Here I survey how we got here, where things currently stand, and what it all means.
Let me begin by reviewing some first principles of what the Fed is all about. How did the cash currently in your wallet get there? You withdrew it from an ATM, perhaps. But these wonderful contraptions don’t just give you the green stuff for free– you had to have deposits in the bank to be able to withdraw the cash. You can think of your account with your bank as credits you can use to get cash whenever you want it.
But where did your bank get the cash? It likely has an account with the Federal Reserve System, which account, just like the one you have with your bank, shows a certain level of deposits that the bank has in its account with the Fed. Your bank can then go to the Fed and withdraw those deposits in the form of cash. So you can think of your bank’s deposits with the Fed as credits it can use to get cash whenever it wants.
And how did your bank come to have those deposits with the Fed? These deposits are something the Fed has the power to create out of thin air. This indeed is its primary power– the ability to create money. That’s a power that could be easily abused, so our system is set up to prevent the Fed from creating deposits willy-nilly. Specifically, the traditional operation of the Federal Reserve was to purchase assets such as Treasury securities from a private dealer, paying for them by simply crediting the dealer’s account with the Fed with new deposits. The Fed hasn’t created any wealth with this transaction, it has simply introduced a new asset (ultimately, money) and retired an old (the Treasuries that were formerly held by a member of the public are now held by the Fed).
Although private sector wealth is unchanged as a result of this transaction, there is one important implication for the Treasury. Before the Fed made this open-market purchase, the Treasury was obligated to pay interest on those securities to someone in the private sector. Now as a result of the open-market purchase, the Treasury is making that payment to the Fed. The Fed in turn returns those payments back to the Treasury. You can see those payments from the Fed back to the Treasury each month in Table 4 of the Monthly Treasury Statement.
But wouldn’t the Treasury want the Fed to simply buy up all of its outstanding debt, and relieve the taxpayers forever of that nasty burden? It might, but to do so would require so much new money creation that it would cause a horrific inflation. To avoid that, we have a careful separation of powers, asking the Fed to take responsibility for inflation and letting the Treasury worry about how to pay its bills.
The Fed could always use its power to acquire assets other than Treasury securities. For example, the Fed could make a loan to a private bank through its discount window. The bank receives the loan in the form of new deposits with the Fed, which again the Fed simply creates out of thin air. The receiving bank presumably used those deposits to pay somebody else, but that transaction simply transferred the Fed deposits to another bank, so that newly created deposits stay in the system until they are withdrawn as cash. The Fed in this case acquired an asset (the loan) whose value by definition exactly equaled that of the newly created deposits.
Alternatively, the Fed might want to add reserves to the banking system temporarily, to satisfy what it saw as a temporary liquidity need. Traditionally it would do so with a repurchase operation, in which the Fed temporarily takes ownership of an asset held by a dealer, and temporarily credits the reserves of the dealer in exchange. Essentially a repo is a collateralized short-term loan from the Fed to someone in the private sector.
There are some other categories of assets the Fed could acquire, and some other potential disposition of reserves it creates on the liabilities side. The chief among the latter that I will mention here is the Treasury’s account with the Federal Reserve. This traditionally was used by the Treasury for cash management of its receipts and expenditures. Some of the deposits that the Fed creates (for example, with a discount window loan) might have ended up being transferred between banks (as individual customers send checks to customers of other banks) and ultimately end up in the Treasury’s account (as income taxes get withheld, for example). Since the Treasury isn’t going to withdraw these funds as cash, they’re counted separately on the liabilities side of the Fed’s balance sheet.
There are a lot of other little categories we could discuss, but historically there was really just one big story– the Fed created deposits primarily by buying Treasury securities, and these ultimately ended up as cash held by the public. The left column of the table below summarizes the assets (factors supplying reserves) and liabilities (factors absorbing reserves) as of December 5, 2007. At that time, 85% of the Fed’s assets were held in the form of Treasury securities, and 89% of its liabilities took the form of currency held by the non-bank public.
|Dec 5, 2007||Dec 17, 2008|
|Factors supplying reserve funds||920.4||2346.7|
|Currency in circulation||819.3||877.7|
|Service and reserve balances||16.0||801.8|
|Factors absorbing reserve funds||920.4||2346.7|
|Off balance sheet|
|Securities lent to dealers||4.5||186.5|
Over the last year, however, there were some profound changes in the composition of the Fed’s balance sheet. These initially were dictated by the desire of the Fed to make more loans (which it thought it needed to do to alleviate problems in the credit market) without creating any new money (which it worried would create inflation). The way the Fed sought to achieve this was by selling off a large chunk of its holdings of Treasury securities, and replacing them with loans and alternative assets. These came in a variety of shapes and colors, but the two biggest categories at the moment are the Term Auction Facility, which essentially is a systematic program to encourage a particular volume of borrowing by banks from the Federal Reserve, and amounted to $448 billion as of last week, and currency swap lines, the biggest factor in the “other” asset category reported on the Fed’s H.4.1, said other category coming to some $682.4 billion last week. Up until September of this year, the Fed was implementing these changes without increasing the total level of assets on its balance sheet. The graph below plots the composition of the end-of-week asset holdings of the Federal Reserve over the last two years.
Beginning in September, the Fed decided it couldn’t afford to sell off any more of its Treasuries, but wanted to lend more and still have no effect on the money supply. To do so it needed to find a way to funnel the reserves created by the new loans it would make into categories on the liabilities side that would not result in more cash held by the public. The first such device was to reach an agreement with the Treasury for the Treasury to simply hold on to a huge volume of Federal Reserve deposits, some $484.6 billion as of last week. The way this worked is that two operations were implemented simultaneously. First, the Fed created a lot of new deposits, for example, $318.8 billion from the Commercial Paper Lending Facility alone. Second, the Treasury borrowed an additional half trillion from the public, forcing somebody in the public to send a check to the Treasury. In the aggregate, the reserves created by the Fed through the CPLF end up just being parked in the Treasury’s account with the Fed, with no creation of money. The graph below plots the composition of the liabilities side of the Fed’s balance sheet over the last year. By definition, the height of the line in the graph below is identical for every date to the height of the line in the graph above.
The second measure that the Fed employed to allow this ballooning of its assets was to start paying banks an interest rate on reserves that is exactly equal to its target for the fed funds rate itself, essentially eliminating any incentive for the banks to lend fed funds and encouraging banks instead to simply let excess reserves accumulate. Last week, banks were sitting on about $800 billion in excess reserves with the Fed, doing absolutely nothing with them. The Fed was in effect lending those funds in place of the banks. I have been quite apprehensive about this scheme, particularly now that we have reached a point where, in my opinion, the Fed in fact does want the money supply to increase so as to cause a little inflation. But the present arrangement makes it quite awkward for the Fed to do so.
And what’s the risk associated with the Fed’s new strategy? Back when the Fed held $800 billion in Treasuries, these were a liability of the Treasury and an asset of the Fed. In effect, the Treasury’s nominal obligation was one for which taxpayers would never owe a dime. Now that more than half of those securities have been lent or sold off by the Fed, and the Treasury has borrowed a half-trillion extra to make this work, that’s more than a trillion extra for which the taxpayers are potentially on the line. If the loans and other assets that the Fed has acquired with those funds do not make a loss, then all is still well and good. But if the Fed’s new loans do not perform, there won’t be a positive receipt in the Monthly Treasury Statement corresponding to interest returned from the Fed to the Treasury. In other words, the federal deficit will rise by the amount of the extra interest the Treasury owes on up to a trillion dollars in new debt.
And how about the Fed’s “free money” from the ballooning excess reserves? If those funds do start to end up as cash held by the public, then the Fed will need to worry again about inflation, in which case it has two options. One is to sell off some of its remaining assets (or fail to roll over some loans). In this case, the consequences for the Treasury are the same as above– that income from the Fed’s earnings is no longer coming back to the Treasury, and it’s as if the $800 billion in excess reserves was again replaced by direct Treasury borrowing.
The second option is just allow the inflation.
The bottom line is that Bernanke has made a gamble with something approaching 2 trillion. If the gamble wins, taxpayers owe nothing. If the gamble loses, taxpayers are committed to borrow a sum equal to any losses and start making interest payments on it.
For the record, let me reiterate my personal position on all this.
(1) I am doubtful of the Fed’s ability to alter interest rate spreads through the kinds of compositional changes in its balance sheet implemented over the last two years. Whatever your prior ideas were about this, surely it’s time to revise those in light of incoming data– if the first trillion dollars didn’t do the job, how much do you think it would take to accomplish the task?
(2) I think the Fed’s goal should be a 3% inflation rate. Paying interest on reserves and encouraging banks to hoard them is inconsistent with that objective, as would be a new trillion dollars in money creation.
I would therefore urge the Fed to eliminate the payment of interest on reserves and begin the process of replacing the exotic colors in the first graph above with holdings such as inflation-indexed Treasury securities and the short-term government debt of our major trading partners.