Scott Sumner  

OK, now I'm really confused (the snake and the peg)

Why India's Ban Against Child ... How not to privatize...

Here's an imaginary dialogue:

Student: So the central bank sets the level of interest rates, right?

Professor: Not really. It would be more accurate to say that interest rates mostly reflect the condition of the economy. Higher inflation and/or higher real growth causes higher nominal interest rates. I.e. NGDP growth is the main determinant of rates.

Student: So the central bank cannot control interest rates?

Professor: That's too strong a statement. They can peg the short term nominal interest rate. But as a practical matter the peg is greatly constrained by economic conditions. There is a fairly narrow band of choices for the short term rate. If they move out of that band then the economy will become highly unstable, moving toward hyperinflation or hyperdeflation. Before that happened they'd have to move rates back within the safe band.

Student: So central banks can move rates slightly higher or lower, but cannot cause huge swings in nominal rates, such as what we saw in the period from the 1960s to the 1980s.

Professor: That's not quite right either. Imagine a time series diagram of the interest rate on 3 month government bonds. Now surround that time series line with a band that is about 1% wide. Now it looks sort of like a snake slithering across time. Let's call that band the snake. My claim is that central banks must (as a practical matter) keep their interest rate peg within the snake. They have a little room to maneuver, but not much. However you forgot one thing, the central bank can move the snake! So they can move rates a great deal over an extended period of time. But only by moving the underlying determinants of the snake, inflation and growth.

Student: Hah! I see where you are going with this. So if the central bank wants rates to rise much higher, it can push rates toward the top edge of the snake, bumping up against the top of the band. The snake reacts by shifting upward to reduce the "pressure", just as a horse responds to pressure put on it by the reins.

Professor: Nice metaphor, but unfortunately this snake is a bit perverse. You have described a Williamsonian snake, but the interest rate band is a Rovian snake. If you want the snake to shift upward, you lower the rate until it falls to the lower part of the band. I know this sounds confusing, but trust me, it's true.

Student. OK, now I finally get it. Within the snake itself the central bank has very little leeway. The current setting of rates is mostly determined by economic conditions. The central bank can also move its short term interest rate peg higher or lower within the snake, but only slightly. Over longer periods of time they can actually move the snake much higher or lower, but only by first moving interest rates in the opposite direction from where they wish to go in the long run. That moves inflation and GDP in the opposite direction, which moves the snake. By moving the snake they can move the peg by a large amount over an extended period of time.

And this means that although we cannot directly see the snake, we can infer that in the short run it is usually moving in the opposite direction from the interest rate peg.

Professor: You were exactly right until "And this means . . ." In fact, the interest rate peg and the snake usually move in the same direction.

Student: Ok, now I'm really confused. Didn't you just say that a change in the interest rate peg causes the snake to move in the opposite direction?

Professor: Good point, that was a bit of sloppiness on my part. What I meant to say is that an exogenous shift in the interest rate peg causes the snake to move in the opposite direction. But the snake usually moves for reasons unrelated to shifts in the central bank's interest rate peg. In mid-2008 most snakes were steadily falling, even as central bank pegs were pretty stable. When central banks did start cutting rates, the changes were mostly endogenous, they were simply following the market. As the snakes fell all over the world, the central banks followed them lower, with an unfortunate delay. So snakes and interest rate pegs usually move in the same direction, even though exogenous changes in the peg cause the snake to move in the opposite direction.

Student. Are there any situations where central banks cannot move the snake by adjusting the interest rate peg?

Professor: Yes there are. When most of the body of a Rovian snake falls below zero he is difficult to budge. That's because the central bank cannot lower the interest rate peg below zero. So they cannot lower it to the bottom portion of the snake, which is what is required to nudge the snake higher.

Student: So in that case there is nothing the central bank can do to nudge the snake upward?

Professor: I'm afraid you are wrong once again. Listen more closely to what I am saying. They can't move the snake by adjusting their peg. But the central bank still has several other options. They can switch to a Kimbellian snake, which has no fear of the number zero. But those snakes are exceedingly rare and inconvenient to work with. Or they can try an alternative approach. The central bank can whisper into the snake's ear that there are some tasty treats up around 3%, perhaps a box full of hamsters and gerbils. Snakes are very trusting, and respond well to this sort of forward guidance. Even Japanese snakes, which have an undeserved reputation for being cynics.

Student: But what if there are no tasty treats up around 3%?

Professor: You better be sure there are! If you make this promise and then don't carry through with it then the snake will never again respond to your sweet talk. In that case you'll have to give the snake a great deal of money up front, and tell him that he can keep all the money if the gerbils are not there as promised. Otherwise he won't budge.

Student: I'm afraid we are getting too deep into metaphors here. What is the real world counterpart to the offer of a great deal of money?

Professor: NGDPLT futures targeting.

Student: What makes that policy so attractive?

Professor: It's not easy to estimate the path of the snake when NGDP is expected to grow at 5% per year. The futures targeting helps the snake to find the path that will lead to 5% growth. More specifically, the futures program rewards all the millions of people who give the snake useful advice on which path is best. It makes their advice more credible.

The level targeting part of the policy helps in two ways. First, it helps keep the snake above zero when he's hit by a sudden shock, such as the crisis of 2008. This is one reason the Australian snake never hit the zero bound. Second, it makes it easier to keep NGDP growth on target, because temporary shocks to NGDP cause future expected NGDP growth to move in the opposite direction. And those changes in expected NGDP growth help move actual NGDP back on target.

Indeed with the futures market and level targeting system, central banks can now ignore interest rates entirely, and just peg the price of NGDP futures. The snake will still be there, invisible to our eyes, but he will no longer matter. Inflation won't matter either, as NGDP targeting will leave only optimal movements in the price level.

And NGDPLT futures targeting will allow us to stop teaching huge portions of macroeconomics. Because expected future AD will always be on target, we will live an a classical world where all the opportunity cost-based classical nostrums apply. No bailouts, no beggar-thy-neighbor, no paradox of thrift or toil, no fiscal stimulus. When the snake is finally tamed the world will be a much simpler and more pleasant place.

Student: I love stories with happy endings.

Comments and Sharing

CATEGORIES: Monetary Policy

COMMENTS (12 to date)
RPLong writes:

I was with you until the bit about "exogenous" versus "endogenous" shifts to the peg. I thought "peg" was supposed to mean policy. If it's policy, then it's endogenous; if it's exogenous, then it's not policy.

I'm obviously even more wrong than the hypothetical student, but I can't figure out why. Maybe I need examples:

(1) An exogenous shift that affects the peg.
(2) An exogenous shift that does not affect the peg.
(3) An endogenous shift that affects the peg.
(4) An endogenous shift that does not affect the peg.

Maybe some of these are impossible scenarios, but just knowing that would be a big help, because my current understanding is that everything that affects the peg is endogenous by definition.

Mike Freimuth writes:

Scott: This is excellent.

RPLong: Let me take a stab.

"Endogenous" and "exogenous" may not be the perfect words to describe the distinction (though it could just as easily be said that policy is by definition exogenous, this is a question of how you define the model). The issue is whether the CB is moving the peg in reaction to the snake moving or in an attempt to cause the snake to move. So if the snake moves down because of some exogenous (not directly related to CB policy)and this puts the CB's previous peg above the snake and they move it down to the upper edge of the snake, this will not move the snake back up.
On the other hand, if the snake does not move to begin with but the CB moves the peg down to the lower edge, it will (usually) cause the snake to move up. It's a question of whether the CB is the "tail" or the "dog."

Based on this theory, when the snake moves down exogenously, the CB could lower the peg so far that it would be on the lower edge and this would cause the snake to move back up (which is presumably the optimal policy) but I think Scott's view is that they just don't usually do this.

RPLong writes:

Mike Freimuth - thanks. That makes sense. I guess I just got lost in the metaphor. :)

Mike Freimuth writes:

No problem, that's pretty understandable it's quite the elaborate metaphor.

Scott Sumner writes:

RPLong, Mike provided the sort of answer I would have given. But the terms endogenous and exogenous are pretty vague, so I don't blame you for being confused.

lxdr1f7 writes:

"Student: So the central bank cannot control interest rates?"

It can control or very strongly influence the nominal short term interest rate on reserves but it only influences to varying degrees all other market rates. The CB has a very direct influence on rates on reserves as it directly participates in this market but mostly participates indirectly in other markets.

" the snake usually moves for reasons unrelated to shifts in the central bank's interest rate peg."

Considering the structure of the monetary system is vital to understanding the endogenous factors causing interest rates to move. If most broad money gets issued banks by during lending then how effective are the lending markets in allocating new money? If money isnt effectively entering circulation Debt to GDP levels will deteriorate resulting in stagnant credit markets while new lending leads to little real growth. This is how the ZLB is reached in our current economy.

On the other hand is the monetary system is one where money isnt created and debt and equally distributed to all debt to gdp levels will improve when money is expanded.

Emerich writes:

There's a problem with futures targeting: to attract any trading interest, and hence liquidity, there has to be volatility. To the extent NGDP targeting is successful, the contract would have no volatility, hence no trading interest, and the contract would die.

Edgar writes:


Trust but verify. Please give us references to research works that articulate your theory in models and present evidence in support of it.

Don Geddis writes:

Scott: love the imaginary dialogue. Very clever!

To make it a little less opaque, could you provide definitions or links to your different kinds of snakes? "Williamsonian", "Rovian", "Kimbellian". (I suspect many readers will not understand the references.)

@Edgar: maybe Sumner's Re-Targeting the Fed paper would help you?

John Becker writes:

That's a very good description of the issues that Milton Friedman and Eugene Fama brought up a long time ago about the Fed not truly being able to set interest rates. You've gone into a lot more detail about it than they did. I had a similar discussion with commenters on your blog about a week or two ago.

lxdr1f7 writes:

Alot of money is available to entities from trading NGDP futures so alot of resources will be devoted to anticipating ngdp. All those resources could be used elsewhere if the fed just automatically expanded money into citizens accounts when ngdp is too low.

Scott Sumner writes:

Lxdr, It's best to treat money and credit as completely separate issues.

Emerich, It's not a problem if no trading occurs, that means monetary policy is right on target.

Edgar, Mishkin's textbook on monetary economics has many of these ideas. The work of Milton Friedman is a also useful here.

Over at the money illusion I have a link to a short course on money which explains my views in more depth.

Don, You are right that the names are too opaque. I sometimes forget I'm not at The Money Illusion anymore. Miles Kimball recommended negative interest on base money. Steve Williamsson argued that raising interest rates can raise inflation. The Rovian snake referred to the blogger Nick Rowe.

Thanks John.

Comments for this entry have been closed
Return to top