Household Spending and Consumer Debt in Canada: Empirical Evidence from the 2005-2009 Surveys of Household Spending

by Anindya Sen


Note: While abstracts are available in both English and French, some research papers are only available in their original language. This document is only available in English.


There have been considerable concerns raised by policymakers on the amount of debt being accumulated by Canadian households. One possible consequence is that an increase in interest rates might adversely affect the ability of households to pay their debt and push them to bankruptcy, thus initiating a return to recession.

In order to assess this possibility, it is important to understand the extent to which households spend more than what they earn and what items form a majority of expenditures. In this regard, there is very limited empirical research. This study attempts to fill this gap by employing the 2005 to 2009 waves of the Surveys of Household Spending.

Based on these data, this paper offers the following stylized facts.

  • A significant portion of the population – between 35% to 39% spend 105% more than what they earn. On the other hand, there has been a drop in such ‘spender households’ – from 39% in 2005 to roughly 35% in 2009.
  • In 2005 the vast majority of spender households - roughly 65% - were those with income less than $50,000. But in 2009, the proportion of such households dropped to 60%. In contrast, the proportion of ‘spender households’ with income between $50,000 and $100,000 increased from roughly 27% to 32%.
  • ‘Spender households’ are also an important constituent of all households within specific income groups. Specifically, a significant portion of all households that earn less than $50,000 – roughly 45%-55% - spend 110% more than what they earn. However, the proportion of more well off households that spend more than 110% of income is non-trivial. The proportion of in debt households earning from $50,000 to $75,000 is either at or a bit less than 30%, while the corresponding figure for households earning from $75,000 to $100,000 ranges from 30% to 25%. The point is that while overspending declines with increases in household income – it still exists at higher income levels.
  • There has been a noticeable increase in average debt levels over time, ranging from $15,000 to $23,000. These amounts are rather sizable given that the vast majority of in debt households earn less than $50,000.
  • While a loss in asset value is correlated with spending more than household income, a significant portion of households that spend more than income experience very little change in asset value over the same time period.
  • Sample means across time demonstrate that households that spend more income than they report – constituting more than 30% of the population - spend a greater portion of their expenditures on mortgage and rent, recreation expenditures and private transport relative to other households. They also allocate a proportionately smaller amount on pension and insurance payments as well as personal taxes relative to households which spend what they earn.
  • Consistent with simple summary statistics, OLS regression estimates confirm that household income is the most significant predictor of whether a household is in debt. Household size is also important – from a statistical perspective. Interestingly, neither marital nor household type dummy variables are consistently significant.
  • OLS estimates are sensitive to household income. The regression results suggest the paradox that less spending on core items is correlated with a higher likelihood of spending more than income. It is possible that an increase in spending on core items might leave less budgetary room for other non-core items and therefore reduce the likelihood of spending exceeding consumption. On the other hand, it is also possible that a preference for non-core items crowds out spending on core-items and increases the probability of being in debt.

Date modified: