I tried to explain it several times, apparently unsuccessfully.
No you were sucessful some people just disagree with you. It happens.
Money supply is defined in different ways, but the best one is the total nominal value of whatever is used as a medium of exchange.
The best one according to you. However the presence of more money has an effect on prices even if it isn't DIRECTLY used in a medium of exchange.
For example I have $2K in my checking account. I purchase a $1K 6 month CD. Did the money supply shrink? Well according to the Federal Reserve the M0 has shrunk but the M2 remains the same. The $2K was never part of the M0.
M0: The total of all physical currency, plus accounts at the central bank that can be exchanged for physical currency.
M1: The total of all physical currency part of bank reserves + the amount in demand accounts ("checking" or "current" accounts).
M2: M1 + most savings accounts, money market accounts, retail money market mutual funds, and small denomination time deposits (certificates of deposit of under $100,000).
In order for FRB to affect money supply, the debt instruments issued must be accepted as a medium of exchange (for simplicity, we'll assume this occurs at par). Having zero maturity (being redeemable on request) is an insufficient condition.
This is your belief but not one universally shared by economists. It is a conclusion you leaped to without evidence.
In the example above the Fed includes my $1K CD in the M2. By your definition a CD wouldn't be included in the money supply. It is hardly used a medium for exchange. Why? Because even indirectly it influences my buying habits. Say hypothetically the bank was offering 20% interest on my CD and it was a $100K CD not a $1K. Every month I am getting about $1600 "richer" (in nominal terms). Now while I may not have access to that $1600 my newfound richness may cause me to spend MORE of my other money.
In other words the rising value of my M2 savings has an effect on how likely I am going to spend my M0/M1 money. In macro-economic view is the aggregate goods and services is growing slower and millions of people are also experiencing the same "richness" they are all more likely to spend (or spend a greater % of their available cash) and that will lead to higher prices (and reducing in buying power of 1 USD).
So the Fed includes CD in M2 calculation. CD (non demand deposits) increase the money supply.
To the OP:
The TL/DR version is money doesn't HAVE to be spent to affect the money supply. Everyday MOST money in the money supply isn't spent but even when not spent it influences buying/pricing decisions and thus is important to consider in macro economics. For example say you like Dr. Crypto soda and consider 1 BTC to be a good value. Not lets also pretend exchanges don't exist so if you want to drink Dr. Crypto soda you need to spend BTC. You consider 1 BTC to be fair. Now if the price rises to 1.2 BTC you may not be as likely to buy it however lets say your wealth (in BTC) also increased by 20% and your paycheck (in BTC) increased by 20%. Now the 20% rise in price is less likely to affect your consumption. Even if you only had 100 BTC in your wallet the fact that you have a 50K BTC savings bond (earning interest and increasing your wealth) in the First International Bank of Bitcoin will influence your willingess to accept higher prices. Likewise if the bank failed and you lost the 50K (money multiplier collapse) you may not accept the 1.2 BTC price as fair. You might not even drink as much Dr. Crypto even at 1 BTC.
To Loney Miner:
Even if your defintion WAS correct (which it isn't) then fractional reserve banking would STILL affect the Bitcoin money supply. Take a Mt. Gox code. Say you wanted to sell a video card for 100 BTC. Someone says I don't have 100 BTC but I have this Mt. Gox code for 100 BTC. Would you not take it? Of course you would. Not Mt. Gox (hopefully) isn't engaged in fractional reserve banking so there is no multiplier effect but lets say they did. Would you take it at par? Most people would but even if you wouldn't would you take it for 80 bitcents on the Bitcoin? 60 bitcents on the Bitcoin? 50 bitcents on the Bitcoin?
Even if you were so cautious you would only accept 100 BTC or 200 BTC Mt Gox code it affects the money supply (if Mt. Gox engaged in fractional banking). The amount you discount it (well aggregated discounting of entire ecnomy) would REDUCE the money multiplier effect but just because it is discounted doesn't mean it magically becomes "non-money".
DeathAndTaxes disagrees.
Of course I do.
Well me, most economist, and US Federal Reserve.