Let’s take an extreme scenario: the majority takes all their money out of circulation and puts it into gold. Gold’s the classic inflation hedge. Why shouldn’t they buy gold if they think the dollar is going to be debased? Obviously, gold isn’t going to absorb 100% of the cash currently in low-yielding financial instruments – but what if it did? What would be the economic consequences?
Well, that money would now be out of circulation. The Federal Reserve would have added a certain amount of money to the economy, by buying a variety of debt instruments. The sellers of those instruments, instead of putting the proceeds of their sales to work generating employment and income, stuffed it in a shiny yellow mattress. The Fed has expanded its balance sheet without boosting the economy in any way – but the market is aware of this expanded balance sheet, and aware that, eventually, it has to be unwound. And thus the net effect is contractionary.
Let’s look at this another way. You’ve increased the money supply by X. The recepients of this new money all proceed to buy gold. But they haven’t "taken the money out of circulation" – they bought something! It’s gold! It’s a shiny metal! More importantly, before these people bought the gold somebody had it! Now all those people have money, and they have to do something with it. Maybe they deposit it in a bank. Maybe they buy stuff. But there’s money out there somewhere. Same with Millman’s next example, a land price increase. If the price of land goes up because people start buying more land, the people who owned the land first now have more money!
Look, his broader point – let’s avoid inflating speculative bubbles – is well-taken. But the only way you can take money "out of circulation" is to lower the federal deficit or have it all deposited at the Fed or something. But otherwise it’s out there somewhere. Whether it’s being circulated at desirable velocity or put to productive use is of course endlessly debatable but investing in gold or land doesn’t make it disappear.