Independent evaluations of the US Consumer Demand Index

LEADING INDICATOR PROPERTIES OF THE "US CONSUMER DEMAND INDEX": FORECASTING THE US RECESSION

Associate professor, Heino Bohn Nielsen,Department of Economics, University of Copenhagen May 26, 2009

Abstract: Using the leading indicator properties of the new US Consumer Demand Index (CDI) we illustrate that the current downturn in retail sales and industrial production was partly forecastable. The CDI appears to be more useful in forecasting the current movements in consumption than conventional measures of consumer confidence.

Introduction and Summary

The Consumer Demand Index (CDI) is a new monthly indicator for American households' planned purchase the coming three month published by the Institute for Business Cycle Research (IFKA). The CDI is conceptually different from the conventional measures of consumer confidence in that it aims at directly measuring households buying decisions rather than expectations or degree of confidence.

This note analyzes the leading-indicator properties of the composite CDI index for changes in industrial production and retail sales. Based on simple linear regression models we consider the in-sample behavior for the period 2001 : 2 - 2008 : 6, and use the final preferred specification to produce forecasts for the downturn 2008 : 7 - 2009 : 3 conditional on observations for sales up to 2008 : 6 and the sequence of observations for the CDI. We conclude that the CDI seems to contain useful information. We also compare with the properties of the consumer confidence index from The Conference Board, and conclude that the forecast properties of the CDI seem to outperform the more conventional indicator.

The Predictive Power of The US Consumer Demand Index

Dr. Carsten-Patrick Meier, Kiel Economics, Wissenschaftszentrum, Kiel May 2009

Summary

We take a fresh look at the question whether The US Consumer Demand Index is useful for forecasting U.S. private consumption expenditure and U.S. retail sales. The current cyclical weakness in the United States has served the cyclical features of the index quite well. The predictive properties of the index have improved significantly as a result of the sharp reduction in U.S. private consumption expenditure and retail sales, which was more or less correctly indicated by the US CDI. In addition, the variance generated in the current crisis allowed the evaluation to stand on firmer ground. While a definite conclusion is still difficult given the short data at hand for the evaluation, the study produces new evidence that the US CDI helps to forecast private consumption expenditure and retail sales. It may also be used to forecast the University of Michigan Consumer Confidence Index.

The full text of the two independent evaluations is available on request - please send e-mail to jt@consumerdemand.com