Measuring the pandemic: How Statistics Canada kept the data flowing amid COVID-19

Chief statistician Anil Arora at the Senate Committee on Banking, Trade and Commerce in November 2018.
Chief statistician Anil Arora at the Senate Committee on Banking, Trade and Commerce in November 2018. The Canadian Press/Sean Kilpatrick

Anil Arora understands that filling out a Statistics Canada survey isn’t a priority for people coping with the fallout from COVID-19. But Canada’s chief statistician says the pandemic has demonstrated the importance of the agency’s work as governments assess their aid efforts and businesses plan for the recovery. 

“Canadians are understanding the value of making sure that data are used to drive decisions, and the role that Statistics Canada plays,” Arora told The Logic, adding that the “pandemic has demonstrated that if we don’t do this, there’s a cost.” In an interview on the release day for its flagship Labour Force Survey (LFS)—which showed a continued albeit slowing jobs recovery in August—the chief statistician discussed how the agency has adapted for the pandemic and its response to calls for more race-based data.

Purchase a subscription to read the full article.

By entering your e-mail you consent to receiving commercial electronic messages from The Logic Inc. containing news, updates, offers or promotions about The Logic Inc.’s products and services. You can withdraw your consent at anytime. Please refer to our privacy policy or contact us for more details.

Already a subscriber?

Talking Point

Statistics Canada has accelerated the release of some economic indicators and begun collecting more race-based data during the pandemic, and chief statistician Anil Arora says COVID-19 has shown the agency’s importance as policymakers and businesses try to assess the effectiveness of their efforts and plan for the recovery.

During the pandemic, StatCan has released flash estimates of measures like GDP and manufacturing sales, for which full reports typically take a few months. “It’s better to get it mostly right and put something out that is useful to people in a time when things are changing so rapidly, than to put out something that is perfect after the time [when], it would be most useful,” Arora said, describing the agency’s rationale. 

Typically, once StatCan’s survey responses are in—the business ones are mostly mandatory, so completion rates are high—the agency collates them, then challenges their initial results by comparing them with complementary indicators, and breaks them down into subcategories like industry or province. To arrive at its high-level estimates for early release, it puts together the data from the “dominant players” for a particular measure; the numbers are typically revised a little in the later, full report. 

The agency has been “obsessed” with improving the timeliness of its indicators over the last three years, and has increased the turnaround period for several, Arora said. StatCan will continue producing flash estimates, which have earned “a lot of kudos” from policymakers and businesses. 

Economists and policy experts to whom The Logic spoke this month also praised the agency’s increasingly detailed analysis of the pandemic’s effect on Canada published alongside raw data on subjects like the labour market. But some expressed concerns about the frequency and speed of its reporting; the pandemic’s quick impact and governments’ rapid launch of programs in response mean that data based on a one-week survey taken a month ago may miss important changes, or simply be stale.

Private-sector data has filled some of the gap. Banks have begun releasing weekly analysis of their customers’ debit- and credit-card aggregated use, while open source information like mobility trends from Google and restaurant reservations from OpenTable have attracted coverage. “We don’t compete per se with other organizations,” said Arora. “We are the national statistical authoritative agency.” A company’s visibility is often limited to its customer base, he noted—a payroll provider might have data on wage trends, but their clients may not fully reflect the city or sector, for example. StatCan’s measures are designed to be representative.

“I would encourage those institutions and organizations to partner with us,” Arora said, adding that the agency is increasingly acting as a “data steward,” integrating and linking data from different sources to provide insights for policymakers’ and private-section decisions.

Economists have also expressed concerns about the dropping sample size of the LFS, noting that the dropoff may not be random. If, for example, the workers most likely to lose jobs because of COVID-19 are also most likely to not respond, they’d be undercounted, leading policymakers and researchers to conclude the pandemic’s economic impact has been less significant than in reality. 

“We’re still very confident in the quality of the information that we’re putting out,” Arora said, noting that the six geographic clusters from which StatCan draws its sample are representative of the population. A team of methodologists, operations staff and subject-matter experts meets daily during the survey week to analyze and report potential biases, and tweaks the interviews accordingly. 

Those conversations are typically conducted in person, but in March, the agency sent frontline and call-centre staff home to heed public health advice. “People working from home were busy, not necessarily putting the same priority on filling out a Statistics Canada survey,” Arora said, but it’s still received responses from about 40,000 households monthly.

In July, the agency started collecting race-based data in the LFS, a survey it has been collecting monthly since 1952. Arora said it’s clear the information is now needed to answer policy questions about how COVID-19 was impacting different subgroups of the population. But StatCan hasn’t always seen the same case for it. “When there are slow-moving phenomena in society that you don’t need a monthly indicator of what’s changing in a particular subgroup, why would you impose additional burdens?” he said. “You have to make priority decisions.”

Share the full article!
Send to a friend


Thanks for sharing!

You have shared 5 articles this month and reached the maximum amount of shares available.

This account has reached its share limit.

If you would like to purchase a sharing license please contact The Logic support at [email protected].

Want to share this article?

Upgrade to all-access now


Racial disparities in health and labour-market outcomes, wealth, education and other social and economic indicators are hardly a pandemic phenomenon, however. In June, University of Alberta professor Malinda Smith told The Hill Times that Canada’s disaggregated data was “very poor.” Arora does not entirely accept that StatCan’s efforts have been lacking, citing its collection of information on visible minority group membership on the census since the 1980s and post-censal surveys of ethnic diversity. “The reason we can say that many of these disparities existed is because Statistics Canada has done these studies,” he said. 

The agency is also seeking to link its existing information stores with administrative data to fill in some gaps. In May 2019, the agency launched the Centre for Gender, Diversity and Inclusion Statistics, which publishes tools and standards so that other organizations can collect data that’s interoperable with the government’s own. It’s also partnering with the Canadian Association of Chiefs of Police to report on Indigenous and other ethno-cultural groups in police statistics and collaborating with provincial health agencies on mortality rates for members of different communities. 

“The context and the situation can sometimes increase the urgency of these kinds of things and the willingness of Canadians and other organizations to start sharing data,” Arora said. “Statistics Canada stands at the ready.”

Other highlights from The Logic’s interview with Arora

On the “intangibles” economy:

“Every single thing that we measure, we have to have a digital lens to it, whether it’s the kind of occupations that are changing and the [National Occupational Classification] codes [or] whether it’s the quality-of-life or quality-of-employment indicators.”

Arora said the agency is well-placed to assess the growing importance of intangibles—assets like data and IP innovation policy experts say must be added to Canada’s benchmarking of its global competitiveness. StatCan is participating in OECD committees that are defining metrics for the economic and social impacts, such as cyberbullying. It was the first statistical agency in the world to estimate the value of domestic investments in data, according to Arora.  

On government data:

“We will continue to play [the] role of data steward, so that we can get the insights and the value that’s locked in those data. I’m a firm believer that siloing data is not in the interest of Canadians. But we also need to do it in responsible ways where they absolutely understand how we’re protecting their privacy and their confidentiality.”

Arora co-authored the federal Data Strategy Roadmap, a series of recommendations on how Ottawa should collect, manage, govern and share information, published in August 2019. StatCan has hosted conferences with experts from other departments, put on courses and seminars on proper data use, and convened chief data officers from across the government to discuss issues. Arora himself has leadership roles on UN and OECD statistical committees. 

On administrative data:

“I had committed that we would put those two [banking-data] projects that were the subject of [the Office of the Privacy Commissioner’s] investigation on hold, and not proceed until we met those requirements. And those have not yet proceeded. We are still in the last stages of working with the Office of the Privacy Commissioner.”

In December 2018, StatCan suspended plans to get aggregated, anonymized transaction data from nine banks and financial institutions as well as to use information from a credit bureau following a backlash from policymakers. Arora said the agency is following the privacy commissioner’s recommendations, including greater public transparency and ensuring any collection is necessary and proportional. StatCan has since added new administrative data sources in other fields, such as using satellite imagery and anonymized, aggregated claims information from insurance companies to assess crop yields.

This story has been updated to include Arora’s comments on collaborating with private-sector organizations.