I looked it up online (here), but I find it hard to believe that money literally "disappears". For example, if I buy 10 stocks of a toaster company for \$100 per share, and then the stock value drops to \$10, I've lost \$900, but the entities I bought from have gained \$900 collectively. I understand that money is supposed to be an abstract concept, but we should be able to trace where the dollars go.
So how does money just disappear when the stock market crashes?
Stocks are not money.
The valuation of a company - the market captialisation - is the number of shares multiplied by the share price.
The share price is the price people are willing to trade at right now. It does NOT mean all the shares have been traded at that price.
If a company issues 1 million shares, at a starting price of £10, the market cap is £10 million. They may only sell a few of these shares, so much less than £10 million actually changes hands. If the price rises to £20, only a few of the shares changed hands at that price - but the market cap is reflected in all the shares, and is now £20 million. People have commented on the flaws in this methodology.
The same happens in a crash. Only a fraction of the shares actually change hands, but the media report on the change in the market cap.
If you want an analogy, consider a house burning down. You buy the house for £200,000. Then it sadly burns down (and without insurance). It is now worth £0. However, no money has been created or destroyed; it's a loss of value of an asset.
If I buy 10 stocks of a toaster company for \$100 per share, and then the stock value drops to \$10, I've lost \$900...
You paid \$1000 to some counterparty for some assets worth \$1000 at time $t$, a fair deal. The value of your assets then dropped to \$100, resulting in a return of -90% at time $t+1$. However, the counterparty didn't lose anything, they still have \$1000 in cash and experienced 0% return over the same time period. It's possible that there are second order effects depending on how the trades are financed, but the activity of trading does not destroy or create any wealth.
The entities I bought from have gained $900 collectively.
So this is not quite right. The counterparty gained nothing by trading. This is because the value of the asset was worth the exact same amount as the value of your cash at time $t$. That the price at time $t+1$ would be much lower was unknown to you both. And of course the counterparty was in exactly the same situation when they originally acquired the asset at time $t-1$: the amount of cash that they paid was exactly the same as the value of the assets at $t-1$. Again, no wealth was created or destroyed through trading. Any return the asset generated was due to it being a share of a wealth-generating business.
The assessed value of the stocks goes down.
Non-stock related example: If my prized family portrait is assessed at 1000 dollars, and is later reassessed at 250 dollars. That value is lost, 1000-250=750 dollars lost. No money ever went anywhere, but I, and as result all of society, is now poorer by $750. This loss is because we believed the painting was worth a lot, and we were wrong.
If the stock market lost say 10billion dollars in value, it does not mean that 10billion dollars literally disappears (despite what the poorly written document you linked to suggests). But it is true that some money may have literally disappeared. This is because many shares are purchased with money borrowed from banks ("margin debt"). If the net amount of borrowing for share purchases went down i.e. some money was repaid, then that repaid money did literally disappear. This is all to do with how fractional reserve banking works - "Just as taking out a new loan creates money, the repayment of bank loans destroys money".
EDIT: After the stock market crash of 1929 the money supply fell by about a third. This was because huge amounts of money that had previously been invested in the stock market was in fact borrowed from banks. CORRECTION: I should not have implied that all of the one third reduction in the money supply was caused by a reduction in margin debt, but certainly part of the reduction was.
EDIT: The idea of the money supply falling after an asset price bubble bursts is not widely reported. After the crash of 2007/8 almost no media outlets have discussed any notion of money disappearing. This is because the media and the general public have little clue that it is even possible for the money supply to decrease. But you can read here the governor of the Bank of England telling UK politicians in Jan 2013 that QE was an effort to counteract the falling money supply: "What we were doing [through Quantitative Easing] is injecting money into the economy, and what the banking sector has been doing is destroying money [as existing loans were repaid]." ... "what we were doing was partially to offset what would otherwise have been an even bigger contraction."
Here is the narrative that I've arrived at:
1) Banks "create" money by lending recursively. This is called the "Multiplier Effect" which is a consequence of the "Fractional Reserve Banking System".
This artificial money can only exist if we assume that all lenders will pay back their loans.
2) When there is an economic recession the public stop spending as much (why this happens is another discussion), so borrowers stop getting revenue and default on their loans. The asset (a debt security) issued by the lending bank is now worthless and written off. There is a domino effect in that if that lending bank's assets decrease, the entity that this lending bank itself borrowed from also loses assets. The money that remains is equal to the "real" assets in the system (used to be gold) and this is why the economy is said to "reset" during a recession.
This begs the question of why allow a fractional reserve banking system instead of insisting on a full-reserve banking system. I would guess that it just makes for a very limited economy, even if it results in less spectacular market crashes. Apparently no country in the world sticks to a full reserve banking system.
