Monetary policy using the asset side of the Fed’s balance sheet

An interesting trend has developed in the Federal Reserve’s asset holdings, a trend that the newly created term auction facility is designed to accelerate.

The Fed seems to have become in some people’s minds an institution with immense and mysterious abilities to solve all our current economic problems. But we should not forget that its primary power and responsibilities come from the simple act of controlling the nation’s supply of money. The Fed creates new money by purchasing assets from private sector holders, paying for these purchases by adding an accounting entry to the Federal Reserve deposits in the recipient’s bank. These deposits can then later be used by the bank to make a withdrawal of currency, so that ultimately the Fed’s asset purchases end up as currency held by the public.

The overnight interest rate on Federal Reserve deposits lent from one bank to another is known as the federal funds rate. Because this rate is extremely sensitive to the quantity of reserve deposits in the system, the Fed can (and largely does) think of its money-supply decision solely in terms of this one interest rate. To the extent that we instead summarize monetary policy in terms of the quantities of assets as opposed to levels of interest rates, we would almost always be focusing on the liability side of the Fed’s balance sheet, namely the reserve deposits credited and currency outstanding.

The table below revisits a summary of the Fed’s balance sheet that I employed in a discussion in September. I noted then that despite the aggressive operations by the Federal Reserve in August, as of mid-September there had been no effect on outstanding currency. Updating those calculations, we see that since September, there has been a $5 billion increase in currency in circulation, but this is hardly something to alarm monetarists (whom I define as those who put more emphasis on the supply of money rather than the level of interest rates), because there is always a surge in currency in November attributable to a big increase in demand for cash during the Christmas shopping season. In fact, this year’s November currency gain was much more modest than in a typical year. In each of the previous three years, the currency increase from mid-August to mid-December averaged $14.5 billion, almost three times what we’ve seen for 2007.

Supply and disposition of potential dollars.
(Based on weekly averages, in millions of dollars). Data source: Federal Reserve Release H.4.1.
Week ended
Aug 8

Week ended
Dec 12

Change from
Aug 8 to Dec 12

Total dollars created




Treasury securities




Repurchase agreements




Less: Reverse repos




Discount window loans








Total dollars held




Reserve balances




Currency in circulation




Although there’s not much special on the liability side of the Fed’s balance sheet this season, the asset side is another story. In years past the Fed has implemented the Christmas currency surge with a big purchase of Treasury bills in November. Over each of the last three years, the Fed’s holdings of Treasury securities increased by $18 billion on average between the second week of August and the second week of December. This year, that magnitude decreased by $16 billion. The Fed has made a huge swap from Treasuries into repurchase agreements this season compared with its fourth-quarter operations in years past.

When I first commented on the switch from outright Treasury purchases to repos last September, I thought the explanation was that the Fed wanted to retain flexibility for dealing with financial market turmoil. As the switch has continued in stark contrast to previous November operations, however, it’s become clear that this was a deliberate policy decision on the part of the Fed. The Fed now appears to be using the choice of the assets it acquires through open market operations, rather than solely the consequences of those operations for the supply or reserves and currency, as an instrument of monetary policy. With repos, the asset that the Fed acquires is essentially a collateralized loan payable by the private counterparty to the transaction. The collateral for the loan could be Treasury securities and their derivatives, or certain debt obligations or mortgage-backed securities issued by Fannie Mae or Freddie Mac.

The Fed’s new term auction facility seems primarily designed to advance this objective further. Under the plan, banks can borrow directly from the Fed using even weaker collateral than is allowed for repos. Other things equal, this new borrowing would create new reserve deposits. To prevent this from having an effect on the supply of total reserves or the money supply, the Fed intends to contract its holdings of Treasury securities even further, announcing this week that it will redeem an additional $15 billion in Treasury bills.

Steve Cecchetti, former Research Director of the Federal Reserve Bank of New York and currently a professor at Brandeis University, is coming to the same conclusion:

At its most basic level, the TAF is simply another mechanism for doing open market operations. It seems like one of those technicalities that we normally ignore as being irrelevant. To understand why they are doing this, we need to think about the fact that the central bank can use operations to either change the size of its balance sheet or the composition of the assets that they hold. The first of these is what we teach and understand. It is the traditional policy directed at maintaining the federal funds rate at its target level. The second is different, and that’s what the TAF is about. This new mechanism is aimed at shifting assets from US Treasury securities (that are purchased for the permanent holding or taken in repurchase agreements) to some of the lower quality stuff that is accepted as collateral for discount loans. And the purpose of this is to try to reduce the risk premia charged in the one-month and three-month interbank lending markets.

So why is it the responsibility of the Fed to try to set not just the level of the fed funds rate but also the spread between the funds rate and the LIBOR rate? One possibility is that the Fed thinks that the market is currently overweighting the riskiness of short-term interbank loans. If so, that seems to be a different vision of the role of monetary policy from that articulated by Ben Bernanke in 2002:

I think for the Fed to be an “arbiter of security speculation or values” is neither desirable nor feasible.

A second possible justification is that the market is correctly pricing the riskiness of these assets, but that the chief risk involves an aggregate financial event that the Fed, through actions like the TAF, could mitigate or avoid altogether.

Could the Fed’s strategy work? Cecchetti writes that he hopes so, but has his doubts. And then there’s Paul Krugman:

anyone who expects the Fed or anyone else to come up with a plan that makes this financial crisis just go away will be sorely disappointed.

Technorati Tags: ,

16 thoughts on “Monetary policy using the asset side of the Fed’s balance sheet

  1. FTX

    Thanks for another very informative post on this issue Prof H.
    (Minor point – shouldn’t the table header be millions of dollars?)

  2. jg

    Question, Professor: what is the train of logic/mechanism by which TAF ‘…reduce(s) the risk premia charged in the one-month and three-month interbank lending markets…’?
    Nice presentation of the changes in Fed operations over this fall, sir.

  3. jodie

    There is a scramble for treasury collateral going into year-end. Three month T-bills are yielding under 3%. It doesn’t make sense for the FED to add liquidity by reducing treasury collateral even further.
    I applaud the FED for thinking outside the box and changing the collateral the use for liquidity injections.

  4. Anarchus

    Interesting to contrast this year versus Greenspan in 1998: In 1998 from August 12 to December 10 the Fed increased Treasuries Held Outright by $12.6 billion (from $441.906 bil. to $454.530 bil.) and increased Repo Agreements by another $0.91 bil. Note: The Fed’s balance sheet in 2007 is about 76% larger than in 1998, for perspective.
    I don’t think that repo activity is precisely pari passu with Treasuries Held Outright — the capital markets I live in are certainly behaving as if the liquidity isn’t as real — but at the moment I lack a rational argument for the claim.
    Last, am I the only reader that found Dr. Ceccetti’s “liquidity trap” argument unpersuasive? Ceccetti channels Keynes in observing that “Central banks have great tools for getting funds into the banking system; but they have no mechanism for distributing it to the places where it needs to go. The Fed can get liquidity to the primary dealers, but it has no way to ensure that those reserves are then lent out to the banks that need them . . . . Lowering the target overnight rate further would just mean providing additional reserves to the same primary dealers. Nothing makes me think that their failure to adequate distribute the funds they are receiving now would be addressed by simply giving them more.”
    Well. First of all, if it’s true that in 3 of 10 trading days since the crisis started the max Fed Funds rate has traded ABOVE the discount rate, it’s silly to argue that lowering the target overnight rate further would have no effect. The Fed might first try injecting enough liquidity to put the dang FF at its target and keep the max trades under control in the first place.
    Money is fungible and it’s a commodity. Let’s get the Hill-acopter away from Clinton in Iowa and deploy it properly in the capital markets and put this fire out NOW, before it gets away from us.

  5. JDH

    jg, I believe the theory may run as follows. A large quanitity of bank holdings are illiquid in the sense that their market would exhibit a huge bid-ask spread. The risk premium on loans to banks in part reflects the risk that banks would be forced to liquidate these assets at unfavorable terms and hence risk default on the bank’s debt. By accepting these assets as collateral, the Fed is reducing this liquidation risk and hence, the hope is, reducing the risk premium the bank would pay to borrow short-term funds.

    On the other hand, if you accept Krugman’s premise that this is at its heart a solvency rather than a liquidity problem, then you wouldn’t expect such Fed operations to do any good. We’ll see if the TED spread comes down any with next week’s TAF operations.

  6. RebelEconomist

    Interesting observation that the Fed is not adding much more base money than usual for the time of year, but is doing it in different ways. Actually, coming from the UK, where repo is the mainstay of monetary policy operations, it seems odd to me that the Fed uses outright asset purchases so much. Repo is surely a more appropriate instrument for seasonal addition of base money anyway.
    Unfortunately, this still leaves my question as to why LIBOR trades so far above Fed funds. What stops a bank borrowing in the Fed funds market and drawing on the resulting funds in its settlement account at the Fed to make a deposit at a higher rate in the interbank market?
    I am a bit of a geek when it comes to monetary policy application, and I think that if there is one good thing to come out of the present turmoil, it is that academics will no longer “ignore” such “irrelevant” “technicalities” – and Ceccetti was research director at FRBNY!

  7. jodie

    The key is term funding. Over night libor isn’t that far off from Fed Funds. Its more the term libors that are trading far wide from Fed Funds.
    Also the Fed Fund market is over the counter so you know your counter party. There are practical limits to how much you can borrow in that market.

  8. jp

    Professor – Since we’re on the topic of outright purchases I’ve always wondered… does the Fed’s use of outright transactions help control the Fed Funds rate or is it just temporary repo transactions that control it? Would the Fed typically conduct an outright transaction at or close to the target for the overnight funds rate or does it have some latitude? Is an outright transaction still considered a repo even though it is permanent?

  9. JDH

    jp, our conventional understanding was that it makes no difference for the fed funds rate how the reserves get into the system, whether by outright purchase, repo, or discount loan. That’s why these latest developments are a little surprising.

    Historically, the choice between repo and outright holdings was based mainly on the term for which the Fed wanted the reserves added. To meet Christmas currency demand, you want to boost the reserves in November and take them back out in January, which is longer than the term for most repos so an outright purchase has traditionally been used. For purposes of targeting the daily rate, the Fed is just intending very short-term additions or withdrawals and so usually relies on repos (to put reserves in) or matched sales/purchases (to take them out).

    I don’t understand your last question. An outright purchase means the Fed buys the Tbill, and that’s the end of the transaction. A repo means that the Fed temporarily assumes ownership of the asset, which really just amounts to collateral for a loan.

  10. RebelEconomist

    JDH, it is interesting how different environments generate different views. In the UK and eurozone, repos of three or more months are routine in monetary policy operations, whereas there is comparatively little very short term government debt, so repo would be used to meet a short-term seasonal peak. I guess there are more treasury bills in the US, so the Fed can obtain their chosen maturity relatively easily.
    Jodie, I agree that term seems to be important, although I believe that LIBOR does consistently trade at a higher rate than Fed funds, which could not be explained by the supply of Fed funds being limited.

  11. GK

    Off topic : Shouldn’t the recession-probability index be updated for the Q3 data? I am sure the probability right now is much higher than 9.5%. Probably around 50%.

  12. PrefBlog

    December 17, 2007

    There has been a lot of commentary regarding the coordinated liquidity injection (discussed on December 12) led by the Fed. First to the plate was Stephen Cecchitti, who has written many high quality essays for VoxEU, the most recent of which was disc…

  13. JDH

    GK, we will wait to put out the 2007:Q3 value until the advance 2007:Q4 GDP figure is released in January. This strategy always makes it look like we’re behind in reporting, but we prefer the benefit of making sure the data and call are accurate before publishing the index.

  14. jp

    Professor, thanks for clarifying.
    I’m curious, why would a commercial bank sell a t-bill outright to the Fed when it can sell the same instrument into the open market? Does the Fed offer better prices?

Comments are closed.