Poor Thinking: Neil Reynolds on Measuring Poverty

Neil Reynolds is at it again in today’s Report on Business (not available on line), defining poverty out of existence by questioning the reliability of standard statistical measures.

His main point in a somewhat confused argument is that poverty rates are over-stated by conventional income-based measures such as the LICO.

Reynolds argues that the poor – according to consumption surveys – spend more than their reported incomes in the same year.  He suggests that this situation arises because the poor have hidden income and are somehow not ‘really’ poor.

However, other plausible explanations include the exhaustion of limited financial assets and running up debts during periods of low income. Anybody familiar with the realities of poverty knows that many families are, by force of social assistance rules, required to liquidate almost all liquid assets before collecting social assistance, and often borrow from others, or at very high rates of interest. The fact that the poor may consume at somewhat above their (low) level of current income over specific periods of time is hardly cause for an argument that they are not poor – especially in periods where poor families have no income at all.   Reynolds also makes no mention of the fact that most poor families have incomes well below the LICO lines.
His data on the consumption of the poor are questionable. Reynolds is probably unaware of the fact that Statistics Canada’s  main surveys of consumption patterns are not very reliable, particularly when it comes to measuring the consumption of the very poor.  Household surveys (formerly the SCF and the SLID) have been shown by Statistics Canada to produce significantly lower estimates of the incidence of low income than Census and tax data, likely  because of under-sampling at the low end of the income distribution. (See Marc Frenette, David Green and Garnett Picot “Rising Income Inequality in the 1990s” in David Green and Jon Kesselman (Eds) Dimensions of Inequality in Canada, UBC Press, 2006.)

Reynolds’ related line of argument is that poverty is best measured in terms of consumption rather than in terms of income. He apparently thinks that the former is absolute rather than relative,  more firmly anchored in the cost of buying necessities. In point of fact, consumption-based meassures of poverty are still relative, measuring what poor households can buy compared to some community norm. This is actually acknowledged quite explicitly by Chris Sarlo of the Fraser Institute, very much in the spirit of Adam Smith who argued that to be poor was to fall short of a community norm rather than an bare-bones survival income. Of course, Sarlo’s basic basket is extremely meagre, but it is still explicitly relative.

(For support for this argument see this long note on defining poverty by the CCSD : http://www.ccsd.ca/pubs/2001/povertypp.htm)

Reynolds seems unaware of the fact that  relentless pressure from the right led the federal and provincial governments to develop a consumption based measure of poverty, the so-called Market Basket Measure. This was intended to calculate what income would be needed to purchase the basics of life for representative families in different communities, apparently in the hope of lowering poverty rates by statistical fiat.

While the MBM poverty rates  generated by this exercise differed from the after tax LICO rates in terms of detail, they were not really all that different in terms of the overall dollar amounts which define poverty thresholds. That’s probably why we have barely heard of the MBM since.

In summary – shifting the focus from income to consumption does not fundamentally throw into question the extent of poverty in Canada.