| Under your scenario, if a bank can lend out $900 for every $100 it collects in deposits, then won't we see runaway monetary growth? Consider the following scenario: 1) Alice deposits $100 in Bank X 2) Bob takes out $900 loan from Bank X and deposits it in Bank Y 3) Charlie takes out $8,100 loan from Bank Y and deposits it in Bank X 4) Eve takes out $72,900 loan from Bank X and deposits it in Bank Y 5) ... the cycle continues ... Also, I would assume that the interest rates that my bank pays for my deposits would be a lot higher if it could lend out 9x as much as it holds in deposits. The model where a bank lends out 90% of its total deposits still makes more sense to me and will result in $900 of total money created per $100 initially deposited. 1) Alice deposits $100 in Bank X 2) Bob takes out $90 loan from Bank X and deposits it in Bank Y 3) Charlie takes out $81 loan from Bank Y and deposits it in Bank X 4) Eve takes out $72.90 loan from Bank X and deposits it in Bank Y 5) ... the cycle continues ... and if you sum the geometric series, which in this case is finite, you get $900 of total additional money created for the initial $100 Alice put in. |
If Bob takes cash from Bank X and deposits in Bank Y, Bank X has less cash, that movement is reflected in their balance and in their reserves.
If Bob makes and electronic transfer from Bank X to Bank Y, at the end of the day, Bank X and Y have to clear their balance with each other. As some people have moved money in the opposite direction, from Bank Y to Bank X, the balance could be compensated.
If for some reason, people only retire money from a bank without never making deposits, you have a bank run and it's an indicator of mistrust in that bank.