Survey non-response in Covid-19 times: The case of the labour force survey

With the advent of the Covid-19 pandemic, labour-force survey non-response rates have surged in many countries. We show that in the case of Canada, the bulk of this increase can be explained by the suspension of in-person interviews following the adoption of telework in Federal agencies, including Statistics Canada. Individuals with vulnerabilities to the Covid-19 economic shock-i.e., the young, low-educated, lowsalary, low job-tenure individuals, and those working in occupations with low telework potential-have been harder to reach and have been gradually less and less represented in the Canadian Labour Force Survey (LFS) during the pandemic. Using exogenous variation in the assignment of individuals to the different LFS rotations, we present evidence suggesting that the decline in employment and labour-force participation have been underestimated over the March-July 2020 period. We believe, however, that these non-response biases have been moderate when contrasted with the unprecedented severity of the Covid-19 disruption. Furthermore, since attrition only represents a minor part of the non-response increase, we argue that one should not expect additional difficulties when using panels as compared to cross-sectional samples, and when using public-use LFS files instead of restricted-access files. All in all, the LFS remains a reliable data source for analyzing the economic impact of Covid-19 in a timely manner.