Thank you.
I'm very pleased to speak to you today on this crucial issue.
I'm an economist who specializes in research related to understanding gender and equality. I have experience using Statistics Canada data.
I'd like to speak first of its importance for economic analysis.
The census is the single most comprehensive source of data for analyzing the socio-economic situation of women and the issue of equality. Economists and other social scientists use the census to examine inequality in education, earnings, incomes, and other considerations. Work done on the feminization of poverty and the ongoing labour market inequality has relied on the census.
To analyze these kinds of issues, three kinds of data are essential in addition to individual income and earnings characteristics. The three kinds of data that are essential and that the census is particularly good for are as follows.
First, unpaid work, which has already been mentioned. It is essential to understanding women's economic inequality. Women’s organizations worked very hard to get questions on unpaid work into the census. Canada is a leader and a model in so doing, and the results have been invaluable for research on women’s equality.
The large sample and wealth of related variables in the census make it an important alternative to the time diary method from the general social survey. The quality of the summary data collected by the census has also been shown to be good in comparison with the time diary method. They're both good, but they're useful for different kinds of questions and they aren't replacements for each other.
The second kind of data that the census is good for in terms of analysis of gender inequality is the data at the household and family levels. You can't just look at the individual to understand the situation of women. Economic outcomes depend on household decision-making. Many other surveys only have data on the individual, which makes it harder to understand the processes that give rise to labour market and other outcomes.
The third kind of data in the census that is essential is data on ethnic origin, immigration status, language, geographic location, disability, etc., and other markers of social location. Women are not all the same, and the census allows one to analyze multiple dimensions of inequality.
In my own case, I've relied a lot on the level of geographic detail that's available in the census. For example, you can look at small communities, rural-urban differences, and you need that kind of reliable sample and detail to get at that level.
A voluntary survey will have a high likelihood of underrepresenting marginalized and vulnerable groups in the population. As the non-response rates have been shown not to be random, we're going to have a skewed sample from a voluntary survey.
I'd also like to speak about policy analysis and the importance of the long-form census.
Policies related to women’s equality require the availability of data to analyze the problem in the first place and demonstrate the need for policy intervention. Without that data, the case can't be made that the inequalities exist, nor can one design effective policy without understanding the underlying causal relationships. We can't plan for the future in policy areas such as health, education, and pensions without accurate demographic information.
Without data one cannot evaluate the impact of various policies on women. This includes policies aimed at addressing inequality and also policies that are aimed at other issues but that impact on women. So virtually every policy that we have does impact on women. For example, an analysis of budgets from a gender perspective is not possible, nor is an evaluation of the gender impacts of programs like EI or pensions, if we don't have this kind of good data.
Of course, advocacy on women’s equality relies on the data from the census. Without it, groups will have difficulty making their points and women’s ongoing inequality will become invisible.
Finally, I will comment on the impact of dropping the mandatory census on other Statistics Canada data.
It has been pointed out by Statistics Canada officials and other economists that the loss of the long-form census impacts the other surveys that the government is saying they can use. In terms of their sampling frames and the weights they use for analysis, the other surveys rely on the underlying population measures that are generated by the census. Those weights allow them to take account of a possible non-response bias in the voluntary surveys. Without reliable population measures, the whole thing becomes less adequate. Without the mandatory long-form census, Canada's position as an international leader in quality of data and research on women’s equality will be lost.
Finally, to make one last point, about privacy concerns, and speaking as a user of the data, Statistics Canada is also a world leader in terms of how hard it is to use their data in any non-confidential way. They are extremely demanding when it comes to protecting privacy, and researchers know this.
In conclusion, along with everybody else I feel strongly that it is very important, regardless of political persuasion, that we understand that good social science in every kind of policy decision requires reliable data.