How Reliable are Cointegration-Based Estimates for Wealth Effects on Consumption? Evidence from Switzerland

Abstract

Summary According to economie theory, the intertemporal budget constraint of households implies that a permanent increase in wealth should have a positive effect on consumer spending. Given the comparatively strong increase in Swiss household wealth over the past few years, the question of the extent to which changes in wealth influence expenditures of households has become of special interest for Switzerland. In this paper, I show that while the link among consumption, wealth and income was quite strong from 1981 to 2000, it has been very unstable since 2001. This fact suggests that the gap among the three variables, i.e., the deviation from long-run equilibrium, that has opened over the last few years is less likely to close. The results apply to aggregate wealth effects as well as to separate financial and housing wealth effects. Furthermore, I document several fragility issues related to the use of the cointegration approach to estimating wealth effects. These issues highlight the importance of carefully checking the robustness of the results, instead of looking just at one cointegration estimation method and only one time period. They also highlight the need for a non-cointegration approach to estimating wealth effects.

Keywords

• Wealth effects
• consumption-to-wealth ratio
• cointegration
• cay residual