Categories
Governance Poverty Transparency

Inaccurate measures of poverty: A brief Canadian history

Low income cut-off (LICO)

For several decades, Statistics Canada used LICO as its primary measure of relative poverty in Canada.  Although it was not formally referred to as the ‘poverty line’, that is the purpose it served for many.   It was a general measure that took the average family’s expenditure on food, shelter and clothing (adjusted for family and community size), added a margin of 20% to it, and that was it.  If a family’s income fell below its respective LICO threshold, it was counted as low-income.  The LICOs were occasionally updated to account for increases in average family expenditure on the three basics.  What the LICO lacked in precision, it compensated for as a historical index:  LICO could be traced as far back as 1968 and it allowed for the creation of historical maps such as the ones published with the 2006 Census income release.

Low income before tax cut-offs, 2005 

Low income cut-off, after tax (LICO-AT)

The LICO was changed in the early 1990′s.  In 1991, LICOs based on after-tax income were published for the first time.  It seemed reasonable at first glance:  Canada supposedly had a progressive taxation system with tax rates that rose relative to income bracket, and the LICOs should reflect that.  Problem?  The Census (and SLID) income data was the basis for the denominator. The Family Expenditure Survey (FAMEX) and later Survey of Household spending (SHS) would be the basis for the numerator.  The Census (and SLID) income tax data was notoriously unreliable, where it was available at all.  If reliable income tax data was not available, then what were the after-tax LICOs based on?

There is no simple relationship, such as the average amount of taxes payable, to distinguish the two types of cut-offs.  Although both sets of low income cut-offs and rates continue to be available, Statistics Canada prefers the use of the after-tax measure…  The number of people falling below the cut-offs has been consistently lower on an after-tax basis than on a before-tax basis.

Basically, the amount of income tax paid was assumed using the tax brackets and aggregated Canada Revenue Agency T1 data, irrespective of how much actual tax a person or family actually paid.  That’s inaccurate, as many high-income and/or wealthy families could defer, offset or reduce a significant share of their earnings through various registered investment plans, capital gains and other tax shelters.  Lower-income families’ earnings on the other hand largely consist of wages and salaries, from which personal income taxes are usually already deducted.  The lowest personal income tax rate is higher than the effective capital gains tax rate on the highest personal income tax bracket in Canada. Even if they do qualify for tax rebates, lower-income families are less likely to apply for or receive them as they  don’t  generally benefit from professional tax preparation services. Some families with little income do not file returns at all.

While it was no more accurate than it’s newly-dubbed ‘before-tax’ counterpart, the after-tax LICO did have one notable effect: by lowering the cut-offs across the board,  it also lowered the number of Canadians falling below them, effectively reducing low-income incidence.

Low income after-tax cut-offs, 2005

Low-income measure (LIM)

Also introduced in 1991, the LIM is a fixed percentage (50%) of median adjusted family income based on family size.  Canadians would only be considered low-income if their total family income was less than half the median same-size family’s income.  Community size was not taken into account .

The LIMs are based on the Survey of Labour and Income Dynamic (SLID), a voluntary survey of a small number of households sampled from the LFS (in turn sampled from the Census).

Oh, and in many cases the more general LIMs also happened to further lower the income thresholds, effectively reducing low-income incidence.

LIM after-tax thresholds, 2006

MBM 

Introduced in 2000, unlike the LICO and LIM, which were relative measures of low-income based on average family income and/or  consumption, the MBM measure was normative:

The MBM and the MBM disposable income were designed by a working group of Federal, Provincial and Territorial officials, led by Human Resources Development Canada (HRSDC) between 1997 and 1999. 

Bureaucrats got together to figure out what/how poor people should live and created a basket of goods based on these assumptions, which, not surprisingly, were rather questionable.  Probably the most questionable of these was:

The MBM thresholds are calculated as the cost of purchasing the following items: …Transportation costs, using public transit where available or costs associated with owning and operating a modest vehicle where public transit is not available.

A low-income family of four (two adults,  two children) living in a major urban area  (CA/CMA) where transit is available gets a couple of transit passes added to its basket for transportation costs.  Because the working poor in areas where transit is available presumably needn’t drive – not the Molly Maids, the delivery drivers, the general labour contractors, etc.  This assumption was plainly wrong and was easily demonstrated to be so by a simple calculation (using Q47 of the 2006 20% sample Census).  Nevertheless, it remained.

In addition to being an inaccurate measure, in many cases the MBMs further lowered income thresholds, effectively reducing low-income incidence.

MBM thresholds for reference family, 2007

‘Progress’

The historical ‘progress’ of low-income measurement in Canada can be demonstrated with an example.  A family of four (two adults, two children) living in Montréal:

LICO-BT (2005)       38,610
LICO-AT (2005)       32,556
LIM-AT (2006)         30,358
MBM (2007)            26,560

With each subsequent new measure, the income cut-off was lowered, cut by nearly a third overall for that family of four living in Montréal. It seems ‘progress’ in Canadian poverty reduction is simply a matter of changing metrics.

Leave a Reply

Your email address will not be published. Required fields are marked *