First of all, one of the things to remember about the data I gave you on the current elderly is that they don't include any of the baby boom generation. The most recent age cohort you would look at would be people born in 1941, I think it is. Right? So we haven't really begun to catch up with the effect of the baby boom.
On the solution side, I realize that in editing myself at the end, the one thing I did leave out is that I think you should pay a lot of attention to the role of old age security in the retirement income system. It tends to get overlooked in these discussions, but it's very important in creating a base of non-income-tested income that is not directly linked to labour force participation. So I'll just leave that thought with you.
The other thing is that, in the third pillar, we are probably relying far too heavily on individual employers to serve as platforms for delivering retirement income. They're not effective platforms because most of them don't have the scale expertise to run a decent pension plan. Most of them will die before their employees do—that is, the companies will die, not the people.