Revisions to 2006 to 2011 income data
Introduction
Archived Content
Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.
Income statistics provide valuable information on the economic well-being of Canadians. These statistics are used by all levels of government as well as by non-government organizations to develop income support programs and social services, by academics and research organizations interested in labour market and social policy issues, as well as by the general public.
Income statistics are more informative when comparisons can be made over time. For example, there is a need to monitor year to year developments in household income to observe the effect of a change in the business cycle, or the introduction of a new government program. There is also the need to study longer-term trends in income, for example, to better understand differences in regional growth patterns, low income or income inequality.
Statistics Canada provides a range of income statistics that are produced using both administrative data sources such as the T1 Family File (T1FF)Note 1 and survey-based sources such as the Canadian Income Survey (CIS). While administrative data provide detailed income estimates for families and individuals, survey data provide richer information at the household level that more fully reflect the economic circumstances faced by Canadians.
Over time Statistics Canada has made methodological changes to the way in which annual survey-based income data are collected and processed. However, efforts are made to ensure that income statistics produced by these data remain as comparable as possible from year to year. The most recent change was introduced in the 2012 reference year when income statistics were first reported by the CIS. The CIS reports on many of the same statistics as the Survey of Labour and Income Dynamics (SLID), which last reported on income for the 2011 reference yearNote 2.
Statistics Canada has advised users that as a result of methodological differences between CIS and SLID, statistics produced by the two surveys are not comparable. These differences were described in “Note to Users of Data from the 2012 Canadian Income Survey” published with the initial CIS release in December 2014Note 3.
An important difference between the two surveys is in their design; SLID was a longitudinal survey in which the same respondents were interviewed each year for a six year period, while CIS is a cross-sectional survey where respondents are only interviewed once. SLID estimates can differ from those of CIS as a result of coverage and response differences. Coverage issues include an undercoverage of recent immigrants in SLID, as new immigrants to Canada were only added to SLID when a fresh panel was introduced. Response differences include the effects of sample attrition over the length of the SLID panel. Sample attrition refers to the fact that, in a longitudinal survey, fewer and fewer members of the original sample are interviewed each year due to refusal to continue participating, or inability to find respondents following a move. As a cross-sectional survey, neither of these issues are present in CIS.
To ensure continued comparability of income statistics over time, Statistics Canada released revised historical estimates for 2006 to 2011 that allow for the comparison of CIS data to earlier years. This note presents these revised estimates and discusses how they compare to pre-revision results, as well as to estimates from CIS for 2012 and 2013. The main conclusions are:
- Revisions to 2006-2011 estimates make it possible to compare results from the CIS to earlier years.
- Revisions to median income values tend to be small.
- Revisions to 2006-2011 increase low-income rates during those years. The upward revisions in low-income estimates are largest in Toronto, and, correspondingly Ontario, and were also concentrated in other large Census Metropolitan Areas (CMAs).
In revising estimates for 2006 to 2011, the objective was to make SLID estimates as comparable as possible to the new CIS data for 2012 and onward. Nonetheless, for some characteristics the data trends could still reveal a “break” due to the change in methodology. Such a break would appear as a noticeable upward or downward shift in the data between 2011 and 2012. It represents a change in the data which is attributable to the two surveys having differences which could not be fully adjusted for in this revision.
This note also provides advice to users regarding the comparison of income estimates for years before 2006 with those from 2006 and after. Users should also be aware of potential breaks when comparing 2006 and later years to income estimates for years before 2006, as pre-2006 estimates have not been revised. Such a break would likely appear as a noticeable upward or downward shift in the data between 2005 and 2006. However, income estimates before 2006 are suitable for analyzing long term, cyclical trends. Analysis suggests that similar conclusions regarding long term or cyclical trends would be drawn using the unrevised and revised series.
- Date modified: