Art Demurs: "You don't break it up that way. The increase in debt shows one long, continuous trend. One trend."
Actually, they're both wrong.
It is very hard to look at an exponential growth curve and see any of the detail. Is the rate of growth changing? If so, which way? Too much for my aging eyes to discern. The answer, of course is to display the data on a log scale. Human mind-eye coordination is adapted to deal with straight lines, and that is what a constant growth rate provides on a log scale. Here is the right way to look at this data.
Contra Art's objections, we can now clearly see that there are distinct differences over time: straight line growth through the early-mid seventies; steeper through the mid-late eighties; then a rather sharp knee followed by another straight line with almost identical slope as the first segment until the Great Recession.
This does not negate the idea that the last few decades have been a speculative bubble. But, OTOH, it gives the idea no support at all.
Update: Prompted by the discussion in comments to this post, here is a graph of total credit debt divided by GDP. I think it's pretty astounding. Until 1980, debt was flat at about 1.5*GDP. Under Reagan it rose to about 2.2 times. It was close to flat under most of Clinton's term. Then it skyrocketed to almost 4 times under Bush. Compare to Mark's linked graph in comments. Debt becomes overpowering when it exceeds ability to pay. Mark's graph and this one show big moves in the wrong direction.