Munir A. Sheikh is executive fellow at the School of Public Policy, University of Calgary and former chief statistician of Canada.
I commend Prime Minister Justin Trudeau's new government for the decision, formally announced Thursday morning, to reverse the cancellation of the long-form census. I am also encouraged by Mr. Trudeau's emphasis on the need to use evidence in decision-making.
The previous government's cancellation of the long-form census was irresponsible. It led to deterioration in the quality of the census data collected in 2011 by the voluntary National Household Survey, which replaced the long-form census. As well, data for about a quarter of Canadian communities is missing because of their particularly poor quality or confidentiality concerns. And all that grief when the replacement, the NHS, cost $22-million more than the long-form census!
All of these problems related to less and lower-quality data from the NHS have been well documented by many researchers. Can you imagine a single Canadian who would behave like the previous government did, choosing to pay a higher price for less quantity of a good of lower quality?
In addition to the lower-quality data collected by the NHS, further damage has been done because of the loss of the census anchor for validating and adjusting many other Statistics Canada household surveys. Keep in mind, too, that the non-response rate threshold beyond which census data is not released was raised from 25 per cent to 50 per cent for the NHS, meaning that with the census standards for dissemination, a lot of the NHS data would not have seen the light of day, given the NHS global non-response rate of over 26 per cent.
The quality of data is likely particularly bad for surveys of small populations, such as Canadians with disabilities and aboriginal peoples, and for small areas, such as individual communities. For example, 40 per cent of communities in Saskatchewan have no data from the NHS. A host of organizations – governments, NGOs, businesses – operate in small communities, and they need local-level data for a wide range of their activities. These were put at risk with the low-quality data the NHS produced at that level.
The average Canadian may not use census data directly in their everyday lives, but organizations that provide goods and services to such Canadians rely heavily on them. Long-form data is a critical factor in decisions made by the federal government, provinces, municipalities, businesses, NGOs, academics, pollsters and the news media, among many others. The long-form census contains questions on income, the labour market, education, housing, transportation, languages, disabilities, citizenship and immigration, aboriginal peoples and ethnicity. Anyone working in any of these areas has suffered because of its cancellation and, as a consequence, all Canadians who receive services from these organizations have suffered.
Consider the following examples. With the lower quality of income and labour-market data produced by the NHS, the federal government would find it harder to correctly read the country's economic situation, changes in Canadians' living standards, characteristics of the labour market, people living in poverty or the effectiveness of federal social programs.
Worse, with the NHS data not being comparable to the 2006 census information, it becomes impossible to observe trends over time. It is harder for provincial governments to plan social programs, determine housing and education needs, or assess transportation requirements, including infrastructure and transit planning. It is harder for municipalities to plan services, such as determining the fate of a library or the building of a firehall.
Businesses are equally affected. A business would find it harder trying to decide whether to open a new store in a community or what stocks to keep. A bank would run into difficulty trying to determine whether to open a new branch or to provide services in minority languages.
Academics are affected, as well. One of Canada's most respected researchers, Charles Beach, professor emeritus of economics at Queen's University, had this to say about the cancellation of the long-form census: "It has certainly impacted my own work on what has been happening to middle-class earnings in Canada," he told The Globe and Mail in January. Its loss "inhibited research into inequality and identifying winners and losers in economic growth, research into understanding the national problems of the have-nots in the economy and research into how best to provision local government services."
With so much at stake as a result of the cancellation of the long-form census, it is appropriate to celebrate this policy reversal. However, we should also ask why and how we got into this unacceptable situation in the first place, with a government interfering in the very technical issue of which questions should be asked on the census form. Could we find ourselves back in the same situation again one day in the future?
The answer is that the Statistics Act, the law governing Statistics Canada, is flawed – it gives the responsible minister final authority in deciding on technical statistical matters. The law also gives cabinet the authority to determine questions that should go into a census. This is simply not right.
Given the law, there is no reason to be sure that a future government would not again cancel the long-form census. I believe the contents of a census should be a decision purely based on a country's data needs and not on the politics of the day.
Indeed, the Liberal Party's election platform promised to make Statistics Canada independent – I hope the new government fulfills this promise in the near future.