Household Leverage and Asymmetric Housing Wealth Effects- Evidence from New Zealand
The Global Financial Crisis (GFC) in 2008/09 highlighted the impact that household debt and leverage have on consumption, above and beyond the important role of wealth emphasized by classical theories of consumer behaviour (Fisher, 1930; Modigliani and Brumberg, 1954; Friedman, 1957).
In this paper, we study the role of household indebtedness in transmitting housing wealth shocks to consumption empirically. Housing wealth and mortgage leverage take center stage of the recent literature on consumption partly because housing wealth accounts for the lion’s share of household wealth in most advanced economies and partly because of the prominent role played by house price dynamics and mortgage debt accumulation in the years preceding the Great Recession.
We use microdata from New Zealand Household Economic Survey (HES) to study the housing wealth effect on consumption while controlling for household leverage. House price dynamics in New Zealand represent an interesting contrast to the US: While the US economy experienced a severe housing market downturn after the GFC, house prices and household leverage in New Zealand have been persistently rising over the last decade. Empirical studies focusing on this period for the US show that household leverage amplified the housing wealth effect on consumption through the collateral channel (Kiyotaki and Moore, 1997 and Iacoviello, 2005).
Our main empirical finding is that, on average, the elasticity of the consumption growth to housing price changes is 0.22%. This corresponds to 3 cents out of one dollar in terms of marginal propensity of consume out of housing wealth. In addition, we find that the housing wealth effect is asymmetric with respect to positive and negative housing wealth shocks. The housing wealth elasticity is 0.23 for negative shocks, as compared to 0.13 in response to positive changes in housing wealth. We contribute the asymmetric housing wealth effect to the influence of household indebtedness. The intuition of the finding is that leveraged gains might mostly be used to pay down debt (precautionary saving effect), whereas leveraged losses have a more direct bearing on consumption (collateral effect).
This finding is based on further investigating the role played by household indebtedness in accounting for the asymmetric housing wealth effect. First, we find that, on average, measures of indebtedness negatively impact on the consumption growth. When we separate positive and negative housing wealth shocks, we find evidence showing that the collateral effect is significant only with negative shocks, while the precautionary saving effect is significant only with positive housing wealth shocks.
This empirical finding implies that the precautionary savings effect and the collateral effect work in the different phases of housing cycles. In a housing boom the precautionary savings motive dominates and high leverage weakens the reaction of consumption spending to housing wealth increases. In a housing downturn the collateral channel dominates and high leverage strengthens the reaction of consumption to housing wealth decreases. In another words, the household leverage reinforces the housing wealth effect in a bust, but dampens the housing wealth effect in a boom. The asymmetric transmission mechanism through household indebtedness explains the asymmetric housing wealth effect that we find in the data without controlling leverage explicitly.