Scott Sumner Can’t Stand When People Use Interest Rates to Gauge the Stance of Monetary Policy
He hits a familiar theme in the opening sentence from a recent post: “There are days where everything seems hopeless. I still find that 98% of the pundits I read don’t even understand that low rates don’t mean easy money.”
But wait, there is at least one other man who gets it. Sumner explains that an FT article gets the analysis right. Sumner provides a block quotation to show this “ray of hope” in understanding. It contains the following two sentences:
Despite German misgivings, low interest rates are no evidence that money is too loose: nominal GDP growth stutters along at less than 3 per cent, a clear sign that the stance is much too tight. In recent years the ECB twice made the mistake of raising rates too soon, and thereby punished Europe with a deeper recession and a worse fiscal crisis.
I have often complained when Sumner would wag his finger at people for equating the ECB’s low interest rates with easy money in one post, then in another post say matter-of-factly that the ECB raising interest rates meant tighter money. But I think this is the first time when I’ve seen someone (whom Sumner endorses on this very point) making those two points in back-to-back sentences.
It does not seem inconsistent to me to say
1) that low rates don’t mean easy money
and
2) that increasing rates will make money tighter than if rates hadn’t been increased.
The reason is that in the MM framework rates are tight/loose entirely in regard to whether they are appropriate for the desired growth rate of NGDP. But within this framework an increase/decrease will (other hing equal) still qualify as loosening/tightening of monetary policy.
The trouble is that low (historically) interest rates are said to not tell us anything… we must look at NGDP growth. In this view, you are scolded for thinking interest rates have any information of value regarding the stance of monetary policy – “only look at NGDP, stupid!”
Then, if rates move ever so slightly higher, we are told that higher rates do matter because that will cause lower NGDP growth (i.e., tighter monetary policy).
Huh?!
Scott Sumner doesn’t think interest rates don’t matter. He thinks that in isolation they can’t tell you whether monetary policy is loose or tight. That’s different from saying that interest rates have no effect on whether monetary policy is loose or tight.
So interest rate changes affect NGDP growth and NDGP growth affects interest rates?
Then we get this stuff where 0% is defined as too tight if NGDP growth is too low, but can’t it be that the FED being too loose is hurting NGDP growth (e.g., in a period of asset price inflation and low PCE inflation)?
At any rate, Austrians don’t have a problem with the idea that different interest rate targets will have different effects depending upon the “real” or “natural” conditions that exist at a given moment in time. But MMs take this further by saying we need price inflation and it needs to be in NGDP-measured spending or else the FED is too tight.
Interest rates in themselves tell us nothing about the stance of monetary policy. But in any given situation a increase/decrease in the target rate will generally be tightening/loosening of policy compared to a situation where the change did not take place.
This strikes me as rather intuitive, but perhaps I spend too much time on reading MM bloggers.
+2
Exactly, transformer: raising a midget’s height from 3’0″ to 3’2″ does not make him tall, but itt certainly makes him tallER.
Transformer and Gene, again, I encourage you to go read Scott’s comments in the blog post. He explicitly disavows what you are saying.
Scott Sumner believes that the rate increases meant tighter money given NGDP growth rates at the time. Were European NGDP growth rates rising to 10% per year, Sumner would castigate the ECB’s tiny rate hikes as loose money. It’s just like when, in the 1970s, a 5% FEDFUNDS was a sign of extreme monetary looseness, but in 2006-7, a 5% FEDFUNDS was monetary tightness so severe as to lead to declining housing prices. I say monetary tightness/looseness can also be usefully gauged by the yield curve and real M1, though these yield different results from the NGDP gauge.
Note: in practice, the Fed over the past three decades has used a Mankiw (how do you even spell that in Cyrillic?) rule to determine whether money was too tight or too loose.
As I don’t yet have tenure, I don’t really have time to read MM blogs, so I’m not really “up to speed” on this stuff. However, Couldn’t one make a Wicksellian argument that “tight” or “loose” policy depends on market rates relative to the natural rate? Is that what Scott’s trying to get at? Of course, NGDP doesn’t seem to be directly related to the natural rate of interest…
Look at this post by Sumner:
http://www.themoneyillusion.com/?p=6624
“So if there is a “natural rate of interest” it would be the rate where inflation or better yet NGDP is optimal, where the macro economy is in some sort of equilibrium. “
Thanks for sharing. This article was also interesting: http://journal.apee.org/images/9/9c/JPE_Fall_13_v2_part3.pdf
He concludes that paragraph:
“Instead, the focus should be on NGDP and inflation expectations. Get those variables right, and then interest rates will also be at the proper level.”
That’s pretty silly, honestly. It assumes that zero information is lost in aggregation and calculation of NDGP.
An assumption built into The General Theory and everything that came as a consequence… including most modern macro-economics.
Well that’s certainly the case. As I noted below, though, NGDP only (ostensibly) includes final transactions. It’s likely to be biased from the actual PQ that “exists.”
Say what?! Where is this assumption made?
It’s implicit, Gene. Interest rates will only be at the “proper level” if the Fed “gets [NGDP] right.”
Good luck accurately counting every “final transaction” (if that can be stated in a way that never creates ambiguity) that occurs in the country.
Additionally, why aren’t non-final transactions counted? Certainly money is used to make those transactions, implying that they are relevant for interest rates.
Sort of.
I think Sumner thinks that if NGDP growth can be kept on target then the actual rate will be close to the natural rate .
The way to keep NGDP growth on target is by adjustments to the size of the monetary base. This does not have to be done via interest rate targeting, but if this is used, then the target rate will always be set so that monetary policy will allow interest rates to trend towards the natural rate.
“This does not have to be done via interest rate targeting, but if this is used, then the target rate will always be set so that monetary policy will allow interest rates to trend towards the natural rate.”
That makes some pretty serious (un-argued) assumptions about aggregation and decentralized economic reality.
Probably.
I really meant that if the MM framework is correct , and this framework is used to set interest rates , then the market interest rate will tend to trend towards the natural rate.
There are undoubetly some pretty serious assumptions about aggregation and decentralized economic reality embodied in the MM framework that are beyond the scope of a short comment to address.
Well another major flaw is that NGDP doesn’t represent _all_ transactions, just “final” transactions.
If only he would explain it properly!
The word “Austerity” means… whatever the nearest Keynesian wants it to mean, right at this minute. See, easy?
“Tight Money” means… exactly the same as “Austerity”. Why do people get themselves all confused over this?
Assuming a zero percent rate of change in the total volume of money transactions (PT) then there should be no systematic tendency for interest rates to deviate too high or too low relative to the rates that would clear the markets for loanable funds on the basis of voluntary savings and investment. There may still be a tendency, depending on the banking system, regulations, how much of the transactions volume is actually determined by private decision makers, and so on, for there to be significant deviations above and below the proper rates. What’s going on inside the aggregate still matters, though, and that should not be forgotten.
From that it’s not too far a leap that the rates *relative* to the growth rate of PT matter, rather than the nominal rates themselves. Still, it’s not possible to know what interest rates would exist absent the growth rate of PT greater or less than zero, so it’s not possible to say whether the rate happens to be the natural rate *given* a PT growth rate. It might be too high or too low, but that can only become apparent after the fact.
Additionally, in a situation where inflation has become so much the norm, unfortunately the growth rates of PT consistent with long term contracts and those consistent with equilibrating markets for loanable funds, *cannot* be the same (the former need to be positive, the latter is a growth rate of zero). But this is the result of people adjusting to living under a central bank, it is not a flaw in the market.
Guys, for those of you trying to defend the FT guy’s comment about interest rate hikes… Please click the link and read the comments at Scott’s original post. In particular, a guy “Benoit” tried to make such a case, and Scott blew him up. Look at my comments later on in the thread, where I’m not snarky and spell out exactly why I think Scott tries to have it both ways with this stuff.
So yes, Transformer, in the grand scheme one could say that Bernanke gave us easier/looser money by cutting rates down to 0%, but that’s still not easy/loose money. However, that’s not what Scott’s position is. He thinks rates would be higher if we had looser money right now.
“However, that’s not what Scott’s position is. He thinks rates would be higher if we had looser money right now.”
He thinks that if we had appropriate monetary policy for the past few years then rates would indeed now be higher (closer to equilibrium natural rate). But I still find it hard to believe that he wouldn’t think that a rate hike would be tightening compared to the rate hike not happening.
I agree his answer to Benoit is a bit weird – but I think he’s making a rather pedantic point about Benoit’s saying ” everything was the same except rates being higher”, when obviously in a literal sense everything would not be the same if rates were higher.
-That might actually be true. Assume the Fed decided one day their inflation target was 10,000%. Interest rates would surely rise (possibly, even real ones!). But in any serious sense, money would be very loose.
“However, that’s not what Scott’s position is. He thinks rates would be higher if we had looser money right now.” Bob, the fact that if monetary policy was looser then rates would be higher doesn’t mean that all possible scenarios where rates are higher correspond to looser money.
Sumner believes that if the Fed did NGDP targeting, then rates would be higher, and he thinks that if the Fed reduced the money supply, then rates would also be higher. But he thinks the first option would be a looser monetary policy, and the second option would be a tighter monetary policy. What’s the contradiction in that?
“But he thinks the first option would be a looser monetary policy, and the second option would be a tighter monetary policy.”
No he wouldn’t, as he has stated before that money supply is not a proper measure. Sumner defines looser and tighter as higher or lower NGDP.
And if the fed increases its target rate this will have a negative effect on NGDP and so be tighter money.
Yes, exactly. The fact that it’s tighter money is a consequence of its effect on NGDP from Sumner’s point of view.
That contradicts your other comment where you said Sumner regards the Fed selling treasuries as tightening money.
Sumner regards NGDP itself as the measure.
You misunderstand me. I don’t mean that by definition, contracting the monetary base means tighter monetary policy. I’m just saying that if the Fed were to start selling short-term Treasury securities right now, Sumner would consider that to be tighter monetary policy.
I am not misunderstanding you at all. I am saying that you are making incorrect claims regarding what Sumner has written.
No, he would not consider thr Fed selling treasuries as tighter money. He would consider tighter money to be below target NGDP, and if the Fed isn’t targeting NGDP, then to Sumner it would mean a “significant” decline in trend NGDP growth.
He defines tighter and looser money in terms of NGDP, not money supply. If the money supply rose, but NGDP fell, then he would consider that as tightening monetary policy. If the money supply fell, but NGDP rose, then he would consider that as looepsening monetary policy.
*loosening
Major_Freedom, we’re in complete agreement regarding what Sumner’s definition of loose and tight money are. I’m just saying that from Snee’s point of view, if the Fed were to contract the money supply right now, that would reduce NGDP growth, and it would thus constitute tighter money.