Lack of census data will imperil policy-making
TheStar.com – Opinion/Editorial Opinion
Published On Sun Oct 31 2010. Munir Sheikh
Statistics Canada recently released population projections for Canada between now and 2036.
The population of seniors is expected to double from its current rate to be about a quarter of the total population by 2036. The ratio of working age population to seniors would drop from about 5:1 now to 3:1 in 15 years and fall further to 2.5:1 by 2036.
These projections confirm the consequences of lower fertility and mortality rates and the impact of Canadians living significantly longer.
This backdrop explains the need for continued immigration, which has been an ongoing Canadian reality. The logic for current and future immigration goes as follows: beyond satisfying other objectives, such as family reunification and humanitarian needs, immigrants provide an economic contribution to the country that helps deal with the economic needs of the indigenous population that, according to projections, is stagnating in number as it ages.
This raises a simple but natural question: have our immigration policies been able to meet this need? Given what we see in our population projections — that immigration would be the principal source of population growth in Canada in future years — this is a fundamentally important question.
Data available from the long-form censuses of the past provide information that can be the basis for research on this question.
Take, for example, the average income levels of groups of immigrants relative to native Canadians from each of the 5-year census cycles beginning with the 1975-79 group and ending with the latest 2000-2004 group.
Data from the long-form census offers two important conclusions. First, economic performance of immigrants, as captured by the earnings of immigrants relative to the native population, is on a downward trend. Second, this gap does not close even after immigrants have been here for up to 20 years. These are unpleasant facts.
A more detailed look at the data indicates some positive recent developments but they appear not to have taken hold. For example, the 1990-94 group of immigrants did marginally better than the 1985-89 group after working for six to 10 years.
Unfortunately, that improvement proved temporary. The 1995-99 group seems to have done better than the 1990-94 group soon after arriving, but their advantage also appears to be slipping away. To make matters worse, the performance of the 2000-04 group has fallen below that of the 1995-99 group.
Governments may look at this long-form census information to monitor the situation and attempt to improve it by changing policies. Some changes have been announced in the recent past, such as those focused on the economic class applicants, the Canada experience classand the provincial nominee programs. Would these changes be effective in improving outcomes and, if yes, by how much?
With the government’s decision to abolish the long-form census, it is not clear how one would get reliable answers to these important questions.
The National Household Survey (NHS) that would replace the census is not likely to give an accurate picture because experience tells us that different income groups do not respond to a voluntary survey in the same manner.
Any seemingly positive developments that the NHS may show on the relative performance of earnings of immigrants may be consistent with improvement in their outcomes or simply be the result of a higher response rate from those immigrants with higher incomes.
Furthermore, any data from the NHS would not be comparable to past census data.
The government has explained its position on the long-form census as achieving a “balance” between the needs for good data and individual privacy. Achieving this “balance” must mean sacrificing some data quality to provide a little more privacy.
The immigration example described above suggests that, in the absence of high quality census data, it may become considerably more difficult to deal with some of the fundamental economic and social issues we face.
Munir Sheikh is a distinguished fellow and an adjunct professor at the School of Policy Studies, Queen’s University, and the former chief statistician of Canada.
< http://www.thestar.com/opinion/editorialopinion/article/883374–lack-of-census-data-will-imperil-policy-making >