Umm, because of math?

Let's say we have 100 people in the labor force. 10% of them are over 55 (10 people). 60% of those over 55 do not participate because of retirement (6 people). This means 6 people out of 100 (6%) is not in the labor force because of retirement.

Now, 15 years later, we still have 100 people in the labor force. 40% of them are over 55 (40 people). Of them, only 30% (an improvement of 30%) do not participate because of retirement (12 people). This means 12 people out of 100 is not in the labor force because of retirement. But our percentage of people not in the labor force because of retirement has still increased (6% to 12%).

Obviously I made those particular numbers up, but they clearly show that just because the percentage in one subgroup increase, that doesn't mean the percentage of the overall group cannot decrease because of that group.

So, yeah...math. In other words, your chart really had nothing to do with anything (like usual).

Ugh, my bad, wrong thread. I answered in the correct thread.