Michael R. Veall
Department of Economics, McMaster University
Hamilton ON L8S 4M4
July 6, 2010
The Honourable Tony Clement, P.C., M.P.
Minister of Industry
House of Commons Ottawa, Ontario, K1A 0A6
minister.industry@ic.gc.ca
Dear Minister Clement,
The Executive Council of the Canadian Economics Association has asked me to write to you to express its concerns at reports that you have decided to replace the mandatory long form in the Census with a voluntary survey.
Important and reliable data have been collected by the Census long form for over a century: for example occupation and school attendance were first collected in 1871, wages and salaries in 1901, working weeks in 1911 and highest level of schooling and the number of rooms in dwellings in 1941. The Census is thus a precious record of our progress as a nation. The proposed change jeopardizes this legacy by risking the quality of the data. For example, it may be impossible to determine whether a new trend in the 2011 data is the consequence of real change or just the different method of collection. The 2011 Census could be costly failure.
We understand that this is a complex issue given the participation resistance from some individuals. (We address this in the attached memo.) But the purpose of the Census is to ensure that public policy is based on the best possible knowledge. Many firms also depend upon the reliable, detailed, small area data provided by the census long form. Therefore, we ask that there be an opportunity for consultation, open to both users and the general public. The discussions would seek to balance the benefits of the data with privacy concerns. If there is inadequate time for such consultation, our view is that the risk of losing the embedded value in the Census is too great, and that the change should be delayed. A number of intermediate measures are possible, including running the new survey in parallel on a smaller sample as a test.
Making this change without consultation will damage Statistics Canada’s currently outstanding reputation inside and outside of Canada and will leave Canada with a Census that is significantly less useful than those of the countries that Canada compares itself against. Please provide an opportunity for consultation. The Canadian Economics Association would be pleased to assist in any way it can.
Sincerely yours,
Michael R. Veall, President-Elect, Canadian Economics Association
cc/ Office of the Prime Minister,
Dr. Munir Sheikh, Chief Statistician of Canada
-----------------------------------------------------------------------------------------------------------------------------------------
And here is the memo referred to in the letter.
Eliminating the Mandatory Census Long Form: Why It is Important to Consult
Canadian Economics Association Executive Council, July, 2010
1. We understand that the reason for the elimination of the mandatory long form was objection from those who did not wish to participate. Some such objections are principled (although there are principled objections to many requirements of modern society, such as paying taxes). But we are concerned that some resistance has been fuelled by untruths that Statistics Canada does not respect data confidentiality and that serious penalties for nonparticipation are widespread. Indeed, perhaps the potential fine should be reduced and the possibility of a jail sentence should be eliminated given that these just give targets for those who wish to register protest. Perhaps also the penalties for confidentiality breech could be increased.
2. One reason to keep the Census completely mandatory is that it provides Statistics Canada with the internal mandate to ensure everyone is included. Without the mandatory provision, data for the lower and higher income groups in particular tend to become unreliable, as there is often significant underreporting*. This is a huge information gap. If we miss the top end, we won’t know much about those who pay the most taxes and make some of the most important contributions to our society. But perhaps missing the bottom end of the income distribution is even more important. This includes some of the most vulnerable. They tend to have disproportionate interaction with government: with the health care system, the criminal justice system, the immigration system and the social assistance system. How can we know how policies are working if we do not have a data tool for use in assessment? Crucially, the long form also provides fine geographical detail for local policy analysis including things like city, school and hospital planning, as well as for private sector use. This is at risk.
3. Estimates from Statistics Canada surveys such as the Labour Force Survey depend upon information from the Census. The data quality issues are broader than just the Census.
4. We understand the privacy concerns. But much of the information is already reported to various levels of government (e.g. local property tax assessors, the Canada Revenue Agency). However, the Census collects the data consistently and at once, so that it is possible to examine relationships between variables such as education and income.
5. A consultation would allow the costs and benefits of potential changes to be considered. As our letter emphasizes, the current Census data is an asset of tremendous value because it allows long-term analysis of Canadian trends and can be compared internationally to the Census of the United States and to those of other countries. If this change goes through, it is possible much of the value will be lost. Certainly if this change is made without consultation, the damage to Statistics Canada’s reputation nationally and internationally will be significant: the perception will be that there was no weight given to data quality in the decision making process.
* Frenette, Marc, David A. Green and Garnett Picot (2004), “Rising income inequality in the 1990s: An exploration of three data sources,” Statistics Canada, Analytical Studies Research Paper Series, No. 219 (
pdf).
