When it comes to the census, who is counted and who isn’t matters.

On a subject that will not die. Consider this a follow-up to the piece Debunking the “homes outpace households” argument.

As always, here’s the TL;DR

TL;DR: Despite media stories and discourse on Twitter, the number of new homes is not rising faster than the number of new households. The discrepancy is due to how the Census undercounts young people, non-permanent residents, and persons who do not have English or French as a first language — all populations that are rising in Ontario.

Quick question: How many people lived in Canada in 2021? The answer should be somewhere in the range of 38 million to 39 million people.

Unless you ask the Census, which states the answer is fewer than 37 million.

Why so low? It’s because the Census undercounts the number of people who live in Canada. Just ask Statistics Canada:

The 2021 Census counted 36,991,981 people in Canada during the national enumeration with reference date May 11, 2021. This count is lower than the preliminary postcensal population estimate of 38,201,103 people calculated for the same reference date. The difference between the two figures is not unexpected and is similar to that which was experienced for previous censuses. This note outlines why there are differences between census counts and population estimates.

The objective of a census is to provide detailed information on the population at a single point in time. In this respect, one of its goals is to enumerate the entire population. Inevitably, however, some people are not counted, either because their household did not receive a census questionnaire (for example, if a structurally separated dwelling is not easily identifiable) or because they were not included in the questionnaire completed for the household (for example, the omission of a boarder or a lodger). Some people may also be missed because they have no usual residence and did not spend census night in any dwelling. In contrast, a small number of people may also be counted more than once (for example, students living away from home may have been enumerated by their parents and by themselves at their student address).

In short, some people get added to the Census twice, and some people don’t get added at all, like those experiencing homelessness.

As Statistics Canada points out in this piece, this under (or over) count is particularly an issue for higher-education students not living with a parent, who get enumerated one of four ways:

• Both at their student address and at the address of a parent
• Only at their student address
• Only at the address of a parent
• At neither address

The difficulties with enumeration is a known issue with the Census, Statistics Canada works to estimate who is being missed by the census, and who is being counted twice. Because students in higher education are one of the most difficult groups to enumerate, and the number of higher education students is on the rise, so too is both the overcount and undercount in the Census, as estimated by Statistics Canada:

The entire piece Coverage of the 2016 Census: Level and trends (where the above chart is taken from) is worth a read. I’ll highlight a few key points.

The StatCan piece contains the following TL;DR on which groups experience high census net undercoverage (CNU) rates; that is which groups are likely to be undercounted by the census:

• Young adults, especially men, show particularly high CNU rates. CNU rates are also higher among individuals who live in the territories, on Indian reserves, who are single, separated, or whose mother tongue is a language other than English or French;

They also note that “[r]ecent immigrants and non-permanent residents (NPRs) are two rapidly growing demographic groups that have particularly high missed rates in the census”. As I’ve discussed in a number of pieces, Ontario’s population boom is due, in large part, to an increased number of non-permanent residents.

Young men, in particular, have a high undercoverage rate (likely, in part, because they’re also a large part of the non-permanent resident and recent immigrant population):

And for likely similar reasons, undercounts tend to be particularly high in Toronto and Vancouver:

Good planners understand that the Census undercounts population, and, in particular, undercounts the population of new Canadians, non-permanent residents, higher ed students, and those who do not speak English or French. For example, Peel Region has a one-pager on this in the context of the 2006 census. They make it clear that understanding who is covered (and who isn’t) by the Census is crucial in planning:

When using Statistics Canada census data, keep in mind that the total population indicated in the census is not the actual total population of an area or geography).

Like other data gathering tools, censuses are not perfect sources of information. When the census is conducted, there is room for error and residences to not be accounted for. Missing data in census counts is called “undercoverage”.

To help cope with this issues, Statistics Canada estimates an “undercount” or “undercoverage” rate to assist estimating the real total population. Undercount rates are typically provided for each Census Metropolitan Area in the country…

The difference between the total and the Census Population [for Peel Region] is significant: over 60,000 people. More people than were reported in Caledon alone for 2006!

When planning for services, keep undercounts in mind!

The undercoverage rate that the Peel one-pager discusses will not be released until September 2023 for the 2021 Census.

Despite the fact we are a year and a half away from knowing the undercount rate at a provincial or local level, we’re seeing pieces like this one at the Globe and Mail assert that number of occupied dwellings is growing faster than the population:

The 2021 Census reported that from 2011 to 2021, Ontario’s population grew by 10.7 per cent and the number of occupied dwellings grew by 12.5 per cent.

If we’re using Census Data for this analysis (as the G&M piece does), we won’t actually know if this is true until the September 2023 undercoverage rates are released. Statistics Canada themselves advise applying caution when using census data.

These results are a reminder that some caution must be used with census data. O’Hare (2019) reported that a relatively low CNU rate can hide significant undercoverage or overcoverage as well as large differences between demographic groups. Furthermore, coverage errors are not calculated for all characteristics available in censuses, so coverage for certain groups likely to be less well covered, like less educated individuals or those with lower incomes, is not well known.

Harder-to-enumerate populations often face physical, economic, social and cultural barriers.

Those barriers are vital here. If we’re basing public policy, such as housing policy, on census data, and if we’re not recognizing that there are groups that the census undercounts, we will inadvertently be embedding those barriers into planning. In particular, we will be underestimating the need for housing, and other public services, that are needed for young persons who are either new Canadians or non-permanent residents, creating further structural barriers.