I have just two quick points.
If something is judged to be invasive, the government shouldn't approve that question for the census. That's a decision of the government. If the government believes some questions are privacy-invasive, let's take them off the long form. That's the first point.
The second point is about whether or not voluntary surveys are more likely to be responded to or not. We don't need to speculate; there is hard, empirical evidence that voluntary surveys have a vastly lower response rate. The long-form census had a response rate in the mid-90% range, and Statistics Canada's working assumption about the replacement survey is 50%. In everybody's experience around the world--the statistical societies have all written to the government expressing their view that this will render the long-form data much less usable, and for some purposes probably not usable at all.