As we can see (click for a much larger graph) debt for the two sectors has gone up dramatically. Financial sector borrowing is mostly re-lent, thus the two sectors aren't totaled up to avoid double-counting (I think - I haven't found a plain-English explanation). However, that financial debt has grown so much is a symptom of the expansion of the "shadow banking system". These businesses borrowed directly in the various debt markets to fund lending instead of using deposits as banks have traditionally done.
From this graph we can see that the household sector has grown the most. State and local governments held their debt fairly level. Federal government publicly-held debt grew under the Reagan/Bush regime, then declined under Clinton, and went back up slightly under Bush. (This data set does not include 2008, which will show a dramatic spike for this sector.) Business borrowing has varied more than the other sectors, but it too has turned up recently.
Now at this point you might be wondering, "Where are the federal intragovernmental holdings?!?" Well, over the next few years they don't really matter. They don't have to be serviced and won't be redeemed, so their immediate impact on the debt markets is negligible. Starting in 2018-2020 they will matter, which is why the nation's various financial problems needed to be fixed yesterday. Yesterday? Wasn't Bush running deficits during the growth phase of the business cycle, making future budget problems much worse? Why, yes, he was and here's the results:
Before we go any further we should take minute to remember that not all debt is bad. Incurring debt in order to invest in something that increases future productivity can be a smart thing to do. Taking out a loan to build a factory which will make more widgets with less inputs is an obvious example. Taking out a loan to build a new road that cuts travel time in half is another. But incurring debt to increase current consumption is bad. If cashflow is sufficient to pay off the loan, it makes more sense to save until the consumption can be paid for in full. A 50" LCD might be neat to have right now, but it is quite unlikely to increase future productivity - probably the opposite. Similarly, taking out a loan to pay for repainting the stripes on the new road every year would be a bad decision.
People have learned to use debt in much more sophisticated ways than the simplistic examples above. And frequently the investment in question will increase productivity only by the smallest of margins. There are also plenty of cases where an investment at one price is productive and isn't at a higher price, but the crossover is hard to determine. Nonetheless, the standard should still be applied every time a company issues a bond or a consumer reaches for the credit card. Unfortunately, the standard has been frequently forgotten of late, to delayed but spectacular effect.