Revisiting the wealth effect on consumption in New Zealand

Release date
30/03/2017
Reference
AN2017/03
Author
Martin Wong
ISSN
2230-5505

Understanding the link between changes in wealth and household consumption is crucial to the Reserve Bank’s assessment of business cycle dynamics. In the Reserve Bank’s pursuit of price stability, changes in the policy rate influence household spending in part via their impact on asset prices and wealth.

Recent trends have prompted a reassessment of the relationship between household spending and wealth. Since the global financial crisis (GFC), per capita consumption growth in New Zealand has been modest even as housing wealth has been rising rapidly. The modest growth in per capita consumption has coincided with an increase in the household saving rate, a slowing in credit growth, and a positive injection of equity into the housing sector by households.

This Analytical Note estimates the response of household consumption (its elasticity) with respect to housing and financial wealth, and income between 1982 and 2016. To examine the stability of the relationship, the model is also re-estimated across a split sample. The results suggest that there has been a change in the relationship between household consumption and wealth. It finds:

  • On a dollar for dollar basis, consumption appears to respond more to changes in financial wealth than fluctuations in housing wealth over the full sample;
  • The response of consumption to wealth has fallen in the sample after 2005, particularly the response with respect to housing wealth.

The change in the response of household consumption to wealth could reflect a reassessment of expected future capital gains from each form of wealth following the GFC, particularly given heightened uncertainty in the housing market. It may also reflect a desire by highly indebted households to pay down debt, along with the influence of low interest rates, indicating a prolonged but transitory period of deleveraging and higher savings. Disentangling the aforementioned factors is beyond the scope of this note but could usefully be addressed with cross-sectional data at the household level.