Mankiw's Baseless Arguments
Greg Mankiw's recent blog post carries a rather risky title: "The Monetary Base Is Exploding. So What?" I really am trying to understand the viewpoint of the wide range of economists (including Mankiw, Paul Krugman, Scott Sumner, Mike "Mish" Shedlock, Bryan Caplan, and David R. Henderson) who think the dollar is not going to fall sharply in the foreseeable future.
But I've yet to see a convincing explanation as to how Bernanke (or his successor) is going to avoid large price inflation, given the corner the Fed and the feds have painted us into. Mankiw's latest post recapitulates many of the standard arguments coming from the "no worries" camp, so it's worth explaining their deficiencies.
The Exploding Monetary Base
Before quoting Mankiw, let's review what the fuss is about. The monetary base (sometimes called M0 or "high-powered money") is composed of (a) actual currency in the hands of the public, and (b) bank reserves, whether in the form of cash in the banks' vaults or on deposit with the Federal Reserve. The monetary base does not include checkbook balances held by the public.
In contrast to other monetary aggregates (M1, M2, etc.), the Federal Reserve can directly control M0 (i.e., the monetary base), at least within very broad limits. If the Fed wants to increase the base, it can buy assets like US Treasuries from dealers in the private sector and pay for them by writing a check on the Federal Reserve itself. The seller of the Treasuries then deposits the new check in his own bank account. His bank in turn takes the check and clears it with the Fed, so that the bank's reserves go up by the dollar amount of the check.
In our scenario, whether the bank makes additional loans or not, the monetary base has gone up by exactly the amount of the check written on the Fed. If the bank makes new loans to its customers, it can affect the total amount of checking deposits, but that figure isn't included in the monetary base. So commercial banks through their lending decisions can affect the broader monetary aggregates such as M1, and M2. But the Fed exercises strict control over the monetary base, M0.
Because of the truly unprecedented buying spree upon which Bernanke embarked in the fall of 2008, the Fed's balance sheet has exploded. By writing more than a trillion dollars worth of checks drawn on itself (i.e., out of thin air), the Fed has caused the monetary base to explode as well:
Naturally, many common-sense analysts are quite worried by the above chart. Not only does it signify that the alleged economic recovery of 2009 is bogus, but it demonstrates a serious threat to the dollar itself.
Yet Greg Mankiw shrugs off these worries. Now that we understand the context, let's examine Mankiw's points.
The Broader Monetary Aggregates Aren't Surging?
Mankiw first tries to defuse the hysteria by making a distinction between the monetary base and the broader aggregates:
It is true that the monetary base is exploding.… Normally, such [a] surge in the monetary base would be inflationary. The textbook story is that an increase in the monetary base will increase bank lending, which will increase the broad monetary aggregates such as M2, which in the long run leads to inflation.
That is not happening right now, however. The broader monetary aggregates are not surging. Much of the base is instead being held as excess reserves.
Mankiw is basically right here. Historically, banks don't keep "excess reserves" in their vaults as cash or on deposit with the Fed. By definition, excess reserves are those above the legal limit established by a bank's total checking deposits (and other items). So if a bank is holding $1 million in excess reserves, it has the legal ability to grant its customers up to $1 million in new loans (which show up on the bank's liability side as checking deposits granted to the borrowers who can then spend the money). To repeat, historically, banks hold very little excess reserves, because they can put them to work earning interest by loaning them out to their customers.
However, these are far from ordinary times, and Mankiw is correct that the banks have let their excess reserves pile up, rather than extend new loans. In fact, there is now a "credit crunch."
However, even though Mankiw is right that the banks are sitting on their reserves rather than extending new loans, he gives the reader the impression that the broader monetary aggregates have been unaffected by the surge in the base. That is not true. Look at what happened to M1 and M2 since the fall of 2008:
If Mankiw doesn't think M1 (the blue line) has surged since the fall of 2008, I'd hate to put him in charge of Iraqi troop levels. Note that M2 (the red line) saw an acceleration in its growth trend at the same time the monetary base exploded, but it admittedly plateaued in early 2009 and is only recently back on an upswing.
Now, I want to offer my sympathies to the reader: believe me, my eyes glaze over too when I read financial reports that are chock-full of charts, each of which seems to reverse the "lesson" of the previous one. But please bear with me for just a bit longer.
One of the main arguments stressed by people in the deflation camp — and I'm not saying Mankiw himself would make this claim — runs like this: "In a modern economy with a fractional-reserve banking system, new money enters the system in the form of loans. It doesn't matter how hard Bernanke pushes, the banks are undercapitalized right now and so they won't extend new loans. There's no way for the money supply to grow until the economy recovers and the banks repair their balance sheets."
I hope the last chart above dispels this myth. The monetary aggregate M1 includes checkable deposits. Even though bank lending — both commercial and consumer — has fallen fairly substantially, M1 has gone through the roof. And the explanation isn't that there is some other component of M1 rising while checkable deposits are falling: this chart shows that demand deposits exploded at the end of 2008, then retreated, but have since resumed their upward trend and are far higher now than they were before the crisis.
How Loans Can Fall While the Money Supply Doesn't
My point here isn't to get bogged down in the various components of the monetary aggregates and come up with a theory of why some have gone up and others have fallen like a stone. All I want to do is get those readers who expect deflation to see that there is something crucially wrong with their argument. It is not the case that falling loan volume translates into falling money supply. The people claiming that Bernanke is literally incapable of expanding M1 and M2 (in the present economic environment) need to explain how he has managed to do so since the onset of the crisis.
Before returning to Mankiw's article, let me offer one possibility to make the deflationists see what could be happening: Suppose that we start with the Acme Bank, which is fully "loaned up," meaning that it has no excess reserves. The Fed then buys some bonds worth $1 million from one of Acme's customers, Bill the bond dealer. Bill takes the $1 million check and deposits it in his account. Bill's checking account balance goes up by $1 million, and Acme's total reserves go up by $1 million. Because of the roughly 10 percent reserve ratio, Acme's excess reserves go up by $900,000. In other words, Acme is only legally required to set aside $100,000 of Bill's new deposit to "back up" his checking account. The Fed's actions so far have increased the money supply (M1) by $1 million, because Bill's checking account now has that much more in it.
Because we assume we are in normal times, Acme spies a hot new real-estate development and gladly lends out the excess $900,000 to the developer, Shady Slick, for 12 months at an interest rate of 10 percent. Slick quickly spends all of the borrowed money on permits, union contractors, building materials, and so forth. Now M1 is $1,900,000 higher than it was before the Fed bought the bonds.
The money that Acme Bank lent to the real-estate developer, Shady Slick, is certainly "in the economy" pushing up prices. We can imagine that the $900,000 in Slick's checking account has now been disbursed around the county into the checking-account balances of various workers, shingle manufacturers, and local government building inspectors.
What if our story takes a sad turn? Suppose a federal bureaucrat is taking a nap on a park bench next to the hot real-estate development, and spots a rare beetle being crushed by a backhoe. The feds come in and completely halt development. All the prospective buyers who were on a waiting list for the trendy new lofts now pull out. Shady Slick is ruined and spends the rest of his days riding the L train and mumbling.
The troubles are not limited to Slick, for Acme is hurt too. Exactly one year has passed since Acme lent Slick the money, meaning that Acme's balance sheet showed $990,000 on the asset side as the loan that was maturing. But of course that entry is now bogus, and so Acme's accountants must replace it with a figure of $0. The write-down causes a direct hit to the equity held by Acme's shareholders, and may cause the bank to run afoul of regulatory capital requirements.
Now here's where the deflationists go wrong: I think many of them assume that somehow the money supply must have shrunk in the economy because of Slick's default on his loan. But that's not true.
The checking accounts of the union contractors, shingle manufacturers, and so forth still have (collectively) the $900,000 that Slick originally spent on their products and services. Thus, even though the total value of outstanding loans dropped by $990,000 because of Acme's write-down, the total value of checking or demand deposits didn't drop at all. This is true, even though it took a loan to originally push up the total value of demand deposits. (In contrast, if Slick had paid off his debt rather than defaulting, and then Acme didn't extend new loans, it is true that the money supply [M1] would have shrunk by $900,000.)
As I said before, I'm not trying to make an empirical case for what is currently happening in the US economy. I'm just pointing out that some of the glib deflationist arguments are incomplete. People who are expecting the money supply to collapse are overlooking some serious gaps in their argument.
Is the Monetary Base "Uninteresting"?
Let's return to Mankiw:
But, you might ask, won't the inflationary logic eventually take hold as the economy recovers and banks start lending more freely? Not necessarily. Recall that the Fed now pays interest on reserves. As long as the interest rate on reserves is high enough, banks should be happy to hold onto those excess reserves. That should prevent a surge in the monetary base from being inflationary.
Here is one way to think about it. The standard way of reducing the monetary base is open market operations. The Fed sells Treasury bills, say, and drains reserves from the banking system, reducing the monetary base. But consider what this means in the [current monetary] regime. An open market operation merely removes interest-paying reserves from a bank's balance sheet and replaces them with interest-paying T-bills. What difference does it make? None at all.
Both reserves and T-bills are interest-paying obligations of the Federal government (including the Federal Reserve). They are essentially perfect substitutes. The monetary base, however, includes one of them but not the other, largely for historical reasons.
The bottom line is that when reserves pay interest, the monetary base is a pretty uninteresting economic statistic.
I am astounded by Mankiw's performance. If I understand him, he's making an argument analogous to someone saying, "A good counterfeiter has no real impact on the economy. Here is one way to think about it: A merchant can sell his goods in exchange for authentic $20 bills, or in exchange for bills made with a laser printer in some guy's basement. So long as the counterfeits are indistinguishable from the real thing, what difference does it make to the merchant? None at all."
In particular, in the second-last paragraph quoted above, I think Mankiw overlooks quite a serious difference between Treasuries and reserves. The reason we include reserves as part of the monetary base is that they can act as the "base" of the monetary pyramid in our fractional-reserve system. That seems like a pretty good reason to me. In contrast, Treasuries do not form part of the base of the monetary pyramid.
When the federal government runs a deficit, that means it is spending more than it takes in through tax receipts. The Treasury can sell bonds to the public in order to cover the shortfall. Thus private citizens can lend the Treasury the money it needs to pay its bills. There is nothing inflationary per se about this. No new money is created in the system. The government has more money to spend, but the lenders have less money.
This is true even if a bank happens to buy the Treasury bonds. Owning such assets doesn't give the bank the legal ability to make more loans to its customers. If customers line up at the bank and want to empty out their checking accounts, the bank can't hand them Treasuries.
Mankiw is right that an individual bank doesn't care whether it holds a Treasury security yielding 1 percent interest, or the equivalent amount of reserves on deposit with the Fed, where they also earn 1 percent interest.
But that's not the end of the story. When the Treasury pays someone interest, it's not inflationary; it simply leaves less money (out of general tax receipts) available for other spending purposes. On the other hand, when the Fed pays interest on reserves, it necessarily increases the monetary base.
Thus, Mankiw's solution for dealing with unprecedented excess reserves is for the Fed to create even more reserves in order to pay bankers not to make new loans. Does that sound like a good long-term plan for the economy?
 The only real limit on the Fed's control of the monetary base occurs if the Fed wants to shrink it significantly. We can imagine a scenario — which may very well come true — in which the Fed expands bank reserves by buying trillions in assets, but when the it tries to unwind the operations, it can't sell the assets for as much as it originally paid. In this case, even if the Fed completely emptied its balance sheet of assets, there might be outstanding reserves that the banks held on deposit with the Fed. At that point, the Fed wouldn't be able to shrink the monetary base any more, because there would be nothing it could offer in exchange to people trading in their reserves.
 I know, I know, Walter Block and other libertarians think that the US Treasury is also a counterfeiter. But two wrongs don't make a right: it only hurts honest people even more when others start printing up phony dollar bills.