Yes.
As I indicated in previous answers, you have to look at this on a program-by-program basis, because the data is often unique to the individual program you're looking at, in terms of the results. You can't go with some generalized data collection that isn't necessarily going to get the information you need for all the individual programs. You need a strategy for each program. That will often involve investments for the program.
One of the critical limiters has been the ability to link it to broader Statistics Canada data, so you can understand its linkages to the broader population. As Frances indicated, Statistics Canada has taken a real step forward with some major investments to allow that to happen.
There's been very steady progress on disaggregated data, but, as the Auditor General's office indicated, we aren't at the end state on this. There's still much progress that needs to come. Again, as somebody who's managed programs in departments, I know there have been some systemic barriers to this beyond the resource one I raised.
Privacy is one. Your counsel will come to you and tell you the collection of that data would be a breach of people's privacy. That's why we issued the guidance to help guide departments through that thicket.
I would say another one, as I indicated in a previous answer, is checking with the communities themselves about whether they're comfortable with the collection of the data. In some cases, they won't be. In some cases, the right answer is to not collect the data—even though it will limit our full understanding of the implications of the program—because individuals feel there would be an undue government interest in specific characteristics they have, which they don't want when they interact with the program.
Those are some of the barriers. I'd say it improves every year, but we are not, in my view, going to get to an end state where every single program has fully robust, disaggregated data. There are going to be some areas where, for that third reason I mentioned, we may not be able to get the disaggregated data.