Funny Data

[Previous] [Contents] [Next]

Chris Sarlo

In the course of studying poverty and living standards in Canada, I have had the opportunity to examine, fairly carefully, the complete income and expenditure databases that Statistics Canada provides. While most of the information conforms to economists' long held understanding about patterns of income and spending, there is clearly some funny data.

The main theory explaining aggregate income and consumption behaviour is "the life cycle hypothesis." It suggests that people's spending is geared, more or less, to their disposable income, but that people generally take a longer view than the current period to determine their disposable income. Thus, young people typically spend more than their current disposable income as they invest in their education and acquire commodities (cars, appliances, homes). They often go into debt to finance their standard of living. This is because they expect that their future income will be sufficient to both pay off their debt and finance a comparable living standard at that time. At some point, people move into the second phase of the lifetime pattern. They begin to save, having already bought many of the items that define their standard of living and having, usually, a higher income than when they were young. The final phase is retirement. Here, people use past savings to finance their typically more modest living standard. This process is referred to as dissaving.

For most people, the gap between their income and their consumption is not huge. In 1996, on average, consumption was about 83 percent of after-tax income and about 66 percent of total gross income. For 80 percent of households, gross income exceeded consumption. For the roughly 2 million households in which consumption was greater than income, fully 72 percent financed extra spending, at least in part, by reducing their net worth (that is, by dissaving). But these are the overall aggregates. Let's get to the funny data.

In 1996, according to the Family Expenditure Survey, almost 44,000 households had a pre-tax income of less than $15,000, yet enjoyed consumption levels in excess of $30,000. In other words, these 44,000 households were poor, using income as an indicator, but were decidedly middle-class in their consumption.

Conversely, other households appeared to be anomalous in the opposite way. How low can consumption go? Unlike income, which can be zero or even negative (due to business losses), consumption cannot fall below some reasonable level. People have to eat; they need shelter and other necessities. Even if they receive some subsidies or in-kind benefits, it would be impossible to imagine a household's consumption falling below, say, $5,000 a year. Yet, in 1996, there were over 14,000 households with consumption below $5,000. In the category of "can't be true," there were more than 1,200 households with negative consumption. While there may be some technical explanation, low, and particularly negative, levels of consumption defy common sense and force us to question the reliability of the entire database.

What are we to make, for example, of the more than 1,000 households who in 1996 had pre-tax incomes of about $800, consumption of over $80,000, and lived in homes worth more than $800,000? According to their income, these folks are desperately poor, living far below anyone's income poverty line. But they are not really poor, are they? Similarly, in 1996, there were just over 1,000 households with million dollar homes who had incomes below the Statistics Canada low-income cut-off (which is incorrectly referred to as a poverty line by many in the social welfare community). Their consumption, however, was solidly middle class.

Even more bizarre, there were another 3,400 households in 1996 who had average reported incomes of about $3,000, average reported consumptions of $97,000, who lived in homes valued at over $320,000, and which consisted of an average of five people. Curiously, these households did not dissave at all but, rather, actually increased their net worth over the year. Again - and this needs to be emphasized - any income-based measure of poverty would classify these households as profoundly destitute. Yet they are clearly not!

When we sum up all the strange cases in which apparently low incomes wildly misrepresent the household's true economic status, we are not talking about a small number of people. Taking just one aggregate, in 1996, according to Statscan's FAMEX, there were almost 22,000 households comprising over 91,000 persons with average pre-tax incomes of $8,618, average consumption of $58,676, and average home values equal to $439,059. Further, there were 364,000 people living in households with below-poverty level incomes (using the Basic Needs poverty line), but having consumption levels above the Canadian average. If we consider that, according to the basic needs poverty approach, there were about 2 million Canadians overall living in households with reported incomes below the poverty line, the overestimate of poverty here is clearly significant. And this assumes that consumption is accurately reported.

In order to measure poverty with reasonable accuracy, we need accurate and reliable data on how well people (and especially people at the low end of the scale) are living. We clearly do not have that now. It seems to me that Statistics Canada has an obligation to provide a truer assessment of the living standards of low income Canadians. Consumption data helps in this regard, partly for its own sake, and partly because it assists researchers in testing the reliability of income data. However, there needs to be adjustments made to incomes to better reflect how Canadians live - adjustments taking into account such things as in-kind and cash gifts, wealth, unreported income, and part-year families. Without this, we simply cannot have any confidence in our estimates of poverty.

[Previous] [Contents] [Next]

info@fraserinstitute.ca You can contact us at the above email address for any comments or information requests. Please report any dead links or technical problems.

If you know someone who would be interested in this web page, please enter their email address below, and we will forward this URL to them: Email Address:

Last Modified: Wednesday June 9, 1999.