Recent Posts

Sunday, August 9, 2020

Who Sets The Yield Curve?

There was a debate on my Twitter timeline whether or not the Fed sets the entire yield curve. Since I am in the midst of writing a MMT primer, the party line is bubbling up in my mind: the yield curve is a policy variable. However, the exact mechanism is not entirely obvious, and the situation is arguably ambiguous.

The following Joe Weisenthal tweet started the discussion:

This got a response from David Beckworth:

I am closer to Weisenthal's view than Beckworth's, although I have reservations about the wording (as I will discuss). I do not want to go through a detailed response to Beckworth's comments here.

What We Can Say for Sure

One of the problems generated by Economics 101 textbooks is that readers have a tendency to refer to "the" interest rate. There is a multitude of fixed income instruments, and each has its own rate of interest/yield. At the minimum, we need to refer to a "risk-free curve." But even that is a simplification - there are a number of cash instruments that qualify as "risk-free" in the United States.
  • Rate of interest on settlement balances at the Fed ("reserves").
  • Rate of interest on inter-bank settlement balance lending ("Fed Funds").
  • Repo rates on Treasury collateral.
  • Yields/rates on Treasury bonds/bills.
(There are also derivatives based on these instruments that should be priced in a coherent fashion as well.)

Only the first instrument listed -- rate of interest on settlement balances -- is explicitly set by the Fed. The others are influenced by open market operations, with the objective of keeping the Fed Funds rate in its band. (The targeting procedure has changed over time.)

Specialists whine whenever spreads appear between these various interest rates at short maturities, and write breathless editorials about the Fed "losing control of interest rates!" However, anyone who understands the concept of duration realises that the returns implications are exceedingly uninteresting to anyone who is not a money market trader.

However, the sloppiness of Fed operations means that when we write "the Fed sets the short-term risk-free rate" implies that it might need to undertake various gymnastics to keep these various risk-free curve in line with each other. Nevertheless, the basic principle remains that short-term rates will stick near the Fed target, barring the U.S. Congress threatening to default (e.g., debt ceiling woes).

The question remains: what about longer-term yields?

Fed Officials versus The Fed Reaction Function

The issue with Joe Weisenthal's tweet is that he wrote that the yield curve is set by "the Fed," when in fact it is the Fed's reaction function, as measured by market rate expectations.

Given that the initial maturities on the risk-free curve are set by Fed officials, forwards reflect the market pricing for future settings. This market pricing may not align with what Fed officials think they will do in the future, which was the alleged "Conundrum" of Al Greenspan.
Figure: Bond Conundrum, My ....

The figure above demonstrates why the worry that bond yields were "too low" in 2005 makes no sense: realised rates over the subsequent decade were much lower than those "low" yields. The market is supposed to be a forecasting mechanism, are we to be shocked that its forecast was at least roughly correct, while a bunch of central planners who were completely blind-sided by the Financial Crisis were wrong?

Government officials shy away from predicting recessions (or that they have completely failed in their duty to regulate the financial system). As such, there is no reason to believe their forecasts of the future. The only thing that can be trusted is where they are setting the short-term rate now; where the short-term rate will be in the future depends upon future economic conditions -- which may be nowhere near where policymakers are publicly predicting.

As such, what matters is the reaction function of policymakers, not their forecasts. If we assume the term premium is zero, we can read out the market implied reaction function by reading off forward rates. (The term premium is obviously not zero, but it is not that large or variable, either.)

Yield Curve Control

The Twitter discussion bled into "yield curve control" -- the central bank setting the entire curve directly by putting price limits on Treasury securities, and intervening to defend those price limits. Doing so is certainly possible, but doing so would obliterate any information in the yield curve. Bond pricing would shift to guessing where the target range would move, coupled with carry analysis to see whether the target curve shape is coherent with projected short rates.

In such a case, Fed officials are arguably setting the entire yield curve, and there is no distinction between them and the reaction function.

(c) Brian Romanchuk 2020


  1. A technical question, from an MMT perspective - you list:

    Rate of interest on settlement balances at the Fed ("reserves").
    Rate of interest on inter-bank settlement balance lending ("Fed Funds").

    Is it correct to say that "reserves" and "Fed Funds" are both central bank money. The central bank alone is able to increase total central bank money, while commercial banks can lend central bank money among themselves?

    1. Yes. Fed funds is the market where private banks lend existing settlement balances among themselves. The Fed creates the settlement balances balances by buying assets or lending against assets.


Note: Posts are manually moderated, with a varying delay. Some disappear.

The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.

Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.