June 22, 1998

NOTE TO EDITORS

The Article Does Consumer Confidence Forecast Household Expenditure? A Sentiment Index Horse Race appears in the latest edition of the New York Fed's Economic Policy Review.

This article looks at consumer attitudes through a comparison of the forecasting power of two popular surveys of consumer attitudes--the University of Michigan Index of Consumer Sentiment and the Conference Board Consumer Confidence Index.

New York Fed economists Jason Bram and Sydney Ludvigson describe the two indexes and examine whether each index contains information that can help predict spending in five categories of household consumption: total expenditures, motor vehicles, all goods (excluding motor vehicles), services, and durable goods (excluding motor vehicles).

The authors conclude that:

  • The Conference Board index contains significant explanatory power for most categories of consumer spending, including total personal consumption expenditures, motor vehicles, services, and durable goods (excluding motor vehicles).
  • By contrast, the Michigan index helps forecast spending in only two consumption categories--motor vehicles and goods (excluding motor vehicles). Even in these categories, however, the forecasting power of the Michigan index is weaker than that of the Conference Board index.

  • The superior predictive performance of the Conference Board index appears to be related to the types of questions that make up the survey. The two Conference Board questions that ask specifically about job prospects in the respondent's area exhibit the most predictive power.

  • Given the article's findings, policymakers and forecasters may wish to pay close attention to the survey questions that generate a large shift in consumer confidence. For example, a surge in consumer confidence that is driven by the questions about job availability may signal greater potential for an increase in consumer spending.

Contact: Douglas Tillett