World Economy :: Consumer Confidence

Matsusaka & Sbordone (1995) used quarterly data from 1953 to 1988 to empirically investigate the relationship between consumer confidence and Gross National Product in the United States. Using vector auto regressions Matsusaka ( 1995) & Afshar, Arabian, Zomorrodian (2007) examined the effect of pure confidence on gross national product by implementing control variables such as the Index of leading indicators & Consumer price index respectively finding that confidence granger caused GNP for 1, 2, 3 & 4 quarter lag models. Through forecast error variance decompositions they concluded that 12%-26% & 8-23% of the variation in GNP can be attributed to consumer confidence. The slight variations in the decompositions can b accounted for by variable orderings and differing time periods. Similarly Utaka (2003) applied the same methodology used by Matsusaka & Sbordone(1995) to empirically investigate this relationship in the case of Japan. By development not only quarterly data, but month ly and semi annual data from 1980q1- 2000q3 they reconfirmed the consumer confidence- GDP relationship for Japan and assemble that it accounted for 9%-11% of variation in GNP. This showed smaller variation for Japan illustrating that confidence indicators influence on GDP movements are country dependent making it unwise to assume uniform relationship strength across countries. Golinelli & Parigi (2004) investigated this matter by analysing the consumer confidence relationship in eight countrys from 1970-2002. They modelled the CCI-GDP relationship using a co-integrated vector auto regression using a common set of macroeconomic variables that were country specific to control for correlation being control by other variables therefore avoiding the limits of the single equation approach found in previous literature. They tested the forecasting power by comparing the RMSE for open and restricted models for 1, 2 & 4 steps horizons. Golinelli (2004), Mourougane & Roma (2002), Taylor & Mcnab (2007) find that RMSE was generally lower in the unrestricted model at short term horizons (1-2 steps) for EU countries, illustrating its importance in short term forecasting. Much of previous literature has aimed at establishing whether consumer confidence indexes provide additional information in comparison to macroeconomic variables and not its forecasting power. However, these have received mixed results in most cases, yet it is acknowledged they maintained an autonomous role in forecasting, (see Mueller 1963, Adams 1964, Suits &sparks 1965, fair 1971 a & 1971b, Adams & Klein 1972). Opinion now acknowledges that the index can help predict economic activity (see Garner 1991, Fuhrer 1993 Carol et al 1994, Kumar et al 1995, Bran & Ludvigson 1998, Eppright et al 1998)