There is no monetary hot potato in standard Keynesian and New Keynesian models. That is because those models make wildly implausible assumptions about people's knowledge. They assume people have perfect knowledge about how much money everybody else is borrowing from the banking system, and how much they are planning to spend from what they borrow.
1. Because their planned expenditures of money exceed their expected receipts of money. They "borrow to spend".
2. Because their desired stock of money exceeds their current stock of money. They "borrow to hold".
In aggregate, actual expenditures of money will always equal actual receipts of money. That is an accounting identity. But this does not mean that aggregate planned expenditures of money will always equal aggregate expected receipts of money. In aggregate, borrowing for the first reason ("borrowing to spend") will necessarily turn out to have been a mistake. Because actual spending must always equal actual receipts. But that mistake creates a monetary hot potato. It creates money that people hold, but did not plan to hold.
Let's consider an extremely simple monetary exchange economy. There are no commercial banks. A central bank chooses a rate of interest at which individuals can either borrow money from the central bank or lend money to the central bank, as they wish. A Wicksellian central bank with a horizontal LM curve. Most economists say that the stock of money is demand-determined under that assumption. I think they are wrong. There can be an excess supply of the stock of money. There can be a hot potato. The Law of Reflux, which says an excess supply of money must flow back to the central bank, is invalid, even in this case.
Start in equilibrium in a stationary economy with no borrowing from the central bank. The stock of currency is constant over time, and equals the desired stock.
Now let's ask what happens when the central bank reduces the rate of interest from 5% to 4%.
Two things happen:
1. The representative individual plans to spend (say) $100 per month more than he expects to receive per month. (The exact amount depends on the interest-elasticity of consumption and investment demand).
2. The representative individual plans to hold an extra (say) $10 stock of money. (The exact amount depends the interest-elasticity of the demand for money).
So in the first month, the representative individual borrows $110 from the central bank. He plans to borrow another $100 in the second month, and $100 again in the third month, etc.. But his expectations will be falsified by events, and his plans may change as a result.
At the end of the first month, the representative individual will be surprised to find that his receipts of money are $100 higher than expected. Because every other individual did the same thing as he did. He holds an extra $110 in his pocket, not the extra $10 that he expected to hold.
What happens next depends on whether he revises his expectations about future monetary receipts.
Suppose he doesn't revise his expectations. He thinks his extra monetary receipts were a temporary fluke, unlikely to be repeated. He now has $100 more in his pocket than he wanted to hold. He was planning to borrow $100 from the central bank in the second month. He now borrows nothing. He again plans to spend $100 more than he expects to receive in the second month.
At the end of the second month the representative individual is again surprised to find his monetary receipts are $100 bigger than he expected, and that he holds $100 more money in his pocket than he wanted to hold.
If he does not revise his expectations, or revise his planned spending, or revise his stock demand for money, this $100 hot potato of excess supply of money will continue to circulate forever, month after month.
Suppose eventually, at the end of the third month, the representative individual revises his expectations. He expects the $100 increase in monthly receipts to be permanent. He revises his planned flow of expenditure, and he revises his desired stock of money.
In the fourth month he plans to increase his spending by an additional $100 per month, which is again $100 more than he expects to receive. And his desired stock of money rises by (say) $20. He now holds $80 more money in his pocket than he desires, so he borrows another $20 from the central bank.
Once again he is surprised. At the end of the fourth month his receipts turn out to be $100 bigger than expected, and there is $100 more in his pocket than he planned and desired. And until he revises his expectations again, this $100 excess supply of money will continue to hot potato around the economy, even with no more borrowing from the central bank.
And so on.
The hot potato process continues until the central bank raises the rate of interest again. [Update: or people stop borrowing to spend for some other reason.]
Any money that people in aggregate borrow for the first reason, "borrowing to spend", will create a permanent hot potato that falsifies their expectations. Any money that people in aggregate borrow for the second reason, "borrowing to hold", will not create a hot potato, and will not falsify their expectations.
Standard Keynesian and New Keynesian models do not allow what I have just described above to happen. They assume an equilibrium in which aggregate planned expenditure is always equal to aggregate expected income for the current period. If the central bank cuts the rate of interest, both planned expenditure and expected income rise by the same amount. In aggregate, people never borrow money to spend. They only borrow money because they want to hold more money. In those Keynesian models, people are never surprised in aggregate to find themselves holding more money than they desire to hold. There is no hot potato in keynesian models
Does the standard Keynesian assumption of continuous equilibrium between aggregate planned expenditure and aggregate expected receipts make sense? Is what I have just described a rational expectations equilibrium? It might be. Even if every individual is fully rational, he does not know that a cut in the interest rate from 5% to 4% will lead to aggregate planned spending in excess of expected receipts, and falsify his expectation. For all he knows, the central bank might have cut the interest rate because everyone else planned to spend less, and the central bank was adjusting the actual interest rate down to match the lower natural rate of interest. For all he knows, aggregate planned spending might equal aggregate planned receipts at 4%.
The standard Keynesian and New Keynesian model assumes that aggregate planned expenditure always equals aggregate expected receipts. People in aggregate never borrow money to spend. They only ever borrow money to hold. It therefore makes totally implausible assumptions about individual agents' knowledge. It assumes a rational expectations equilibrium in which each individual knows the current plans and expectations of all agents. It is that wildly implausible assumption that ensures there is no monetary hot potato in standard Keynesian and New Keynesian models.
If agents in aggregate "borrow to spend", actual income will not equal what agents had expected when they planned their expenditure, and actual money balances will not equal desired money balances. The economy will be "off" the IS curve, and also "off" the LM curve.
Update: basically, in Keynesian terms, the hot potato story is a story about the process by which the economy moves from one ISLM equilibrium to another. It is a story of "false trading" (or "false borrowing" in this case) -- because people's expectations and plans do not adjust instantly to the new Hicksian temporary equilibrium where those plans and expectations are mutually consistent, so they make trades they would not make in equilibrium. And when false trading happens, and stocks change as a result, there is always the chance that the economy may move to a different equilibrium, or even away from equilibrium. People may not always succeed in learning the new equilibrium. And their learning may itself change that equilibrium.
(I think this is related to what the interwar pre-General Theory monetary economists were on about. Robertson, Hayek, etc.)
(And I thank the commenters on my last couple of posts, especially the critical ones, for forcing me to try to think this through. It was hard.)