Income Research Paper Series
Canadian Income Survey: Population rebasing, 2006 to 2013
by Income Statistics Division
Skip to text
Periodically, income statistics are updated to reflect the most recent population estimates from the Census. Accordingly, with the release of the 2014 data from the Canadian Income Survey, Statistics Canada has revised estimates for 2006 to 2013 using new population totals from the 2011 Census. This paper provides unrevised estimates alongside revised estimates for key income series, indicating where the revisions were significant.
The Canadian Income Survey (CIS) is an important source of Canadian data on family, household and personal income. Starting with the 2012 reference year, annual individual and family income data has been produced by the CIS. The CIS is a cross-sectional survey developed to provide information on the income and income sources of Canadians, along with their individual and household characteristics. CIS reports on many of the same statistics as its predecessor, the Survey of Labour and Income Dynamics (SLID), which last reported on income for the 2011 reference year. Together, the results from these surveys provide the Canadian income estimatesNote 1.
Statistics Canada has long had a process in place by which survey estimates are revised once a new series of postcensal population estimates becomes available. Just as a historical revision of the Labour Force Survey was done in 2015 (Statistics Canada, 2015b), a revision to the Canadian income estimates was also required. Previous to this revision, the Canadian income estimates were based on 2006 Census population counts. With the release of new census data in 2015, all survey data are now based on population estimates from the 2011 Census. As a result of this revision, the total target populationNote 2 for 2013 was revised downward by 0.3%.
With the availability of new Census information, surveys can improve the representativeness of their samples by adjusting the survey weights. Since some of these surveys' population estimates for certain population sub-groups correspond to known population totals for a given year, after reweighting, income estimates are of better quality and the various estimates from the survey and from external sources are more comparable (Tremblay, 2005). To adjust the weights, these surveys use various provincial counts: counts of individuals by age group and gender, counts of economic families and households by size, and counts of individuals by wage class based on administrative tax data (LaRoche, 2007).
In January 2015, the Demography Division within Statistics Canada released the new population, economic family and household estimates derived from the 2011 Census (Statistics Canada, 2016a). Individual counts and the distribution of economic families and households by size were all revised. The total Canadian population fell slightly following the revision, and there were slight changes to the economic family and household sizes. These new population estimates were used to recalculate the survey weights between 2006 and 2013, inclusively.
As part of this revision, geographic boundary changes have also been introduced. Census Metropolitan Areas (CMAs), Economic Regions (ERs) and Census Agglomerations (CAs) are now based on the 2011 Census boundaries rather than the 2006 boundaries.
This paper presents a summary of the changes in the key national and provincial estimates produced by CIS and SLID as a result of the revision. Section 2 briefly describes the geographic boundary changes. Section 3 discusses the direct effect on population counts for individuals, households and economic families. Section 4 deals with the indirect effect of the revision on the main survey estimates, namely population counts by family type, income and low income. Lastly, various tables and charts comparing revised and unrevised estimates by province are provided.
2.0 Geographic boundary changes
With the change to the 2011 Standard Geographical Classification (SGC) from the 2006 SGC, boundaries were modified for some of the sub-provincial areas such as Census Metropolitan Areas (CMAs), Economic Regions (ERs) and Census Agglomerations (CAs). Specifically, seven existing CMAs had boundary modifications between the 2006 SGC and the 2011 SGC. These are, in Quebec: Saguenay, Québec, Sherbrooke, Trois-Rivières, Montréal, the Quebec part of Ottawa-Gatineau; and in Ontario: Guelph. Also, as a result of the move to the 2011 SGC, there were small boundary changes for three ERs: Campbellton-Miramichi and Fredericton-Oromocto in New Brunswick; and North Coast and Nechako in British Columbia. Sub-provincial data from 2011 to 2013 were revised to reflect these changes.
Overall, the modifications to the CMA and ER boundaries resulted in minimal changes to the population counts, income and low income estimates.
3.0 Impact on population counts
In 2015, the Demography Division of Statistics Canada revised the population counts. As a result, the target population counts for the years 2006 to 2013 were revised using the 2011 Census base. At the Canada level, little difference was observed (Table 3.1) between the revised and unrevised populations (-0.3% in 2013). Overall, the population estimates were revised downward for each of the years.
At the provincial level, Manitoba (-1.2% in 2013) and British Columbia (-1.7% in 2013) both experienced the largest downward revisions, while Newfoundland and Labrador experienced the largest upward revision (3.2% in 2013). Prince Edward Island, Ontario, Manitoba and British Columbia each experienced downward revisions for all of the years, while Newfoundland and Labrador and Quebec experienced upward revisions for each of the years.
At the Canada level, as with the changes to the population counts, the number of households changed little between the revised and unrevised data (Table 3.2). One person households and households with three or more persons witnessed relatively small changes, while households with two persons experienced a relatively larger change. In general, the number of two person households was downwardly revised for all the years (ranging from -0.2% to -2.2%), while the number of households with three or more persons was unchanged or upwardly revised for all the years (up to 0.7%).
The number of economic families with three or more persons observed little change in each of the years (Table 3.3). While economic families with two persons experienced consistently downward revisions, ranging from -0.2% to -2.3%, the number of persons not in an economic family experienced consistently upward revisions, ranging from 0.3% to 3.9%.
4.0 Impact on survey estimates
The Canadian income estimates have been recalculated to take into account the introduction of the 2011 Census population counts. In general, these revisions had only minor impacts on the estimates. In evaluating the effects of the historical revision on the Canadian income estimates, the first focus was on population estimates by family type followed by income and then low income. As estimates are based on probability samples, differences between estimates are reported only where they are statistically significant at the 95% confidence interval. Also, since the impact of the revision is generally similar in magnitude for all the revised years, the analysis will be limited to 2009, 2011 and 2013.
4.1 Family type
Overall, there were a large number of statistically significant changes in the estimates by family type (Table 4.1), although they tended to be small in magnitude. At the Canada level, the number of persons in economic families and persons not in an economic family was downwardly revised slightly (-0.3% in 2013). If this grouping is separated, the number of persons in economic families experienced a downward revision (-0.9% in 2013), while the number of persons not in an economic family experienced an upward revision (3.4% in 2013). In 2013, the downward revision in persons in economic families was largely driven by couple (-2.7%) and lone-parent families (-3.0%). The upward revision in the number of persons not in an economic family was mostly driven by non-elderly persons (3.6% in 2013).
At the provincial level, there was an even split between the provinces that experienced a downward revision compared to those that experienced an upward revision for the broad grouping of number of persons in economic families and persons not in an economic family (Tables 7.1 to 7.10).
Overall, the historical revision had a minimal effect on the average and median income estimates. The largest impacts of the revision were on the median market income and median after-tax income (Table 4.2). The median market income was downwardly revised slightly for the broad category of economic families and persons not in an economic family (-0.2% in 2009 and -1.0% in 2011 and 2013). A similar pattern was observed in the revisions to median after-tax income (-0.4% in 2009, -1.0% in 2011 and -0.7% in 2013).
At the provincial level and for the broad category of economic families and persons not in an economic family, the median after-tax income in Ontario was revised downward for each of the years. In general, the Atlantic Provinces experienced no statistically significant changes within this broad category. In Manitoba, both the median market income and the median after-tax income were revised upward in some years, while in Alberta, both medians were revised downward in some years (Tables 7.11 to 7.20).
While the revision brought about slight changes to the income estimates, as mentioned above, there was very little impact on the trends (Chart 4.1, 4.2 and 4.3). In general, the historical revision produced a time series that is comparable to the original ones, regardless of the economic family type analysed.
4.3 Low income
Low income statistics can be produced using the following three low income lines (Statistics Canada, 2016b): Statistics Canada's Low Income Cut-offs (LICO) and Low Income Measure (LIM) as well as the Market Basket MeasureNote 3 (MBM) developed by Employment and Social Development Canada (EDSC).
In general, low income rates based on these measures for all persons were revised slightly (Table 4.3 and Chart 4.4). All three were similarly affected and all, in general, experienced either no change or only very small changes. At the Canada level, for all persons, the only low income rate that had a statistically significant change was the one using the LICO after-tax (LICO-AT) thresholds.
The provincial low income statistics, for all the revised years, can be seen in Charts 6.1 to 6.10. Correspondingly, the numeric values for the provincial low income statistics for the selected reference years (2009, 2011 and 2103) can be found at the bottom of Tables 7.11 to 7.20. Over all the years, and for all persons, Nova Scotia (in 2013), Quebec (in 2011) and Manitoba (in 2009 and 2013) were the only provinces that experienced a statistically significant change using all three measures of low income.
LaRoche, Sylvie (2007). "Longitudinal and Cross-Sectional Weighting of the Survey of Labour and Income Dynamics, 2003". Statistics Canada, Income Research Paper Series. Catalogue no. 75F0002MIE— no. 007, October 2007.
Hatfield, Michael, Wendy Pyper and Burton Gustajtis. (2010). "First Comprehensive Review of the Market Basket Measure of Low Income". Applied Research Branch paper. Human Resources and Skills Development Canada. Summer.
Statistics Canada (2015a). Revisions to 2006 to 2011 income data. Catalogue no. 75F0002M, no. 3. December 17, 2015.
Statistics Canada (2015b). Improvements to the Labour Force Survey (LFS), "The 2015 Revisions of the Labour Force Survey (LFS)". Catalogue no. 71F0031X. January 2015.
Statistics Canada (2016a). "Population and Family Estimation Methods at Statistics Canada". Catalogue no. 91-528-X, March 2016.
Statistics Canada (2016b). "Low Income Lines: What they are and how they are created". Catalogue no. 75F0002M, no. 2. July 2016.
Tremblay, J. (2005). "Aperçu de la stratégie de calage harmonisée des statistiques du revenu de Statistique Canada". Proceedings of the Survey Methods Section of the Annual Meeting of the Statistical Society of Canada, June 2005.