If Canada wants to help the poor, it needs a new way to measure poverty

This article was first published on June 6, 2018 in the Globe and Mail.
Written by Michael Wolfson.

Poverty is hard to measure. There are many aspects besides living on low income, including having disabilities or costly health problems, not being able to find decent housing, not being able to understand and communicate in an environment with increasing technological and legal complexity and being unable to find nutritious food at reasonable prices.

Still, the federal government has embarked on formulating a major poverty-reduction strategy and it would presumably like to have meaningful ways of measuring and monitoring progress toward the goal of reducing poverty in Canada – what Social Development Minister Jean-Yves Duclos has called the 3Ms.

It’s time the federal government established an official poverty line – a dollar amount of income below which a person or family would be deemed to be “poor.”

Many Canadians may be surprised Canada doesn’t have one already. Statistics Canada, while it has been producing various low-income lines since the 1960s, has steadfastly refused to call these “poverty” lines, for the simple reason that poverty involves more than living with low income. This is a reasonable position for a national statistical agency because there is no purely statistical method for deriving a poverty line (nor a low-income line for that matter). However, it is entirely reasonable for a government to specify a poverty line, even it embodies arbitrary judgments – that is well within the purview of elected representatives.

Having a clearly defined poverty line enables a government to set targets, such as reducing the prevalence of measured poverty by 50 per cent over the next decade. This kind of target enables a government to focus its policy agenda on the interventions and program changes most likely to be effective in meeting the target. It further allows Canadians to monitor the government’s progress toward meeting its target.

A critical challenge is whether it is possible to construct a meaningful measure of poverty. A poverty line defined by an income level should only be a starting point because it needs to be complemented by a dashboard of other measures in domains such as access to assets and social support, disability, literacy, housing and food insecurity.

What kind of income line would best fit the requirements for a meaningful official poverty line? At present, Statistics Canada produces three sets of lines. All of them are flawed.

The oldest is the “low-income cut-off” (LICO). It is based on arcane statistical estimation and has not been revised since 1992 since it can bounce around in ways that defy logic. The LICOs should be completely abandoned.

Starting in 1992, Statistics Canada began publishing a “low-income measure” (LIM). The LIM is very simple and transparent. It is anchored at half the median family income, after taking account of family size. The LIMs vary for families of different sizes according to an “equivalence scale,” which is the same as is widely used internationally.

More recently, starting in 2000, at the request of social-affairs ministers from across Canada, Statistics Canada began publishing a “market based measure” or MBM. While the LIM fails to make a distinction along an urban-rural spectrum, the MBM has different lines for each of 50 municipalities and geographic regions. This regional variation is important to reflect differences in costs across the country.

The MBM uses the same equivalence scale as the LIM. However, the MBM uses a complex mixture of items to make up its market basket. The mixture is not at all transparent and when one examines the nearly 100-page document describing the most recent 2010 detailed revisions, many will find the specific items arbitrary.

As a result, both the LIM and the MBM have strong and weak points. Fortunately, it is easy to imagine a “new Canadian poverty line” (CPL) that would combine the advantages of both, while avoiding their major limitations.

Like the LIM, this line would be anchored at half the median (family size-adjusted) income. But this median income would be averaged over the past few years to provide more stability. And like the MBM, it would be differentiated by the same 50 geographic regions – though instead of the complex and highly arbitrary basket of expenditures underlying the MBM, only shelter costs (rents on two-and three-bedroom apartments) would be used to reflect geographic variations.

While not perfect, this CPL would be a sufficient start for the government’s poverty-reduction strategy: It is meaningful, it is practical for Statistics Canada to measure and it would provide a solid basis for monitoring progress toward the fundamental objective of reducing poverty.