Paul Krugman Wonders Whether He Is Dumb. I Say: "NO!!"

He writes:

Fed funds question (seriously wonkish, and possibly dumb too): The target Fed funds rate is now 2.25%. Everyone expects it to be reduced further; Citi economists predict that it will be down to 1% by mid-year. But I have a possibly naive question: can the Fed really cut the Fed funds rate that far?... The Fed actually conducts monetary policy through open-market operations in Treasuries: the FOMC tells the open-market desk to buy or sell Treasuries from banks until the Fed funds rate is close to the target. Normally this puts Treasury interest rates close to the Fed funds rate, since one short-term loan to a very safe customer is a lot like another.

But right now Treasury interest rates are much, much lower than the Fed funds rate -- around half a percent on both 1-month and 3-month bills. Weirdness like negative rates on repos aside (I'm still trying to wrap my mind around that one), basically the Fed can only drive Treasury rates down by about another half-point -- which would still seem to leave Fed funds well above 1%.

How is it possible for the Fed funds rate to be higher than the Treasury rates? Well, one interpretation is that banks don't trust each other.... Fed fund loans, after all, are unsecured. In other words, the Fed funds rate may be more like LIBOR than the Treasury rate -- and it may be being held up by a premium similar to the TED spread.

Am I being really stupid here? Or is it possible that the fear factor will soon make it impossible for the Fed even to achieve its target on the interest rate it supposedly controls?

No, this is not stupid. As Clouse and Elmendorf (1997) http://www.federalreserve.gov/pubs/feds/1997/199730/199730pap.pdf write: "Because funds-market trading is typically not collateralized, the funds rate can also differ across borrowers according to their perceived riskiness." This has in fact been happening since last August--what the (average) fed funds rate is on any given day depends on who is doing the borrowing.

As Jim Hamilton wrote last August:

Econbrowser: Their objective is defined in terms of a volume-weighted average of all the transactions during the day (referred to as the "effective" fed funds rate), with the target for the effective rate currently declared to be 5.25%. The Fed usually makes at most one such intervention early in the day, and is then content to allow the actual fed funds rate at which banks choose to borrow or lend to each other fluctuate above or below the target. Often at the end of the day, and especially on the last day of the two-week period in which banks have to complete the satisfaction of their required average holdings of reserves, one will see the fed funds rate spike up, if some bank finds itself unexpectedly needing funds at a time when everybody else is finished trading for the day, or fall to practically zero, if some bank unexpectedly finds itself with excess funds that nobody else is interested in borrowing. As William Polley noted, this last week these intraday fluctuations have been particularly dramatic, with one trade last Wednesday (a settlement day) as high as 6% and another on the same day for only 0.25%.

Why does the Fed allow so much intra-day variability in the interest rate it is intending to target? One reason is that the Fed does not want to be in a position of subsidizing individual banks that choose to make unusually risky investments. If a bank knew that, no matter what it did, it could always obtain an unlimited source of funds at a 5.25% rate, the bank would have an incentive to borrow a huge quantity of such funds and use them to make higher yielding, but potentially quite risky, investments.... The Fed intentionally allows different banks, in different circumstances or at different times of the day, to pay a higher or lower rate for fed funds than do other banks, as one way of making sure that banks face immediate consequences of any extra risk-taking....

Of course, there is an inherent tension between the goals of serving as lender of last resort and making sure that banks are disciplined for risky behavior, and this tension is at the heart of the current policy dilemma facing the Fed. To help to achieve these twin objectives, the Fed has a separate tool, the discount window, through which it offers to lend directly to banks, temporarily giving them newly created reserves while holding high-quality assets as collateral for such loans. Again there have to be some institutional checks to prevent excessive risk-taking for banks using this facility. Historically, the Fed achieved this by placing additional limitations and regulatory oversight on banks that borrowed too much or too frequently at the discount window. Partly as a result of these, the discount window acquired a certain stigma...

Comments

[stan.jonas]

Forgetting the fact that FED FUNDS are almost always higher than Treasury rates...[a rather strange discovery for an economist concerned with monetary policy].. and why shouldn't they be?

IS THIS ANY WAY FOR ANYONE TO CHARACTERIZE WHAT THE FED DOES TO CONTROL THE FED FUNDS RATE?

"the FOMC tells the open-market desk to buy or sell Treasuries from banks until the Fed funds rate is close to the target. Normally this puts Treasury interest rates close to the Fed funds rate, since one short-term loan to a very safe customer is a lot like another"

GOD IS IN THE DETAILS...and the impression given.. is that at least this eminent Economist is not in control of those...
As a Good Bayesian.. I'm forced to say if he doesn't have the details right.. likely he doesn't have the conclusion right either..

[Frank Howland]

The current spread between T bill rates and Fed Funds is highly unusual. On my desktop I was able to readily compare the Fed Funds and 3 Month T Bill rates since 1954. Most of the time, the spread is well below 1%. The spread was above 2% in 1969, 1974, and 1980-81. In all those episodes nominal interest rates were much higher for short term instruments than they are today.

[anon]

The Fed controls the funds rate by a mechanism that is normally independent of the level of bill rates, repos rates, or any particular rate spread relationship against the funds rate. It does so by controlling the level of bank cash balances with the Fed. Because the Fed controls aggregate system balances, and because the banks must compete for their share of those balances in order to meet their individual reserve requirements, the Fed therefore controls the supply-demand conditions that determine the clearing rate for Fed funds. It is a simple economic process to increase reserves or decrease reserves when the effective (actual) fed funds rate departs from the fed’s target rate on a daily basis. This does not mean that the Fed won’t experience occasional difficulty in getting the funds rate to closely track its target on a particular day, but it does mean that there is an iterative process for the Fed to always be correcting for the deviation of the actual daily funds trading level from its actual target. In this sense, the Fed won’t lose control over the daily overshooting or undershooting. There can’t be a sustained deviation over or above the target rate. But this is all in normal times.

These are not normal times, of course. The problem is that the target funds level is gradually getting closer to 0, and treasury bill rates are getting very close to zero. There is a lower limit of zero for all of the bill rate, the funds rate, and the spread between the two. The question is whether or not this process at some point will cause banks to respond “uneconomically” to the Fed’s injection of excess reserves, should the actual trading rate for funds begin to drift higher that the target rate on a sustained basis. This might happen for example when rates reach the point where treasury bill rates are basically near-zero, and banks no longer consider the bill-funds rate spread to be sufficient to compensate them for banking system credit risk, and therefore essentially refuse to lend out their excess reserves at the target funds level.

This would represent a total seizing up of banks’ confidence in each other. But this seems very unlikely, since some level of excess reserves would surely cause those banks with individual excess reserve levels to lend to the most credit worthy of their peers – at least until the most credit worthy banks dropped their bid for funds to below the target rate level. The opportunity cost of not lending to the most credit worthy banks increases as excess reserves pile up, even if the funds rate is low. Spread economics should still work down to the zero rate level. Don’t get in the way of the Fed.

(In the other direction, there is no mathematical ceiling for interest rates, so the Fed has total control in terms of withdrawing excess reserves and forcing funds to trade at a higher level.)

[rhw]

Remember from Econ 101: the Fed can monetize just about anything. So if nobody wants to sell Treasuries, the OM Desk can buy mortgage backed securities -- they might find a few willing sellers there.

[Matt]

In other words:

The regional member banks are no longer master princes for the current time, they are dealing with a defacto, but disguised competitive monetary system with other the large funds asserting their monetary authority. The Fed chairman is being forced to act more as arbitrage than as senior member of a national banking system.

We need to systemize this process, make it the explicit standard, not an implicit one. Then we get what Paul wanted all along, for the member banks to get off their butts and do something useful.

[knzn]

rhw has a point, although (s)he doesn't state it quite right. The question is not whether anyone is willing to sell Treasuries: the Fed can offer a premium to redemption value (i.e. a negative yield), and with any premium more than a tiny amount, surely nearly anyone will be willing to sell. (Why would anyone settle for $1,000,000 in 3 months when they can have $1,002,500 today?) The theoretical problem is that the Fed may buy up all the Treasuries in circulation before it manages to get the funds rate down to its target. Once I actually say that, it sounds kind of implausible on the face of it: before the Fed has to start buying MBS, it can buy all the 3-month bills, then all the 6-month bills, then all the 2-year notes, then all the 5-year notes, and so on. Surely that's not really going to happen. Surely, with all those bank reserves, the demand for federal funds will go down to the point where the marginal lenders will be the courageous ones, who will be willing to accept 1 percent in the federal funds market.

[Barkley Rosser]

Well, it is my understanding that the favored market for actual intervention is the repo market, and that has been exhibiting negative nominal interest rates very recently. This does not strike me as a good sign.

[anon]

when the target was around 3.00% recently, the actual rate was around 3.19% -- that is, per my imperfect recollection -> the "actual" daily fed funds rate was higher than the "target" rate by a noteworthy amt --> what is the history of the variance between target and actual?

the fed has a different level of control over actual rate of the "fed funds" rate vs. the "discount rate" -- which rate do banks borrow more under - the fed funds rate or the discount rate?

[Barkley Rosser]

Historically the Fed has been pretty able to keep the fed funds rate pretty near its target most of the time by their various operations, again, mostly through repo market interventions. However, it is a rate that is very sensitive to short term market fluctuations and can vary wildly away from the target during a day. Remember, the Fed basically makes only one decision per day regarding what to do, during the famous "11 o'clock phone call" that supposedly transpires each trading day about that time between some folks in Washington and some folks in New York. So, events of less than a day frequency can get out of control In particular, aside from the goofy case of Dec. 1986 when the ffr supposedly flucutated between - 1/2 percent and about 18 percent, there have been triple witching days on which the ffr has shot up as high as about 100 percent for very short periods of time.

[Joe]

I've read a number of commentaries that state that the Fed is constrained by its balance sheet as to the amount of Treasuries or other securities it may buy. Is this correct? I thought the answer was no, and the Fed can monetize unlimited amounts of securities without Congressional or other approval.