[I can't decide whether this thought-experiment has any policy relevance. I hope not, but who knows what the future will bring? It came to my mind after reading a (more policy-relevant) post by Simon Wren-Lewis.]
1. Helicopter money is when the central bank prints money and puts it in people's pockets. Vacuum-cleaner money is the exact opposite; it's when the central bank takes money out of people's pockets and burns it.
2. Paying 5% interest on holding money is when the central bank pays you $5 per year for every $100 you have in your pocket (or in your account at the central bank). A 5% tax on holding money is a negative 5% interest rate on money; the central bank charges you $5 per year for every $100 you have in your pocket (or in your account at the central bank).
3. Let's put 1 and 2 together. "Positive interest helicopter money" is when the central bank prints new money to pay you interest on the money in your pocket. "Negative interest vacuum-cleaner money" is when you pay the central bank interest on the money in your pocket and the central bank burns it.
That was just to explain the terminology. Now let's do the thought-experiment.
Suppose a central bank is broke (Simon says "policy insolvent"). It has no assets. And it can't borrow against its future seigniorage revenue from printing money because there won't be any. And the government won't bail it out. But the inflation rate is in danger of rising above target (which is why it won't print money and earn seigniorage revenue). The central bank cannot buy back money (do an open market sale of bonds, or "negative QE", in the silly modern terminology) because it doesn't have any assets to sell. What can it do?
It can implement negative interest vacuum-cleaner Gesellian money for broke central banks. It imposes a tax on holding central bank money, and burns the money it collects in taxes.
"Now hang on" you say. "Isn't the idea of a tax on holding money to reduce the demand for money (increase the velocity of circulation) and so raise the equilibrium price level? Isn't that why Gesell suggested his tax?"
You would be quite right. And that is what would indeed happen if the money collected in taxes on money were put straight back into circulation (as Gesell proposed). It would raise the price level, so it wouldn't be a very sensible policy if we wanted to lower inflation. But we are going to burn it instead. So a 5% tax on holding money means the stock of money will be falling at 5% per year, which will have a disinflationary effect.
The ongoing negative 5% growth rate in the stock of money causes the equilibrium inflation rate to decline by the same 5% (5 percentage points). But the 5% tax on holding money, and 5% lower expected inflation rate, have equal and opposite effects on the real stock of money demanded M/P, so P does not jump, but instead P declines at 5% per year, along with M, relative to what would otherwise have happened. It's just like an ongoing annual 100-for-105 reverse stock split. Every year you hand in 105 old shares, and get 100 new shares in return, so the shares get 5% more valuable each year (relative to what would otherwise happen).
(Just for once, the Neo-Fisherites would be right, albeit for the wrong reasons.)