Table 2: Income Inequality
The table is intended to allow the following analysis:
- A ratio of incomes at different percentile and decile thresholds (Deciles 1 through 10, 99th percentile top 1% e.g. top 10% to bottom 10%, etc…); and,
- A comparison of the value of the share of total after tax income received by the population above or below any given decile or percentile threshold (e.g. the top 1 percentile earned x% of all income, while the bottom 4 deciles earned y%).
Note that the deciles for this indicator are created using adjusted after-tax family income, in which a family's total income is divided by the square root of the family size and assigned to every member of the family, including children. The deciles are calculated across the population of individuals, who are all assigned this adjusted family income.
Procedures to build the income inequality table:
- Exclude persons in Yukon, Northwest Territories, Nunavut, non-residents and Indian Reserves and dead filers in the year of death
- Calculate income decile groups for each individual within each geography as follows:
- Assign the adult equivalized Census Family after tax income to every person. This income, which is assigned to all family members, including children, is calculated by dividing family income by the square root of family size. This is the same method for calculating the income deciles used in the NHS profiles and for calculating the Low income measure.
- Determine the decile thresholds for each geography across all individuals.
- Group individuals into decile groups using these thresholds and their census family income.
If you are planning on using the Census Subdivision level of geography, please read the 'T1FF_CSD_Quality_Note.docx' document.
Place Names are included in the data tables.
Note that "Rural Postal Code (not in city)" and "Rural Community" are interchangeable.
The working poor and the income inequality tables exclude populations on Indian reserves and in the 3 northern territories, since both the income and the expenditure patterns can be quite different for the population in those areas.
Geography notes
This table is available for Census Subdivision geographies. Due to the relatively poor quality of the coverage for some CSDs (mostly in rural areas and low population areas), T1FF data at the CSD level should be used with caution. Refer to the T1FF/population count comparison table to determine the quality of coverage.
Statistics Canada does not accept responsibility for any direct or indirect consequences associated with under coverage or over coverage in T1FF data at the CSD level.
The document "T1FF_CSD_Quality_Note.pdf" describes the methodology used to link postal geography and census geography.
The tables also include FSAs (true complete FSA, both urban and rural, not split by postal cities; listed as Geo level 03) and all rural postal codes (listed as Geo level 06). Geo Level 3 also includes residual counts listed as “Other Provincial Total”. This is where FSAs with less the 100 tax filers are grouped together at the province level. Since FSAs almost always have more than 100 tax filers, some provinces will not have an “Other Provincial Total”. The assignment of Geo Level 03 and 06 can potentially lead to confusion since these are slightly different than same levels of geo in the other tables produced for the client and in standard tables. In the Working Poor and Income Inequality tables the Geo Level 03 is for true FSAs both urban in rural, while in the standard tables Geo Level 03 is only for partial urban FSAs (postal codes not associated with regular residential mail delivery are excluded) and these are sometime split by postal cities. As for rural postal code listed as Geo Level 6 in the Working Poor and Income Inequality tables, this is actually a combination of the Geo Level 06 (rural postal codes in a postal city) and Level 09 (rural postal codes outside a postal city).
Suppression
Users have noted that some cells in the Income Inequality tables are suppressed. Below are the reasons given by Statistics Canada for suppression in these tables:
Issue 1: Counts of observation per cell.
As a basic suppression rule, and data cell that represents less than 15 individuals gests suppressed in our custom or standard T1FF tables. This means that if you are creating deciles that counting individuals, you need at least just over 150 indicial in that area if you don’t want all the deciles to be suppressed (150 pop divided by 10 = 15 individuals hence all the deciles have to be suppressed.
Issue 2: Programming and allocation of individuals and income values for areas just around 150 individuals.
Let’s assume that we have an area where the unrounded counts is 157 individuals (the rounded count would be 160). When you allocate this by deciles, what the programming does (without getting into too much of the technical details) is the equivalent of assigning 16 individuals to some deciles while others only get 15 individuals. The ones for which only 15 individuals were assigned will be suppressed, while the other that have been assign 16 individuals (meets the minimum rule described in issue 1) will not be suppressed and be rounded to 20 individuals.
Issue 3: Dominance for top and bottom deciles.
Our suppression rules for dominance dictates that if one value within a data cell (extremely low or extremely high) has too much of an impact as compared to the other values within that cell, then that cell is suppressed. A good example of this would be if we have one extremely rich individual in a small rural region. For example, if you have 19 individuals the top decile and the individuals in that deciles have incomes around $40,000, but the 20th individual has an income of a few million dollars, then the cell gets suppressed. This is done in order to protect the confidentiality. Someone familiar with the area who knows that only one individual there truly has a high income would be able to breach the confidentially that we guarantee to our survey respondents or, in the case T1FF, tax filers.