Certainly. The research that was done in the fall for the ministry was the first time the databases that are available out there actually put pension plans and group RRSPs in the same study. They're among some of the first studies that look at what other assets people have, perhaps investing in their family farm or a small business, or in their home equity, and trying to factor. It's very difficult to factor assets into income, because it very much depends on the choices that individuals make. It becomes much more complex, but there has been some very good work done in NBER—the National Bureau of Economic Research—in the U.S. and other places to try to begin to build models like that.
As I intimated, there is some ongoing work at StatsCan that for the first time is trying to look at not the replacement ratio of individuals but the replacement ratio of couples. Very, very often, retirement decisions and retirement savings are not factors made by individuals; they're made by a couple. So if one half of a couple has a very good defined benefit pension plan, the other person may not spend nearly as much time worrying about retirement savings. They may put the education savings in that spouse's name.
So there's a lot of evidence that people make decisions about retirement as a couple, and we're just beginning to have data that looks at household level income and retirement savings and replacement ratios rather than looking at it for individuals. In terms of the individual data, if you have a homemaker who has no income, she or he could have been named on a spousal RRSP over the years, or maybe the spouse's pension is sufficient with spousal benefits that they've determined that's not necessary.
There are a lot of reasons where zero might be the right number if you're only looking at an individual and not at a household. So those are the kinds of questions we cannot get answered, given the state of the data.