Open this photo in gallery
Nunavut is an extreme example of a broader trend. The national Labour Force Survey response rate has dipped below 70 per cent – almost 20 percentage points lower than the monthly average in 2019.
Emma Tranter/The Canadian Press
Statistics Canada is struggling to get residents of Nunavut to respond to its flagship labour survey amid the pandemic, a situation that has worsened over time and raised concerns about the quality of employment data for the territory.
In Nunavut, the household response rate to Statscan’s Labour Force Survey has fallen to roughly 16 per cent from a prepandemic monthly average of 75 per cent. Nationally, the response rate has ebbed to historic lows of about 70 per cent in recent months.
Labour Force Survey response rate in Nunavut
0 20 40 60 80% Mar. 2020 Apr. May June July Aug. Sept. Oct. Nov. Dec. Jan. 2021 Feb. Mar. Apr. May June July Aug. 75% average from Jan. 2019 to Feb. 2020 61.4 THE GLOBE AND MAIL, SOURCE: statscan
data
share
×
month Nunavut 2020-03-01 61.4 2020-04-01 56.8 2020-05-01 49.4 2020-06-01 51 2020-07-01 52.3 2020-08-01 47.2 2020-09-01 46.6 2020-10-01 44 2020-11-01 42.1 2020-12-01 37.9 2021-01-01 30 2021-02-01 42.4 2021-03-01 33.6 2021-04-01 27.9 2021-05-01 25.9 2021-06-01 19.8 2021-07-01 15.5 2021-08-01 15.7
Labour Force Survey response rate in Nunavut download csv
×
Share this chart:
https://s3.amazonaws.com/chartprod/FSGNKTcCtGDBtMasY/thumbnail.png
The survey is among the most important economic reports in Canada and contains some key indicators, such as the unemployment rate. The monthly report not only informs government policy, it also determines who qualifies for jobless benefits through employment insurance – and for how long.
Its results hinge on the participation of Canadians. Every month, Statscan seeks to interview (in person, by phone or online) about 56,000 households on their work activities. Those households generally remain in the survey for six consecutive months.
However, it’s harder to get a reply. Since the outset of the pandemic, the agency has suspended the use of face-to-face interviews, which normally account for about a fifth of all interviews. As a result, survey response rates have dropped to historic lows.
The impact was more pronounced in Nunavut. Almost all first interviews in the territory are normally conducted in person, Statscan said in a technical note to researchers. Without that method, it’s much tougher to reach people there than in any other province or territory.
“Nunavut really is a special case,” said Vincent Dale, director of Statscan’s centre for labour market information. “More than half the dwellings, we don’t have a phone number.”
In particular, there has been a sharp drop in responses from people without a university degree, who in other regions have been disproportionately affected by pandemic-related layoffs.
As a result, Statscan is now advising people to use a 12-month moving average of labour estimates for Nunavut, up from the standard three months for the territories.
The situation is worrying those in academic circles.
Story continues below advertisement
The response rate in Nunavut “is a number that I would be very, very concerned about,” said Pierre Brochu, a professor of economics at the University of Ottawa.
“Up north, I feel like you’re flying blind in understanding what’s happening in their labour market. It’s a strong statement, but I truly believe that.”
Nunavut is an extreme example of a broader trend. The national Labour Force Survey response rate has dipped below 70 per cent – almost 20 percentage points lower than the monthly average in 2019.
Fewer responses do not inherently mean the data are worse. The concern, however, is that non-respondents have been affected differently by the pandemic, thus skewing the results.
And that may be the case. In a working paper, Prof. Brochu and Jonathan Créchet (also an economics professor at the University of Ottawa) suggest that Canada’s employment losses were underestimated in the initial months of the pandemic because vulnerable demographic groups – the young, low-paid and less-educated – were tough to reach.
Despite that, Prof. Brochu said the Labour Force Survey data quality is generally “outstanding.”
In Nunavut, however, the results have been volatile. The unemployment rate (expressed as a three-month moving average) was 3.5 per cent in June, the lowest on record. The next month, it was 5.9 per cent. Then, in August, it jumped to 12.4 per cent, closer to historical averages.
Nunavut unemployment rate
Three-month moving average
0 5 10 15 20% Mar. 2020 Apr. May June July Aug. Sept. Oct. Nov. Dec. Jan. 2021 Feb. Mar. Apr. May June July Aug. Historical avg. 15.6 THE GLOBE AND MAIL, SOURCE: statscan
data
share
×
2020-02-01 13.3 2020-03-01 15.6 2020-04-01 17.8 2020-05-01 16.8 2020-06-01 15.5 2020-07-01 17.9 2020-08-01 16.4 2020-09-01 16.5 2020-10-01 13.5 2020-11-01 12.2 2020-12-01 9.3 2021-01-01 7.1 2021-02-01 6.5 2021-03-01 7.5 2021-04-01 6.2 2021-05-01 5.4 2021-06-01 3.5 2021-07-01 5.9 2021-08-01 12.4
Nunavut unemployment rate download csv
×
Share this chart:
https://s3.amazonaws.com/chartprod/jkbXEdNdhwLBq5e4g/thumbnail.png
But those estimates come with substantial margins of error – large enough that the true unemployment rate may not have actually changed between July and August.
“These data are just not very high quality,” said Mikal Skuterud, a professor of economics at the University of Waterloo. “You can’t really make anything out of them, even when the unemployment rate doubles.”
In a statement, a Government of Nunavut spokesperson said the territory presents a range of challenges for data collection.
“Cost factors, small community-level populations, language barriers, reliance on southern methods of communication, lack of community buy-in, and in some cases, no or limited access to internet or phones, all impact data collection,” said Angela Petru. “As a result of these challenges, the [Government of Nunavut] cannot rely on data to guide policy decisions to the same extent as other provinces and territories.”
To bolster response rates, Statscan is sending people into the field to procure contact info for phone interviews. However, this hasn’t been done in Nunavut, Mr. Dale said. Travel to the territory requires authorization from the Chief Public Health Officer, which Statscan has not sought. Statscan employs three interviewers in the territories, but none in Nunavut.
Making contact is a pressing concern. In the territories, dwellings remain in the survey sample for two years (rather than six months in the provinces) and residents are interviewed every three months (rather than every month). Thus, some people in the Nunavut sample were contacted in person before the pandemic – and will soon be vacating the survey, potentially adding to response issues.
Statscan is hoping to get interviewers on the ground before the end of the year, Mr. Dale said.
“We feel we’ve exhausted all efforts” to match phone numbers to dwellings, he said. “Really, it’s a question of when are we going to be able to send interviewers back into the field.”
Your time is valuable. Have the Top Business Headlines newsletter conveniently delivered to your inbox in the morning or evening. Sign up today.