Greg Mankiw argues against the view that the rapid increase in the monetary base over the last 18 months will create inflation. I agree. However, I disagree with most of his arguments. I don't agree with his view that the monetary base is of little interest when the Fed pays interest on banks' reserve balances.
When the central bank serves as "Lender of Last Resort," it's key task is to respond to changes in the demand for base money. Banks have greatly increased their demand to hold reserve balances, and the Fed has properly responded by increasing the
quantity of reserve balances. Reserve balances at the Fed are about $1.1 trillion. If the demand for reserve balances should fall again, the Fed can and should reduce the quantity of reserve balances.
There was a very large decrease in the money multiplier. The Fed has increased bank reserves enough to offset that decrease in the money multiplier (and then some.) If economic recovery results in higher loan demand, so that banks choose to hold fewer reserves and make more loans, then the money multiplier will rise. The lender of last resort should reduce the quantity of reserves to offset that increase in the money multiplier.
Those who worry that the high level of the monetary base implies higher inflation in the future simply do not understand the role of lender of last resort. Perhaps they are too focused on the public finance view of money creation. From that perspective, the government creates money and spends it. This is generally done at a moderate rate, and the result in an inflation tax on real money balances. That an increase in base money might be reversed is foreign to this perspective.
Some have argued that the Fed will fail to decrease the quantity of base money as the economy recovers for fear of choking off the recovery. Such a mistake is certainly possible. While weekly money supply figures will make it possible for the Fed to offset a change in the money multiplier, if might fail to do so.
However, it should be noted that thinking about a "certeris paribus" decrease in the monetary base, and imagining "tight money" is an error. In particular, bank credit does expand as banks substitute loans for reserves in their balance sheets. And while interest rates should increase, this is in response to growing loan demand rather than a decrease in loan supply. If the Fed reduces reserves as the demand for reserves falls, the result is not a decrease in the quantity of money and supply of credit. It is rather than the increase is the supply of bank credit is less than it otherwise would be and the quantity of money doesn't decrease but rather doesn't increase.
There is a worry that the Fed will not be able to decrease reserves as the demand for them falls, because its asset portfolio is illiquid (or perhaps unsound.) Currently, the Fed owns a large portfolio of mortgage backed securities. The Fed owns about $909 billion in mortgage backed securities, which is about half of the $1.8 trillion of securities it owns. It currently owns about $18 billion in T-bills (about 1% of its total securities.) It owns $708 trillion in Treasury notes and bills. About $64 billion have maturity dates within the next year--the equivalent of T-bills. About $148 billion mature before 2012. On the other hand, it has bonds maturing in 2039! The concern that the Fed's security portfolio is illiquid is well founded. The link is here.
How does the payment of interest on reserves fit in? Paying interest on reserves increases the demand for reserves. The Fed could use the interest rate it pays on reserves as a policy tool. For example, if the demand for reserves should decrease because of increased loan demand, the Fed could pay increased interest rates on reserve balances to offset that change.
Still, that doesn't make the quantity of base money irrelevant. It is has always been true that both demand and supply must be considered, so that simply looking at changes in the quantity of the monetary base can be misleading.
Base money remains central because the unit of account, the dollar, is defined in terms of base money. The dollar price of base money is fixed at one dollar by definition. Given the nominal quantity of base money, the real quantity changes to meet the demand to hold it through changes in the dollar prices of all other goods and services. That interest is paid on base money simply is one more factor influencing its demand.
Where did Mankiw go wrong? Suppose one does macroeconomics by thinking about "the" interest rate paid on "bonds." If base money pays interest, it might be natural to imagine that it pays "the" interest rate and so it is exactly the same as bonds. But this is a mistake. Thinking about "the" interest rate is often misleading, and completely useless when money pays interest as well.
The Fed should pay interest on reserve balances--a floating rate that is perhaps 1/2 percent below the 4 week T-bill rate. Clearly, reserves would not be equivalent to "bonds." The rule would mean that the interest rate on reserves is never the same as the interest rate on "bonds."