Follow the impact of a $100 cash withdrawal through the entire banking systems, assuming that the reserve requirement is 10 percent and that banks have no desire to hold excess reserve.
If we didn't have a fiat monetary system, there would be no need for the government to now insist that a bank hold any percentage amount of currency in reserve. Well-run banks would back paper with gold or silver; customers wishing to exchange paper for metal would be able to do so. Poorly run banks would lend more than they have in reserve; customers wishing to exchange paper for metal would discover the shortfall, and a run on the bank would occur. The rigors of the market would punish it for its poor practices. Customers, of course, would only deposit money where they would be guaranteed convertibility. Bad banks and practices should be allowed to evolve away.
Yeah, I can't really argue with anything #2 says either. I will, however, point out that money might not end up back in a bank; if you are dealing with small businesses like the hundreds of little shops in New York, their "bank" might be a safe in the back room. In that case, the money will only change value as the currency as a whole changes.
#2 gives a great description of how this works. Essentially, the basis of your question revolves around banking practices and how they lend out 90% of the money they receive and keep only 10% of it as a reserve. The big debate at the moment, after the financial crisis, is whether banks should be legally enforced to keep more than that 10%.
As I understand it, part of the reason we are in such bad shape right now is that so many financial institutions are in fact wary of lending money. For a time it even actually looked as if the entire credit system would "freeze up." Although such total paralysis was avoided, it was apparently avoided very narrowly.
It is important, by the way, to point out that people will be paying this money back at interest rates significantly higher than their loan rates. So it wouldn't really be eaten away, but would lead to profits for the bank, which of course, is why they loan it out in the first place. Historically, the problems emerged when there were no reserve requirements, and no insurance for banks, so when investors attempted to withdraw their funds en masse, it triggered panics.
Eventually at some point, the initial withdraw of the $100 will be lost to the hold the different banks have on the money deposited and loaned out. The money will be eaten away, little by little, and nothing will actually be held in the hand of the consumer (down the line).
If every customer went to the bank and withdrew $100, would the bank have enough? It depends on whether or not the bank is holding that much cash, and how many customers that bank has. What if a customer took more? It would be a disaster.
Post two has given a great example of how this works. Basically, banks will hold 10% and the rest they will seek to loan out and if all works well, then the money will be in all different banks consistently. However, the important point to consider is the amount of reserve banks need to have in reserve. Some are arguing that it should be higher in case of defaults. This is the current debate.
Basically what happens here is that you get the multiplier effect. If I withdraw $100 from my bank, I will presumably then spend it. Assuming I spend it in one place, the person who I pay it to will deposit it in their bank. Their bank will hold on to $10 and will then lend the other $90. The person who borrows it deposits it in their bank, which holds $9 and lends $81. This goes on until the money is essentially all in various banks.