Except that banks don't really loan currency that way. That is to say, banks don't really loan out their deposits of Federal Reserve Notes, but instead issue loans by expanding the money supply. Say that same $10,000 in hard currency is deposited into bank A, but this time, the bank doesn't just loan out $9,000, keeping $1,000 in reserve. Rather, the $10,000 in deposits is the reserve from which the bank can then loan out an additional $90,000. It is still meeting its requirement to keep 10% in reserve. So if the deposits don't account for the loans, where does that loaned "money" come from? Why, it is simply created out of thin air! The bank punches some keys on a computer and—POOF!—extra digits show up on a borrower's account statement. Neat trick, huh? That's the magic of fractional-reserve banking.
The bank hasn't turned on some printing press and created more Federal Reserve Notes to place in its vault to represent the amount of the loan, the $90,000. The "credit" was just signed into existence when the borrower put his John Hancock on the loan agreement. So the "money" that was "borrowed" never existed in the first place. But the borrower can still gobuy a car or a house or whatever, because the seller will accept those digital numbers being transferred to his or her own account as value for the item sold. Works great, doesn't it? Well, sure, except that the borrower now owes the principal plusinterest on the "money" the bank "loaned" him by creating it out of thin air, and, of course, if he doesn't repay it, the bank will take the house he bought with the "money" he borrowed—which is to say, in either case, that the borrower must repay something of real value representing the fruit of his labor in return for having borrowed something of no real value representing no labor or production.