Thursday, August 25, 2011

Unconventional Monetary Policy

Scott Sumner is great because he's extremely logical and impervious to facts. This makes him the go-to source for the nonsense that is Monetarism, because he will happily make the baldly ridiculous statements. One can then judge the value of the model for oneself.

This recent piece on MMT is a fine example. The key question gets isolated in the comments:
"How can the Fed put money into circulation if the people don’t want it?"
Scott never responds in this post, but in the past he's written about "unconventional monetary policy".

"Unconventional monetary policy" means the Fed targets some level of nominal GDP and keeps buying things until that level is met. Usually, the Fed buys and sells short term Treasuries to influence the clearing price of reserves in the overnight interbank market. The clearing interest rate in the overnight interbank market is the "Federal Fund Rate" and this intervention is the mechanism it uses to manage the level of excess reserves in the system (as it can drain reserves by converting them to Treasuries). Scott argues that the Fed can buy other things, like road repair services, bridge building services, etc. etc. and therefore hit any NGDP target it chooses.

"Road repair services" and "bridge building services" are very unconventional monetary policy instruments indeed -- but they are utterly conventional fiscal policy instruments. Indeed, much of the Obama stimulus (for reasons for politics and ignorant atavism) was focused on lame "shovel ready" projects of exactly this sort. So Scott is advocating what most of us think as "fiscal policy" but renaming it "unconventional monetary policy". Why?

I think most of us think of "monetary policy" as setting "interest rates", but Scott might be thinking of "monetary policy" as policy conducted by the monetary Authority, which is the Fed. Therefore, whatever the Fed does is "monetary policy" because, by Gods, it's being done by the Fed! This kind of makes sense, and at least we now have a transmission mechanism by which the Fed can, actually, hit an NGDP target -- direct purchasing of products and services.

But when stated in this clear way, I don't know how happy people would be if the Fed starts engaging in fiscal policy -- they believe that to be a matter for Congress. The Fed would pay for this by expanding its own balance sheet (the Fed simply marks accounts up and down to buy and sell things) so it would not show up in the Government deficit number, but there's no reason the Fed cannot use this same ability and move Congressional spending "off balance sheet" in the same way if it so chose.


Blogger NeilW said...

Yes, but the Fed is where the Gods of Olympus reside - and their Solomon like Wisdom is renown across the globe.

It is unthinkable to True Believers that such an auspicious organisation could ever do something as lame as buy the wrong thing.

After all it is Independent and that bestows magical powers of foresight.

12:01 PM  
Blogger Peter Drubetskoy said...

So, is Sumner an MMTer but just fails to realize this? You're not the first person saying that NGDP targeting and functional finance are similar. What do we need to do to make Sumner understand? That thread you mention is quite frustrating because it feels like two groups of people talking past each other and failing to address each other's points. But then such attempts are made, there seems to emerge some common ground. for example, the commenter W.Peden who likes to say that MMT is bunk seems now a bit more thoughtful in his comments after being exposed to some MMT logic. Or maybe I am deluding myself...

7:18 PM  
Blogger winterspeak said...

Neil: I'm not sure Sumner has even gone that far. I don't think he's ever said that "unconventional monetary policy" is "fiscal policy conducted by the Fed".

Peter: No, Sumner is a monetarist who has no idea how the economy works. The difference is that he'll take monetarism to its absurd logical conclusions, while others will fudge when it becomes obviously ridiculous.

6:35 AM  
Blogger NeilW said...

He won't. Cognitive dissonance won't allow it - as Bill Mitchell pointed out in <a href=">his post today.</a>

So they will keep talking in monetary terms even when they are clearly proposing buying stuff.

7:10 AM  
Blogger JKH said...

"Scott argues that the Fed can buy other things, like road repair services, bridge building services, etc. etc. and therefore hit any NGDP target it chooses."

I doubt he's ever said that. That's a purchase of current GDP output, which indeed is fiscal.

But where has he said that?

7:44 AM  
Blogger winterspeak said...

JKH: I hope I haven't put words into Sumner's mouth. I believe it is fair to say that he claims the Fed should stabilize the macroeconomy by "targeting some measure of nominal spending" although the mechanism remains opaque (unsurprising as he beleives banks lend reserves).

He calls this "unconventional monetary policy" and, to my recollection, it meant going beyond buying what the Fed usually buys (Treasuries) and buying other stuff as well.

But, I cannot find an actual definition of what he means on Google, so I am relying on memory. This is usually a bad idea, so I stand ready to be set straight.

5:00 PM  
Blogger Greg said...

But dont you see? If the Fed does it they wont be printing money and devaluing the dollar they will simply be replacing one asset ( a bridge) for another (reserves). It will be a simple asset swap!

The real question then becomes will they do a swap for MY house or will Jamie Dimon and friends be the only ones with access to this buyer of last resort.

12:42 PM  
Anonymous Anonymous said...

If I'm not mistaken, "Unconventional Monetary Policy" is merely using means besides adjusting short-term interest rates - which is the "Conventional" technique to accomplish the Central Bank's policy goals and the one that had historically been used in more normal economic conditions. Lowers rates yield higher expected NGDP paths, but obviously, there remains the problem of the zero-bound if even that low rate cannot produce the desired path in the short-term.

At this point anything else you do to increase NGDP (that is, "inflation") is "unconventional" and the most obvious way to accomplish it is by monetizing government debt and that's a "monetary" policy action, not a "fiscal" one - since the Fed isn't making any particular direct decisions on what the government should be doing with spending or taxation - only filling the gap with new dollars.

Until the Federal Reserve owns all the Treasuries in existence this particular method will not be exhausted. Even if it were exhausted, the federal government could lower taxes to zero and the entire budget could be monetized Zimbabwe-style. Eventually the Quantity Theory of Money's prediction of the price-level takes over. So it's really not necessary to explore even more extraordinary and unconventional potential purchases like normal goods and services, which, I agree, would be more "fiscal" in nature.

If a 5%-NGP path had been followed for the last 5 years, since 2006, we would be at $17 Trillion instead of $15 NGDP, so, an additional compounded inflation rate of 13% or what would have been 2.5% more each year. Whatever you think of the wisdom of pursuing such a policy, (which would have helped reduce the real burden of excess household and government debt in the economy, still a major problem) I don't think that kind of additional level of inflation would have been too difficult for the Fed to achieve, and it certainly wouldn't require any "fiscal-like" unconventional measures.

5:11 PM  
Blogger NeilW said...

"only filling the gap with new dollars."

Except that it is not filling the gap with new dollars. That is a myth.

In fact it is removing dollars from the system.

Bank reserves at interest are exactly the same as bonds at interest. They are financial liabilities of the government sector.

The only possible effect is if there is a lowering of the average interest rate due to the elimination of high paying bonds, and there is little evidence that occurs. And that effect has to be greater than the elimination of the *fiscal spending* stream of interest coming from the government sector.

Unless you get extra Private Sector credit spending (a silly idea in its own right when private sector debt is up over 300% of GDP), then messing around with interest rates can have no effect on the real economy (other than short term timing effects).

The entire comment shows total faith in monetary mechanisms and completely misses the simple point that all the effects on the real economy (reducing taxes and initial government spending) are fiscal in nature.

Unless the Fed buys real stuff it is pushing on a string.

12:28 AM  
Blogger winterspeak said...

glowplant: I know "unconventional monetary policy" is something OTHER than draining the reserves to manage the OIB market. The question is *what* other thing?

I don't think that swapping reserves (one Fed liability) for Treasuries (another Fed liability) is monetizing anything, nor do I see what gap it fills. It does not change the quantity of outstanding liabilities, merely re-orders their term structure.

The question isn't "wouldn't higher NGDP be better"? It clearly would be. The question is "how will altering the term structure of a fixed quantity of outstanding liabilities impact NGDP"? Or, "how can monetary policy actually impact anything"?

NEIL: The Fed can *not* be pushing on a string if it pays the "wrong" price for the stuff it buys. CBOC actually uses this mechanism, so there is a precedent.

10:31 AM  
Anonymous Anonymous said...

My understanding is that the Fed can purchase any financial asset, not any real asset.

But that doesn't stop Wall Street from creating "Infrastructure Bonds" (IB's) and starting a whole new bubble, with a willing Fed as buyer.

Whether the proceeds from IB's are actually used to hire anyone or not would be irrelevant, as long as we can put together a formula to show that their will be some income stream somewhere in the distant future, we good to go! We deserve a bonus! Maybe even donate to political parties!

Thanks, Scott!

2:06 PM  
Anonymous Anonymous said...

The Fed cannot purchase "any financial asset". It can only purchase a small group of assets as defined in the Federal Reserve act -- basically government guaranteed assets, agricultural paper, and bills of exchange. It can also lend to its member banks, but this too is prescribed by the act (e.g. section 23).

7:18 PM  
Blogger winterspeak said...


I am sure what you say is technically true, but does it hold up in real life?

For example, would you count the Fed purchases of residential mortgage backed securities among your three categories? How about the commercial paper facility they set up I think in 2008? Or the open swap lines between foreign governments -- is the Mexican peso now guaranteed by USG?

7:07 AM  
Blogger JKH said...


Sumner’s idea is to use monetary policy to target NGDP.

Not fiscal policy.

His desired monetary policy involves a combination of intervention in an “NGDP futures market” plus standard OMO’s.

If you haven’t already read about it, you don’t want to know about it.

I don’t recall him using the phrase “unconventional monetary policy” much.

9:13 AM  
Blogger winterspeak said...

JKH: Maybe I'm thinking back to one of the few times he was able to get Krugman to take notice of him. "Unconventional monetary policy" was his hook, and I haven't been able to get clear on what that means.

The NGDP futures market is a profoundly stupid idea.

"Credit easing" might work, but I call making a loan that won't be paid back a transfer, and now we're firmly in the world of fiscal.

10:46 AM  
Anonymous Anonymous said...


A distinction needs to be made between the collateral acceptable for loans and outright purchases of assets. There is more wiggle room in the former than in the latter. Hence the CPF and TALF. Those were loans that were subsequently unwound. I personally dont believe that the MBS program is legal. Nevertheless there is a difference between purchasing "pseudo" government guaranteed assets such as agency debt and being able to purchase any financial asset.

12:27 AM  
Blogger JKH said...


The Fed's argument for not "rescuing" Lehman has always been that L. didn't have sufficient collateral for continued/bridge support through the LLR function.

Apart from outright purchase of Treasuries (or Agencies more recently), the Fed's primary functional asset role is that of lender, not outright asset purchaser.

Interesting, the treasuries they purchase are supposed to serve as collateral against currency issued.

Although this latter interpretation would cause a loyal MMT'er to spew his soup.

5:22 AM  
Blogger winterspeak said...

windyanabasis & jkh:

Both good points which I accept.

Also, good to ponder actual reality when solyndra files for bankruptcy. I believe they enjoyed a half-billion dollar loan guarantee from the Federal Government, making their paper eligible for Bernanke.

Do we know if the bearded one loaded up when he could? Should Brian Harrison's salary be considered fiscal or monetary in nature?

Also, and you both may find this hard to believe, I'm not categorically opposed to the Fed being an asset purchaser.

9:49 AM  
Blogger W. Peden said...


Not delusional- just misinformed. You have been taught that everyone except MMT'ers subscribe to the money multiplier story; that the money supply is endogenous; and that MMT is a decent theory of how money works in a real-world modern economy. In my experience, those whose primary source of knowledge about a subject has been cranks are always surprised when they find that the non-crank views aren't as ridiculous as they were told. They may even confuse this realisation with successful persuasion.

For example, I've know people to say that I'm "accepting MMT" when I say that loans create deposits. The same people are usually amazed to find out that purchases of securities create deposits as well, or that a bank might sell cash for a security even if it doesn't want to lock up that cash in a vault forever.

9:25 AM  
Blogger W. Peden said...

* Accept cash for a security.

10:09 AM  
Blogger W. Peden said...

* Or hold so much cash as a proportion of its balance sheet.

11:37 AM  
Blogger sparc5 said...

@WP No, not only MMTers say loans create deposits

Comment on this site, New Economic Perspectives of KC, and Billy Blog to get highly informed responses to your question. Not every student is equally knowledgeable, and the teachers often respond.

Lastly all you did was rant without addressing the substance of this post. No wonder no one answered you.

9:43 AM  
Blogger Scott Sumner said...

I'm very late to finding this post, I'll just say I never said anything even close to what you claim I said about bridge building.

7:43 AM  
Blogger winterspeak said...



Clearly there is great confusion (at least in my head) about just what constitutes "unconventional"monetary policy.

Buying and selling Treasuries to control short term interest rates is conventional monetary policy -- yes?

Buying and selling Treasuries to control longer term interest rates is unconventional monetary policy -- yes?

Buying and selling non-Treasury Government backed securities, such as RMBS, is unconventional monetary policy -- yes?

The Fed buying solyndra government guaranteed debt would be unconventional monetary policy -- yes?

Buying a bridge -- no longer monetary policy!

Am I correct in this now?

7:54 AM  
Anonymous Anonymous said...


scott sumner believes monetary policy is accomplished through changing expectations. he doesn't really care about what actually happens in any of those transactions you've listed. it's simply good enough for the fed to say it is targeting a certain NGDP level and everyone will jump sufficiently. partly why he doesn't care about banks or really make much effort to understand financial markets

see nick rowe on this as well:

a beauty from rsj:

"As an aside, arguing that "CB communications strategies" is what is needed to address mass unemployment has to be the logical end-game of our modern day rabbit hole economics. Keynes would be weeping, and with good reason."

7:39 PM  
Blogger winterspeak said...

hasn't Bernanke already been quite open about wanting higher NGDP?

Was this considered "unconventional"? Is it working?

6:38 AM  

Post a Comment

Subscribe to Post Comments [Atom]

<< Home