Income Research Paper Series – Research Paper
Low income Measurement in Canada: What do different Lines and Indexes tell us?
- Main page
- A comparison of LICO, LIMs and MBM
- Low income indexes under alternative lines
- Who fall between the lines?
- Who contributes more to overall low income? A decomposition analysis
- Summary and conclusions
- Tables and figures
- Appendix 1 Methodology
- More information
- PDF version
A comparison of LICO, LIMs and MBM
Research at Statistics Canada led to the establishment of the Low Income Cut-Offs (LICO) in the 1960s and the variable Low Income Measures (LIM) in the early 1990s. The Market Basket Measure (MBM) line was developed in the late 1990s by Human Resource and Skill Development Canada in consultation with a Federal-Provincial-Territorial working group of officials on social development research and information. The methodologies of these low income lines are well known and a brief description is contained in the appendix.
The fixed LIM line has not been implemented in Canada. It was recommended by a popular report on social indicators for the EU (Atkinson et al. 2002). The line can be easily constructed, and practices in several European countries suggest that it is a useful measuring rod for low income and poverty (Corak, 2005). The methodology of the fixed LIM is similar to that of the variable LIM. But it has several unique characteristics: (1) in the base year selected, the thresholds of the fixed LIM line are identical to that of the variable LIM line; (2) outside of the base year, the thresholds of the fixed LIM are obtained by updating the base year thresholds with consumer price index; and (3) the base year is re-set periodically, say, every five or ten years.
The fixed LIM line can be a useful complement to the existing lines. If we view the variable LIM as a purely "relative" low income line reflecting a contemporary standard of living, then the fixed LIM represents an "anchored" standard of living. When we examine low income statistics with the fixed LIM line, we actually compare the well-being of individuals relative to the median of an anchored income distribution. This is different from the variable LIM line which is based on a "floating" income distribution. It is also different from MBM and LICO. The former represents the contemporary costs of a fixed basket of goods and services, while the later is based on a given pattern of spending on food, shelter and clothing of an average family.
Table 2 compares the characteristics of the four low income lines, largely from an operational point of view. First, the low income lines differ in terms of the complexity of the methodology which affects the transparency of the measure and its ease of communication. In this respect, MBM, variable and fixed LIMs have advantages over LICO. With MBM, one specifies the basket of goods and services that are deemed essential to maintain physical health and to reasonably participate in community activities. The costs of this basket are then calculated for different communities. With variable and fixed LIMs, one defines a family as being in low income if its adjusted income is below half of the population median income. But with LICO, one has to estimate a spending model and derive the income needed for a given level of spending on food, shelter and clothing.
Second, the low income lines also differ in terms of international comparability. LICO and MBM are designed as Canadian-specific lines and hence are not comparable with those of other countries. But variable LIM and fixed LIM are comparable with those used in other developed countries. Since the LIM lines are strictly based on income distribution and family composition, we can compare low income in Canada with that in any country in which an income survey is conducted, and for which a LIM-type line can be easily derived, if that country does not have one already.
Third, the low income lines differ in terms of the underlying assumptions and choices. While it is necessary to make assumptions and choices in creating any low income line, there are differences in making implicit and explicit assumptions. More specifically, for MBM, variable LIM and fixed LIM, virtually all assumptions and choices are explicit, but many implicit assumptions and choices are associated with LICO such as, (a) the propensities to consume (food, shelter and clothing) are assumed to be the same between families in the bottom and those in the top of the income distribution; (b) individual's age, health, labour force status and so on, have no effect on a family's spending on food, shelter and clothing; (c) the estimated effects of family income, family size, and community size are all significantly different from 0; and they are the same across different regions.
Fourth, the costs to produce and maintain these lines differ significantly. MBM is the most expensive one to produce, while LICO, variable LIM and fixed LIM can be produced and updated with little cost. To derive the MBM thresholds, extensive price data for different communities have to be collected and processed and the costs of data collection as well as the periodical basket re-design can be quite high. But to produce the LICO thresholds, one uses data from a household spending survey to fit the spending model and calculate the thresholds using the estimates, no separate data collection is necessary.1 Similarly, the variable LIM thresholds can be easily obtained from any regular income survey. The fixed LIM thresholds are based on the variable LIM thresholds of a chosen year in combination with the consumer price index (CPI).
For example, to construct the fixed LIM thresholds based on the 1992 income distribution for 2007, we need the 1992 variable LIM thresholds, which can be obtained through the Income Research Paper Series of Statistics Canada (top panel of Table 1).2 These are then adjusted by the CPI indexes of 1992 and 2007 to obtain the 1992-based fixed LIM thresholds for the year 2007 (bottom panel of Table 1).
Fifth, the lines differ in the degree to which they reflect regional variations in the cost of living. By design, MBM reflects different regional costs of living through the use of regional or community specific thresholds. While LICO reflects different regional costs of living between rural and urban areas and between urban areas of different sizes, it does not take into consideration the provincial difference in costs of living. In contrast, variable and fixed LIM thresholds only capture the difference in costs of living between families of different sizes. They do not account any differences in the costs of living in different communities.
Finally, the rebasing frequencies are different. By definition, there is a clear updating rule for the variable LIM: it is rebased every year. But for the other low income lines, there are no agreed-upon rules and updating has occurred sporadically in the past. The first set of LICO thresholds was based on 1959 spending. They were rebased subsequently with data from 1969, 1978 and 1986. The current LICO thresholds are based on 1992 data. The MBM thresholds are based on a 1997 basket, and are currently under revision with a new basket.
In this study, we investigated the behaviour of the 1992 fixed LIM together with those of LICO, variable LIM and MBM. We chose 1992 to anchor the LIM thresholds in order to have the same base year with the current LICO thresholds.3
- However, from 2010 on, Survey of Household Spending (SHS) will become a monthly survey, implying that to produce a new set of LICO thresholds using the same methodology, a special annual spending survey is needed.
- Statistics Canada (2004).
- We also tested the 1982 and 2002 fixed LIM lines and found that the choice of base year does not affect low income trend. Results based on the 1982 and 2002 fixed LIM thresholds are available from the author upon request.
- Date modified: