© 1994 Karen Selick
 Mired in Poverty?  No, Mired in Poor Stats
An edited version of this article first appeared in the December, 1994 issue of Canadian Lawyer.  If you wish to reproduce this article, click here for copyright info.


 Mired in Poverty?  No, Mired in Poor Stats

If you approached a dozen people at random on the street and asked them what characterizes living in poverty, most of them would probably say it has something to do with not being able to afford the basic necessities of life: food, shelter and clothing.

Strangely, this is not the way most politicians, the media, and poverty lobbyists approach the issue.  They rely instead on "poverty lines" calculated by Statistics Canada or the Canadian Council on Social Development (CCSD), even though those organizations themselves have said that their low-income lines should not be construed as such.

These unofficial poverty lines make no attempt to measure how many people cannot afford to buy the necessities.  On the contrary, Statistics Canada's low-income lines measure the number of families who have to spend more than 56 percent of their income on food, shelter and clothing (which leaves as much as 44 percent for non-necessities--also known in some circles as luxuries).  Similarly, CCSD's line is simply fixed at an arbitrary 50 percent of the average family's income (adjusted for family size), which produced the implausible poverty rate of over 25 percent in 1986.

Defining poverty in these ways produces some intriguing results.  These were pointed out by economics professor Christopher Sarlo in his 1992 book Poverty in Canada and his 1994 update, recently published by Vancouver's Fraser Institute. Here is a sampling of some of the anomalies.

Using the CCSD's methods, even if every family's income doubled, tripled or quadrupled in real terms over a period of years, it still wouldn't make a dent in the 25 percent poverty rate.  By using relative rather than absolute standards, this method of measurement ensures that there will always be a "poverty problem."  Someone whose income is a princely $1 million would still be defined as poor if the average income were $2 million.

Furthermore, if this method were applied in Third World countries, many of them might well show lower rates of poverty than Canada.  Says Sarlo: "...recent reports from Cuba suggest that the typical diet in 1993 is virtually at subsistence level, with less than half the recommended caloric intake.  If these reports are true, then more than half (probably closer to 75 percent to 80 percent) of Cubans are currently poor using a basic needs approach.  However, if most incomes in Cuba are clustered around a low, inadequate average, then a relative approach would reveal almost no poverty in Cuba."

Another curiosity is this:  if you were to index Canada's average family income for the year 1951 according to increases in the consumer price index, and then apply current Statscan or CCSD poverty lines, you would learn that in 1951, the average Canadian family was what we would now be required to label "poor."  This would come as a great surprise to everyone I know who can remember back to 1951.

My favourite part of the book, though, is the appendix.  Here, Sarlo tackles the perennial "quintile problem."  Most readers have undoubtedly seen those troubling statistics showing, for example, that in 1988, the bottom one-fifth of Canadian households received a scant 4.6 percent of the country's total household income, while the top one-fifth received a whopping 43.2 percent.  

The first noteworthy fact is the remarkable stability of these percentages over time.  In 1951, these figures were 4.4 and 42.8 percent respectively--almost identical to what they were in 1988.  Yet those 37 intervening years wrought enormous changes in Canada's social welfare legislation.  Massive transfer payments have been syphoned out of middle- and upper-income households and directed into programs designed to raise the lowest quintile.  Why don't the figures reflect this?

Sarlo points out that incomes vary drastically according to age, even among people in the same social class.  A high school graduate starting his first job will earn much less than his 50-year-old father, who is in his peak earning years, but perhaps more than his 75-year grandfather who is retired on a pension.  Each generation moves through various quintiles at different stages of their lives. 

He constructs a simulation model which assumes that every individual in society follows an identical life pattern.  Everyone does everything at the same age: starts working, marries, has children, retires and dies.   Everyone gets the same starting salary, the same annual increases over their working career, and the same pension.  In other words, everyone earns the same lifetime income.  The only difference among people is the year of their birth, which is staggered evenly among the population.

It's difficult to imagine an income structure more uniform than this.  Yet when this fictitious population is divided into income quintiles by household, the lowest 20 percent still receive only 9.21 percent of total income, while the highest 20 percent receive 31.57 percent.

When the model is modified to include two categories of employment (ordinary workers who start working at age 21 and professionals who start at age 26 with higher salaries), the bottom quintile ends up with 7.19 percent, and the top quintile with 38.38 percent.  

This is not far removed from the real world results, yet it takes no account of individuals' varying tastes for work, leisure, responsibility and so on.  Adjusting for those factors might well demonstrate that there is nothing inherently unfair about the quintile figures.

 - END -


..... ..... 


June 07, 2000