Population Projections Project for Canada's Census Areas

How can society intelligently participate in the setting of local to national priorities if the population does not understand where we are and where we are heading on current trends?1

The Population Projections Project (PPP) is an encyclopedic reference with which to understand where we are and where we are heading on current demographic trends. In particular, trends that impact population change, namely, fertility (is the rebound continuing?), life expectancy (has the increase come to an end?), and migration by age and sex (are patterns changing?) not only for the nation and provinces and territories, but for regions within, to cities, towns and villages.

Why refer to the Population Projections Project for Canada's Census Areas? The Population Projections Project:

1) is based on open data (not withheld and manipulated by a small group of secretive people who need not explain how they change the numbers),2

2) as reliable as possible (Canada's Census of Population, arguably one of the best sources of demographic trends for any nation in the world, providing an estimate of error), these are counts, not estimates,3

3) uses as straight forward and clear a method as possible (clear instructions provided - tab above),

4) therefore, the projections can be replicated, not only to ensure that correct methods and data accompany "findings" rather than incorrect methods and data accompanying unsupported numbers, but to understand the strengths and weaknesses of the methods and data - making clear what the numbers represent,

5) and calculates several well defined population projection scenarios directly from the Censuses of Population rather than just one possible future scenario pretending to foretell the future4,

6) not only for the nation as a whole, and Canada's provinces and territories, but for the regions within to the cities, towns and villages across the country,

7) not only providing total population numbers but also projections of the number of people by age and sex,

8) using maps and charts to provide proportions and perspectives,

9) previous projections remain available and open for review rather than removing previous projections (making it impossible to check how close or far off they were); therefore, the projections calculated after each census release can be compared with the next census results,

10) provides a reference independent from the political direction of the current government with which to compare other population forecasts/projections, to help safeguard against the intrinsic risks and danger of monopoly that may result in unsupported / false numbers being used to guide public services.

These standards and unique qualities make the Population Projections Project a reliable reference with which to be aware of where we are and where we are heading on current demographic trends.

Now that the 2016 census counts, by age and sex, are available, comparisons between the 2016 projection scenarios calculated directly from the 1996 to 2011 censuses versus the actual 2016 counts can be made. The original PPP page will remain showing the projection scenarios based on the 1996 to 2011 censuses to provide a comparison with the actual 2016 census counts (rather than removing the original projections, thereby not allowing a check to see how useful the projections were).

Comparisons of the actual 2016 census counts with the 2016 projection scenarios continue to be added to the PPP census areas pages, as well as being posted on the Population Analysis Articles page.

In accordance with Statistics Canada's (STC) practice, rather than publishing only one possible scenario that pretends to foretell the future, a set of well defined scenarios are provided.

"Forecasting pretends to foretell the future, while the projection is an analytic tool which -- within the constraints of a tightly specified model -- enables the analyst to consider the implications of alternative scenarios."
"The issue of scenarios is important: in order to underline the analytic character of projections, it is our policy to publish always a set of possible projections, each corresponding to a well defined analytic scenario." (Ivan Fellegi, 1999)5

While the projections are based on the census counts, over the last 20 years, available from Statistics Canada, Statistics Canada did not examine, test, review, nor contribute to the Population Projections Project (PPP) in any way. The PPP adapts Statistics Canada's census counts available online (link below) then uses the cohort change ratios method and the child women ratios to calculate scenarios that are valid and can be replicated. Report any problems.

By referring to the Census of Population and making the methods clear, community members, (including people providing public services as well as elected representatives) have a resource with which to understand where we are and where we are heading on current demographic trends. You can participate in sharing this information with people in your area.

Endnotes and Sources:

1 This question, as well as the rest of the introduction, was adapted from "The United Nations Economic and Social Council, Statistical Commission and Economic Commission for Europe, Conference of European Statisticians, 47th plenary session, "Analytic Activities at Statistics Canada", submitted by Statistics Canada, prepared by Chief Statistician, Ivan Fellegi, 1999." www.wminfomatics.com/WP/ANALYTIC_ACTIVITIES_AT_STATISTICSCANADA.pdf

2 See the summary of the introduction to "Population Projections for Community Members" presented at the 2013 Congress organized by the Federation of Humanities and Social Sciences, Methods for Projections session hosted by the Canadian Population Society. .

3 This project refers to Canada's census area's age/sex counts from the Census of Population and Dwellings available on Statistics Canada website at http://www12.statcan.gc.ca.

NOTE TO THE READER - A word of caution

These are counts not estimates. The census misses some people and can count others more than once. Counts do not include net undercount (overcount minus undercount). Estimates do include an estimate of net undercount using a procedure referred to as a reverse record check. As well, counts are randomly rounded to 0 or 5. Estimates are not randomly rounded.

Why use counts then?

Counts are the base for estimates and we are looking for trends from the national to local (municipal and smaller) areas. If we see trends in the counts that are not seen in the estimates then we must ask why and require a clear explanation. See point #1 regarding open data.

The Population Projections Project does not over look nor by-pass the census counts, but rather starts with the census counts because they are raw and genuine.

They are so genuine that they are very slightly altered to protect individuals living in very small population areas from being identified through the census by randomly rounding to 0 or to 5. For example, if there was 1 person in the 80 to 84 year age group in a small population area, the counts for the 80 to 84 year age group would show either a 0 or 5 generated randomly. Random rounding is done for counts for all census areas, from the national to local levels.

Why are estimates not rounded to 0 or 5? Primarily because estimates are only provided from the national to the census division level (regional districts in BC, counties in Ontario etc.) and are not provided to the census subdivision level including small towns and villages. Also, estimates include net undercount which has a margin of error as well. For example, between 5% and 10%. In 2016 STC began providing one figure, for example 5.1% rather than between 5% and 10%, but it is helpful to be aware that this is an estimate and can be considered in terms of a range.

Also, postcensal estimates (estimates produced annually for each year after the last census until the next census counts) get adjusted after the release of the next census. This is why the postcensal estimates are referred as preliminary.

There can be mistakes in the counts; reviewing the counts is fundamental to quality assurance of the counts and of the estimates, and occurs shortly after the census release.

As well, we are looking for demographic trends from the national to local census subdivision (municipal) levels. Estimates from STC only go from the national to the census division level - not the census subdivision (municipal) level.

The projections calculated directly from the Census of Population counts for Canada's census areas are marginally lower than the projections calculated directly from Statistics Canada's estimates but are otherwise very similar. As mentioned, any differences in the trends are important to consider.

To see the comparison of demographic trends using counts versus estimates visit the Population Analysis Articles page. Scroll to the November 4th, 5th, 6th, and 7th 2016 articles comparing trends in total population, broad age groups, dependency ratios, as well as children 0 to 4 and 5 to 9 years of age.

To get total population estimates that take into consideration net undercount, simply add Statistics Canada's estimates of net undercount (Global non-response rate (GNR)) to the total count. For example, the estimated 2016 GNR for the Capital Regional District = 5.1%. To figure out the estimated total population for the Capital Regional District just add 5.1% to the count.

The PPP does projections by 5 year age groups for each census year into the future (i.e. 2021, 2026 ...2046). Nonetheless, the PPP can and does produce annual postcensal counts as well as postcensal estimates referring to historical estimates of net undercount provided by STC, for both 5 year age groups as well as single years of age.

A word of caution regarding population projections from any source:

Important considerations when looking at population projection scenarios are:

1) they must be verifiable and replicable thereby safeguarding the integrity of the population projections (if they are not replicable then they maybe made up even if they are from an official statistical office - especially if they are from an official statistical office or a consultant hired by a special interest group including tax funded consultants) and

2) they must include more than one possible future scenario because one scenario (according to former Chief Statistician for Statistics Canada, Ivan Fellegi) pretends to foretell the future, while several well-defined projection scenarios highlight the analytic character of projections (these are not ordained by a god or gods or a senior government official) and allows us to consider the implication of alternative scenarios.

4 For an example, compare the Superintendent of School's forecast of 0% change in the number of school aged children by 2027 with the "Vancouver Island Health Authority forecast of a 30% increase in total population" for the same area, to 2030 both providing only one future possibility - both pretending to foretell the future.

5 "Analytic Activities at Statistics Canada", page 5.

Report a problem email information [at] wminfomatics [dot] com

Website content, code, and design by William Warren Munroe. Copyright 1999 - 2019 ©