions on numerical methods. We also thank seminar participants at the Universitat Pompeu Fabra EC2 conference, the National Bank of Belgium, the Society for Nonlinear Dynamics and Econometrics Oslo meetings, and the Board of Governors of the Federal Reserve. The views expressed herein are those of the authors and do not necessarily reflect the views of the National Bureau of Economic Research. NBER working papers are circulated for discussion and comment purposes. They have not been peerreviewed or been subject to the review by the NBER Board of Directors that accompanies official NBER publications.

Estimating the U.S. output gap with state‐level data

“…Stock and Watson () argue that the width of the confidence set of the cycle, in this type of models, reflects two sources of uncertainty: the signal extraction uncertainty conditional on the values of the parameters of the model and the parameter uncertainty. The model with state‐level data contains more information than the model with aggregate data, and hence the signal extraction uncertainty is smaller in the former.…”

“…The model that I propose is similar to the one for the U.S. put forward by Stock and Watson (), which uses different inflation categories as the source of cross‐sectional variation to estimate common and category‐specific permanent and transitory components of inflation. In their paper, the common permanent component is denominated as trend inflation.…”

“…One advantage of introducing information about either multiple cross‐sections or variables to estimate a common component—the aggregate output gap in this case—is that one can gain efficiency compared with a model that considers a single cross‐section or variable (see Basistha & Startz, ; Fleischman & Roberts, ; Stock & Watson, ). However, the larger the cross‐sectional dimension or the number of variables, the more parametrized the model becomes, which can add uncertainty to the estimated common component.…”

Constructing and evaluating core inflation measures from component‐level inflation data

“…A potential drawback of their approach is that including all of the component‐level data on the right‐hand side of the regression risks overfitting and potentially compromises the out‐of‐sample forecasting performance. Stock and Watson () also employ a data‐rich method to estimate core inflation. They apply a dynamic factor model to 17 component‐level price series of the personal consumption expenditure deflator index.…”

“…The 50 components and 17 components are listed in Appendices A and B, respectively. The 50 components are the same as those used by Bermingham and D'Agostino () and the 17 components are the same as those used by Stock and Watson (). We use both sets of components to forecast and then compare the forecasting results from both within and across sets of components.…”