Disaggregate Wealth and Aggregate Consumption: an Investigation of Empirical Relationships for the G7*

Abstract

To date, studies of wealth effects on consumption have mainly used aggregate wealth definitions on a single-country basis. This study seeks to break new ground by analysing disaggregated financial wealth in consumption functions for G7 countries. Contrary to earlier empirical work, we find that illiquid financial wealth (i.e. securities, pensions and mortgage debt) tends to be a more important long-run determinant of consumption than liquid financial wealth. These results imply potential instability in consumption functions employing aggregate wealth. Our results are robust using SURE; when testing with a nested specification; and when using a linear model.

Consumption in Developed and Emerging Economies

“…Higher wealth effects, especially in the short run dynamics of adjustment, are thus likely both for illiquid financial assets (equities, bonds, pension assets) and non financial tangible wealth. Byrne and Davis (2003) highlighted that illiquid as well as liquid financial wealth is likely to become important in determining consumption in the G-7, and indeed showed in rolling regressions that there has been a rise in the long run impact of 6 Similar studies have been performed by Davis and Palumbo's (2001) study of the US consumption function, which attempted to determine whether changes in wealth as well as income affect the growth rate of consumer spending. Ludvigson and Steindel (1999) also examined wealth effects in a quarterly loglinear long-run US consumption relationship and found a common trend and a statistically significant wealth and income effect (Barrell and Davis 2007) illiquid financial wealth on consumption.…”

Housing Wealth, Credit Conditions and Consumption

“…For most countries, Byrne and Davis (2003) find liquid asset effects smaller than those from illiquid assets, and typically negative for the US, and especially the UK. Since liquid assets are defined as gross liquid assets minus debt, this is a classic symptom of omitted variable bias.…”

“…Other relevant controls are all missing. Two studies by Barrell and Davis (2004) and Byrne and Davis (2003) estimate equations for the G-5 and G-7 countries, respectively, employing no controls for shifts in credit conditions, interest rates, unemployment rates or expected income growth. The former paper aggregates wealth into net worth in log form.…”

14 Housing and Personal Wealth in a Global Context

“…Recent empirical studies of the housing-consumption link on macro data include Case et al (2005), Catte et al (2004), Iacoviello (2004), Barrel and Davis (2004), Dvornak and Kohler (2003), Byrne and Davis (2003), Ludwig and Sloek (2002) and Boone et al (2001). Earlier studies include Hendry et al (1990), Brodin and Nymoen (1992), Kennedy and Andersen (1994), and Muellbauer and Murphy (1995 …”

“…They estimate both single country equations and pooled equations imposing common long-run coefficients. Byrne and Davis (2003) estimate equations for G7 countries with no controls for shifts in credit conditions, interest rates, unemployment rates or expected income growth. They do not distinguish housing wealth but test for differences between liquid and illiquid assets effects.…”