Appendix I: Component changes in population rebasing
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.
Population estimates are produced using what is commonly referred to as the component method. This method involves adding and subtracting components of population change to the base population. These components include births, deaths, immigration, emigration, interprovincial migration, and non-permanent residents (NPRs). The result of this is the total population for all age groups.
The target or sampled population for the LFS includes the population aged 15 and over, and excludes people belonging to the following categories: population living on reserves, institutional residents, and full-time members of the Canadian Armed Forces.
Any changes in the methodology of estimating the components are also introduced whenever a new series is produced. The revised population estimate series was produced using a different methodology for producing the estimates of emigration 1 and the population living on reserves.
The difference between the unrevised and revised population estimates prior to July 2006 is mainly explained by the population living on reserves as this component is estimated differently, although the difference is relatively small compared to the other periods.
The most noticeable change in the methodology of estimating the population living on reserves is the use of new Net Census Undercoverage (NCU) rates specifically for the reserve population, as opposed to the rates used for the whole population. The new NCU rates resulted in an overall higher estimate of the population living on reserves. In addition, the reclassification to the 2011 Standard Geographical Classification (SGC) involved modifications to the status of several Census Subdivisions (CSDs). Some CSDs were reassigned from Indian Reserve status to other types, and vice-versa, resulting in some differences between the two series. By excluding the higher estimate of the population living on reserves in the total population, the sampled population decreases.
The main cause of the difference in the Canadian sampled population between the unrevised series and the revised series after June 2006 is due to the differences that arise between the unrevised series and the adjusted population count from the census, or the error of closure. Thus, when comparing between the unrevised series (based on the 2006 census counts) and the new revised series (based on the 2011 census counts), differences can arise in the base population as the revised series is corrected for the error of closure. This is the biggest reason for the difference between the unrevised and revised series.
After June 2011, the difference in the base population, caused by the error of closure, was retained and continues to explain most of the difference in sampled population between the revised and unrevised series. Two other factors also contribute to the difference between the two series, both relating to the issue of data availability.
First, the estimates of institutional residents are based on the data from the latest census. As the unrevised series did not have the 2011 census data, a difference would naturally arise between the two series.
Second, the estimates of non-permanent residents (NPRs) are based on a 12-month moving average centered on the reference date in order to avoid spikes in the series. As the unrevised estimates were produced on every reference month, it uses NPR stocks from the previous year. The revised series, however, are estimated with all available data at the date of production. As such, the estimate of NPRs differs between the two series. The fluctuation of the difference in sampled population after June 2011 is observed for this reason.
The sampled population estimates of all provinces were affected by the rebasing of the population estimates, although the extent varies from province to province. The leading causes of the differences between the unrevised and revised series are the same as mentioned previously for the Canadian sampled population.
Between 2001 and 2006, Manitoba displayed the largest absolute and relative difference between the two series. Other provinces such as Saskatchewan and British Columbia were also affected for this reason. The differences are explained by the changes to the methodology of estimating the population living on reserves.
From 2006 to 2011, the gap between the unrevised and revised series at the Canada level grows relative to the previous period, as observed in Chart 1. This gap is mainly the difference between the unrevised estimates and the 2011 Census counts, as the revised series has adjusted the base population in order to correct the error of closure.
This difference has the largest absolute impact on populated provinces such as British Columbia, Ontario, and Quebec. In the Atlantic, Newfoundland and Labrador displayed the largest difference, as the error of closure for that province was the highest. Quebec was also affected by the differences in the institutional residents 2 at this time, while the estimates for Saskatchewan continue to be affected by the population living on reserves.
From 2011 onwards, the factors mentioned in the previous periods continue to impact the provincial estimates. The gap in the NPR estimates between the two series affects the movement of the sampled population for each province, but this gap is especially noticeable in Alberta.