Low Income in Canada: 2000-2006 Using the Market Basket Measure - October 2008

4. The Results

Comparisons of Low Income Incidence, Depth, and Persistence for the Period 2000 to 2006 Using the MBM and the LICOs-IAT

In examining low income, several dimensions are important. This report will deal with incidence (the share of people living in low income in a given year); depth (how far below the low-income cut-offs low income families fall) and persistence (whether persons who experience low income during the 2000-2006 period do so for most of that period or for only one or two years).

4.1 Incidence: 2000 to 2006

Between 2000 and 2006, as measured by the MBM, the number of persons living in low income declined by 572,000 to reach 3.8 million in 2006 (Chart 1). As a share of the population, the rate of low income fell from 14.6% in 2000 to 11.9% in 2006. This gradual decline, however, conceals large annual movements into and out of low income. Each year during this period about 1 million people entered low income, while about 1.1 million exited low income.

As can be seen in Chart 2, this trend in the incidence was similar using the LICOs-IAT. Except for the decline between 2000 and 2001, none of these year-to-year changes was statistically significant.¹⁰ This pattern was experienced by each of the three main age groups (<18, 18-64, and 65+) using both measures. (Table 7).

The gradual downward trend over the 2000 to 2006 period for both the MBM and the LICOs-IAT reflects an improving labour market over this period, significant increases in benefits for both children and seniors, and significant reductions in income taxes.

As can be seen in Chart 2, in 2006 the percentage of persons in the 10 provinces living in low income based on the MBM was 11.9%, compared to 10.5% using the LICOs-IAT. This difference in the overall incidence of low income was accounted for by the more stringent definition of MBM disposable income. If the same disposable income definition used for the LICOs-IAT had been compared to the MBM low-income thresholds in 2006, the incidence of low income would have been 9.8% rather than 11.9%.

There was a somewhat wider gap in the incidence of low income for children under age 18 using the two measures (Table 7). A key reason is because the MBM accounts for child care costs, whereas the LICO does not.

The situation was reversed for persons 65 and over, where the incidence measured by the MBM was smaller. This was despite the fact that out-of-pocket medical expenses, which tend to be higher for households where the MIR is 65 or over, are subtracted from disposable income before comparing it to the MBM thresholds.

This upward pressure from uninsured health costs on the incidence of low income for the elderly using the MBM is more than offset by the impact of the Low Income Measure equivalence scale used for this measure. Except for those living in rural and small urban communities, the Low Income Measure equivalence scale calculates that a smaller fraction of the disposable income of a family of four is needed by one and two-person families to have a similar standard of living than do the implicit equivalence scales of the LICOs-IAT.¹¹

Since most elderly persons live in one- and two-person families in medium to large urban communities, their low-income thresholds and, consequently, their low-income rates tend to be lower than those of the reference family using the MBM than the LICOs-IAT.

The significant overall decline in low-income rates for persons 65 and over reflects the fact that persons reaching age 65 after 2000 were more likely to qualify for employer-sponsored pension plan benefits and/or to live in couples where both partners had earnings prior to retirement (and consequently two retirement benefit cheques from the Canada/Quebec Pension Plans following retirement). The decline between 2005 and 2006 also reflects an increase in the Guaranteed Income Supplement for low-income, unattached seniors which took effect in 2006.

The MBM thresholds are more sensitive to geographical differences in the costs of shelter and transportation than the LICOs-IAT. A key result of this greater sensitivity is that within each province the differences between the low-income thresholds for larger urban communities and for rural areas and urban communities too small to be served by public transit systems are smaller than using the LICOs-IAT. This is because operating a used car is more costly than purchasing adult public transit passes and taxi rides. Combined with the fact that incomes tend to be lower in rural areas than in large urban centres, this results in significant differences using the two measures in the geographical distribution of the low-income population. In 2006, according to the LICOs-IAT, only 5.3% of Canada's low-income population lived in rural areas, while 60.4% lived in the nine Census Metropolitan Areas (CMAs) with populations of 500,000 or more.¹² However, in the same year, according to the MBM, 11.2% of Canada's low-income population lived in rural areas, while 51.6% lived in the large CMAs.

4.2 Depth: 2000 to 2006

In addition to the percentage of people living in low income, it is also important to know how far below the low-income thresholds their incomes fall. Two populations might have the same incidence of low income. However, if one low income population, on average, has an income just below the low income thresholds while the second, on average, has an income that is only half the low income thresholds, the first population is definitely better off than the second.

For those families with disposable incomes below a low income threshold, the depth of low income is the difference between their disposable income and their low income threshold, expressed as a percentage of that threshold. For example, a depth of low income of 20 means that the person lives in a family whose disposable income is 20% below its low income threshold

In general, the annual changes in depth are slight and not statistically significant; national depth fell a mere 0.8 percentage points between 2000 and 2006. More noteworthy are inter-group comparisons of depth. In 2006, the depth of low income was considerably less both for the low income elderly (19.8%) and for low income children (25.0%) than it was for the working-age population 18 to 64 (34.3%). This reflects the greater generosity of government transfer programs to seniors (such as the Old Age Security Pension, the Guaranteed Income Supplement, and the Canada and Quebec Pension Plans) and to families with children (such as the refundable Canada Child Tax Benefit, the National Child Benefit Supplement, and the Universal Child Care Benefit).

Comparing the MBM to the LICOs-IAT, it appears from Chart 4 and Table 9,that working-age persons and children in low income in 2006 experienced a smaller depth of low income using the MBM than using the LICOs-IAT. However, these differences are also not statistically significant (see footnote 11). There is a wider difference in the measure of depth for seniors because their income distribution is more compressed, making depth among them more sensitive to the differences in the low income thresholds used by the LICO-IAT and the MBM.

It is difficult to determine why year-to-year changes occur in the depth of low income. In any given year, large numbers of persons and economic families move above and below the low income thresholds of any measure. For example, between 2005 and 2006 there was a net reduction of 120,000 persons living in economic families below their MBM thresholds. However, over that same period, 2 million persons either moved into or out of low income economic families.

The change in the depth of low income from year to year is the net result of a combination of influences: 1) how far below the thresholds those moving into low income fell; 2) how close to the thresholds those moving above the thresholds were before they exited low income; 3) whether the disposable incomes of those who remained below the thresholds in both years moved closer to or farther from the thresholds; and 4) the low income status of families added to the sample.

4.3 Persistence: 2000 to 2006

The negative consequences of living in low income, particularly for children, are likely to be more damaging the longer one lives in such circumstances. Thus it is important to know to what extent persons experience persistent, as opposed to short-term, low income.

A person is said to experience persistent low income using the MBM if the total of the annual disposable incomes of the families to which they belonged over the period of years being examined was less than the total of the low income thresholds for those families for those years.

It is important to understand the difference between experiencing low income in a given year during a period of more than one year, experiencing low income every year during the period, and experiencing low income persistently during the period.¹³ Consider, for example, an unattached young woman living alone in 2000 and 2001 who had a disposable income of $12,500 and a low income threshold of $15,000. In 2002 she married and the combined incomes of the two spouses totalled $21,000 compared to their low income threshold of $22,000. In 2003 she became pregnant and stopped earning, but her spouse obtained better-paid work so family disposable income remained at $22,000. In 2004 the baby arrived, so their low income threshold rose to $25,500, but their family income, with her parental benefits, went up to $27,500. So, over the five-year period from 2000 to 2004, this woman lived in families with a total disposable income of $95,500 while the total of the low income thresholds for those families was $99,500. Therefore she is counted as living in persistent low income for the period from 2000 to 2004, although during the last year of the period, the disposable income of her family was above the low income threshold.

Of all persons aged 18 to 60 in 2002, 22.2% experienced low income using the MBM at least one year between 2002 and 2006; while 7.5% (just about one-third of those experiencing low income at least one year) experienced persistent low income over this five-year period. The comparable numbers for the period from 2000 to 2004 were that 21.7% experienced low income at least one year during the period, while 7.9% experienced persistent low income using the MBM during those years. Thus, there was a slight decline in the incidence of persistent low income between the earlier and later period.

Looking at children under age 14 in 2002, 26.5% were in low income at least one year between 2002 and 2006 using the MBM; while 11.0% (under half of those experiencing low income at least one year) experienced persistent low income. The comparable rates for the 2000-2004 period were that 26.2% experienced low income during at least once during these years while 11.6% experienced persistent low income. Again, there was a slight decline in the incidence of persistent low income among children between the earlier and later period.