Practice Makes Perfect: Learning Effects with Household Point and Density Forecasts of Inflation

53 Pages Posted: 13 Nov 2024

See all articles by James Mitchell

James Mitchell

Federal Reserve Bank of Cleveland

Taylor Shiroff

Federal Reserve Bank of Cleveland

Hana Braitsch

Federal Reserve Bank of Cleveland

Date Written: November 13, 2024

Abstract

This paper shows how both the characteristics and the accuracy of the point and density forecasts from a well-known panel data survey of households' inflationary expectations – the New York Fed's Survey of Consumer Expectations – depend on the tenure of survey respondents. Households' point and density forecasts of inflation become significantly more accurate with repeated practice of completing the survey. These learning gains are best identified when tenure-based combination forecasts are constructed. Tenured households on average produce lower point forecasts of inflation, perceive less forecast uncertainty, round their uncertainty but not their point forecasts, report unimodal densities, and provide internally consistent point and density forecasts.

Keywords: inflation expectations, surveys, forecaster heterogeneity, combination forecasts, density forecasting, learning

JEL Classification: C53, D84, E31, E37

Suggested Citation

Mitchell, James and Shiroff, Taylor and Braitsch, Hana, Practice Makes Perfect: Learning Effects with Household Point and Density Forecasts of Inflation (November 13, 2024). FRB of Cleveland Working Paper No. 24-25, https://doi.org/10.26509/frbc-wp-202425, Available at SSRN: https://ssrn.com/abstract=5019644 or http://dx.doi.org/10.2139/ssrn.5019644

James Mitchell (Contact Author)

Federal Reserve Bank of Cleveland ( email )

East 6th & Superior
Cleveland, OH 44101-1387
United States

HOME PAGE: http://https://www.clevelandfed.org/en/our-research/economists/james-mitchell.aspx

Taylor Shiroff

Federal Reserve Bank of Cleveland ( email )

Hana Braitsch

Federal Reserve Bank of Cleveland ( email )

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
13
Abstract Views
131
PlumX Metrics