Over at Econlog, I have a new post discussing Ricardo Reis's proposal for a market-based price level target, which relies on shifts in money demand. (In contrast, my NGDP futures targeting paper contemplates using markets to adjust the money supply until NGDP expectations are on target.)
When thinking about these two approaches, it might be useful to compare a money demand shock with a money supply shock. We all know from EC101 that a shift in supply has the opposite impact (on value) from a shift in demand. Money is no different. Those who are used to taking a quantity theoretic approach to money might consider the following three examples:
1. Suppose there is a huge spike in the demand for US currency in foreign countries. That will increase the demand for US base money and, other things equal, will reduce the US price level. But you can also think of this sort of demand shock as reducing the supply of currency within the US. With less currency circulating in the US, the price level falls for standard "quantity theory" reasons.
2. Now let's consider an economy whose currency is of no interest to outsiders - say Canada. If Canada imposes draconian taxes, then the Canadian public might hoard currency as a way of hiding income from the government. That would result in a huge spike in Canadian currency demand. But you can also view that as a money supply shock. The supply of currency actively being used as a medium exchange, a.k.a. Canadian "transactions balances," would decline as hoarding balances increase. The quantity theory of money applies better to transaction balances than to all money balances.
3. With the advent of interest on reserves (IOR) and enormously bloated levels of excess reserves, the Fed has shifted somewhat away from a supply of money approach and towards a demand for money approach. On the other hand, the original monetarist model treated "high-powered money" as being equivalent to the monetary base. With IOR, the entire monetary base is no longer high-powered money. Instead, the currency stock is the new high powered money. So a change in IOR that impacts the demand for base money can also be thought of as impacting the supply of high-powered money. Thus, if the Fed suddenly cut IOR to zero, banks would try to get rid of some of their excess reserves, which would flow out into circulation, boosting the supply of currency, and hence the price level.
[Some people who understand accounting but not macroeconomics will tell you that banks as a whole cannot get rid of reserves - that reserves withdrawn from one bank will be deposited into another. Don't believe them. A monopoly bank could easily get rid of unwanted ERs by buying assets, then making bank deposits unattractive enough so that all the base money doesn't get redeposited into the banking system. The entire banking system is no different. People who treat macro as a branch of accounting will only confuse you - stay away from them.]
There's nothing new in this post, even more than a quarter millennium ago David Hume knew that an increase in money demand is equivalent to a decrease in the money supply:
If the coin be locked up in chests, it is the same thing with regard to prices, as if it were annihilated." David Hume - Of Money
PS. The thought experiments in this paper were merely to illustrate basic concepts. In the real world, the Fed would accommodate a shift in currency demand from overseas hoarding, or domestic tax evaders. Hence, there would be no significant macroeconomic impact.