Not sure it was correct, but it's interesting how much of the focus traditionally was on the lending side of the equation, whereas today it's mostly on the borrowing side. Today people will look disapprovingly on someone who borrows more than they can repay, but traditionally people would look disapprovingly on someone who lent to someone who couldn't repay it. Regular debt amnesties every N years also seem to have been a common traditional feature, making it impossible for debt to persist beyond short-term loans (and discouraging lending to someone who you weren't very sure was going to repay it soon).

It does seem that approach has some structural benefits, in that it doesn't produce long-term webs of debt that will never actually be resolved, but with everyone carrying on as if they will. Forcing a debt reset every 7 years or 20 years or whatever keeps things somewhat honest, making sure people are really talking about assets they have, as opposed to assets they supposedly are promised but will in reality never get.

I think you are confusing two separate issues. One is the probability of debt repayment, and the other is debt lifetime. The two are not necessarily related.

Payday loans have a relatively low level of debt repayment (necessitating high interest rates) and also a very low debt lifetime (days to months). In contrast, many AAA corporate and govt bonds have long lifetimes and high repayment rates (until recently, mortgages did as well). There are also low repayment probability, long term debts (subprime mortgages) and high probability, short term debt (commercial paper).

I don't think debt lifetime is that important a factor. Long term debt makes sense to finance long term investments.