Kling on NGDP targeting
Here’s Arnold Kling’s response to my recent advocacy of NGDP targeting:
Because of a cold, I will be missing an invitation-only event featuring Scott Sumner. In the spirit of sharing my ideas without my germs, let me offer some thoughts on nominal GDP targeting.
That’s no excuse, I also had a cold. 🙂
Kling continues:
1. For the Fed, a target represents a justification for taking action. Some people may be upset with what your action does to the exchange rate or the interest rate. Having a target allows you to justify your action.*
2. I would like to see the Fed use a target to justify its actions.
3. If the Fed were to use a target, then future nominal GDP would be an excellent choice.
4. In the current environment, a target for future nominal GDP could be used to justify expansionary actions.
5. There are plenty of expansionary actions available. The Fed could charge a penalty for holding excess reserves. The Fed could be buy foreign bonds (this might require a change in current law). There still are plenty of long-term Treasuries out there for the Fed to buy.
6. In the best case, the Fed would hit a target for future nominal GDP, unemployment would fall fairly quickly, and we would live happily ever after.
7. In the worst case, we would begin to shift to a regime of high and variable inflation. The Fed would have to undertake strong contractionary measures in order to keep nominal GDP on target, while unemployment remains high.
I believe we would all live happily ever after, and that inflation would not become high and volatile. In order for inflation to be high (on average), RGDP growth would have to average much less than 3%. Let me repeat; much less, not slightly less. A trend rate of 1% or 2% RGDP growth will not get you high inflation on average, unless you think inflation was still high after Volcker brought it down to low levels.
Much more importantly, I don’t think high or variable inflation is a problem as long as NGDP growth is on target. All of the problems that are widely believed by economists to flow from high and variable inflation; actually result from high and variable NGDP growth:
1. Excessive taxation of capital in a non-indexed tax system results from high nominal rates of return, associated with high NGDP growth.
2. Unfair borrower/lender redistributions actually result from volatile NGDP, not volatile inflation.
3. Distortions to the labor market (when nominal wages are sticky) are caused by NGDP shocks.
4. Even the “shoe leather” cost of inflation may be better described by NGDP growth, assuming real interest rates and real GDP growth rates are strongly correlated.
George Selgin has much more to say about the advantages of using NGDP.
In the end Kling argues the NGDP targeting is worth a shot, and he also has some interesting things to say about the possible reasons why the Fed hasn’t yet taken this step:
(a) They do not want to be embarrassed if they are unable to hit a target
(b) This is what Tyler Cowen would call a Straussian situation, in which the insiders must never reveal their true agenda, or horrible demons will be let loose, leading to social breakdown and bloodshed.
(c) They fear that announcing a target would create “lock-in” and cost flexibility.
(d) A target would make many of the departmental functions and rituals (such as FOMC meetings) long cherished at the Fed seem pointless.
(e) The Fed is institutionally more concerned with the stability and profitability of the banking system than with macroeconomic variables.
I would add that most private macroeconomists also prefer inflation targeting to NGDP targeting. Or they favor flexible inflation targets, but don’t plump for NGDP targeting because it seems too crude. I believe that is because inflation and RGDP play an important role in their macro models. Unfortunately, the ‘inflation’ in their models bears little resemblance to real-world inflation indices.
Tags:
18. December 2010 at 09:51
Scott, Could you expand on that last point? How is ‘inflation’ in their models different from real-world inflation indices?
18. December 2010 at 10:50
Alexander, Real world indices use sticker prices or list prices, which are very different from transactions prices (the theoretically relevant concept.) Thus the rent someone pays on an apartment is not a price at all, but more like an installment payment on a loan.
Government data showed housing prices rising (even relative to other goods!) between 2008 and 2009, during the biggest housing crash in US history. Those prices made up 39% of the core CPI. The core CPI is nearly worthless.
18. December 2010 at 11:22
Scott,
I’m sold on the advantages of NGDP targeting but as an intermediate step price level targeting (PT) might be more achievable. It’s been done before (Sweden in the 1930s) and it might be on the verge of being done again (Canada). After all, Rome wasn’t built in a day.
I disagree strongly with Kling’s seventh point but on the other hand he totally hits the nail on the head as to why the Fed has not pursued NGDP targeting thus far, and those reasons apply equally well to PT.
The FOMC is a deeply insecure organization. Full transparency coupled with a pure rules over discretion policy would either reveal monetary policy’s impotence in the face of a “liquidity trap” (I don’t truly believe that) or render the FOMC as obsolete as phonograph cylinders at the iTunes store.
18. December 2010 at 11:34
Its been a pleasure watching you and Kling wrestle with the manifestations of money as information, e.g. “recalculation” vs. “NGDP targeting”.
My intuition as a noneconomist, informed in part by the errors of an insular global warming modeling community (I have no position pro or con. regarding the ultimate scientific facts), is that Kling’s concern about time nonlinearities in measurement and response leading to “high and variable” inflation requires more than “I don’t think” or “I believe” in terms of persuading your audience. The time nonlinearity in the business cycle may be like high voltage current. Handle with extreme caution. It seems to keep shocking the pros despite all of their specialized knowledge.
As anyone applied the physics spin glass phase transition models to the macroeconomy? Amit in the 80s did a lovely job extending these models to neuronal processing. Topics like symmetic vs. asymmetic resonance, local “nearest neighbor” v. global interactions, and temporal lags in response can be addressed in slightly more concrete terms. “Recalculation” directly maps onto the “synaptic weights” between individual agents. Would be curious to have someone discuss the role of money supply as it relates to the gain function/randomness driving phase transitions between local and global minimums.
18. December 2010 at 11:57
Scott:
“In order for inflation to be high (on average), RGDP growth would have to average much less than 3%”
That assumes that the Fed succeeds in hitting the NGDP target. I think Kling’s concern is that the Fed will use the target to justify actions that will ultimately result in overshooting the target. All that’s necessary is that the Fed badly misestimate the AS curve; and Kling’s whole “recalculation” idea seems to be arguing that we are generally misestimating the AS curve.
The main issue here, I think, is instrument instability, but I believe it can be solved by having a sufficiently long lag on the target, so that the Fed has time to correct its course smoothly when it makes a mistake. There’s room for a lot of disagreement about how long is sufficiently long. I’m guessing you and Arnold Kling would be on opposite extremes of that disagreement, and I would be somewhere in the middle.
The other issue is that there may be no RE equilibrium that corresponds to the Fed’s target path for NGDP. (That, I think, is what Krugman and a lot of New Keynesians would be concerned about. I’m not sure how it fits into Kling’s thinking.) In that case, you could get lot of volatility, and the Fed will seldom come close to its target, even if it has time to adjust. Personally, I’m not terribly concerned about this issue because I think it’s only really a problem when market participants have homogeneous opinions. In practice, even if the target path is not an equilibrium for the average market participant, there will be enough diversity out there to get a liquid market on the target path.
18. December 2010 at 12:37
http://themoneydemand.blogspot.com/2010/12/monetary-policy-and-minsky-cycles.html
18. December 2010 at 13:45
If Kling is completely right about recalculation, then inflation would have been substantially higher over the last few years, assuming NGDP remained on target. There would have been a few quarters with inflation as high as 7 percent (-2 percent growth in real GDP because we don’t want houses but have no idea what else to produce) which at 5 percent money expenditure growth generates 7 percent inflation. More quarters would have had 3 percent or even 4 percent inflation. But mostly, we would be running at about 2 percent inflation.
AS really has nothing to do with hitting the NGDP target (at least, I don’t think it does.) AS has to do with how it splits between inflation and output.
18. December 2010 at 13:50
Correction to my last comment: What’s necessary is not that the Fed badly misestimate the AS curve but that it badly misestimate the impact of policy on NGDP. In traditional monetarist terms, one would say they miestimate the velocity of money. “Velocity” is presumably conditional on the Fed’s policy, and there may be some point in the policy space at which velocity suddenly increases unexpectedly.
18. December 2010 at 13:57
if you use futures markets to estimate NGDP, there should be no foreseeable high inflation except in periods of low RGDP growth. So of course you should use future markets to estimate NGDP.
What interests me is that it seems like you could build a policy rule into futures markets. As far as I can tell, Scott’s proposal is not a proposal for NGDP targeting, it’s proposal for an [expected NGDP growth]-[interest rate] rule. Which seems like a good rule to have! For example, it leads to the profitability of an investment depending only on its utility returns (assuming utility is elasticity-1 in income.)
The way you build a policy rule into futures markets is by creating instruments like TIPS that have a savings vehicle aspect and a futures market aspect. Then the Fed agrees to buy and sell them at a certain price.
18. December 2010 at 14:00
Oh, and also, all the reasons you give apply to NGDP per capita, not NGDP.
18. December 2010 at 14:51
I wonder what Bernanke thinks about NGDP targeting. He seems well-aware of the Japan problem, and he is no Nipponista (speaking of words, this is my new description of the anti-QE crowd).
It is too bad we have such a large country and so many livelihoods at stake–if would be nice to experiment with NGDP targeting for 10 years, and then examine the results. However, casually conducting an economic experiment on 300 million people may strike some as cavalier.
Still, I think NGDP is an idea whose time has come. The only problem I see is that if people knew there would not be recessions, asset prices might really get bid up, and then crash anyway.
18. December 2010 at 16:44
I’m finding the time these days to find inner peace. The Rosary does wonders to steady the mind. Thank God for winter break.
18. December 2010 at 21:57
“In order for inflation to be high (on average), RGDP growth would have to average much less than 3%. Let me repeat; much less, not slightly less.”
This is an excellent conversation to have with those referred to as “conservatives” here. It’s likely a misnomer, however, as much of the rumblings about QE appear to me to serve as ‘can the Fed’ propaganda (which has been rather effective) by those who are on the fringes of anarchy or would have something to gain rather than from the conservative mainstream. Those disseminating apocalyptic visions of doom due to QE, to a deeply unsatisfied and unwitting audience who can see the grass being greener on the other side of the fence, in order to get what they want do not mention a replacement strategy or only vaguely refer to ‘sound money’, whatever that means .The reality is that there is far more politics involved in what they dub rather ambiguously as ‘sound money’ as there has been involved at the Fed. The debates over silver vs. gold and the fixed price raged for decades in the late 19th and into the 20th century. They were driven by mining lobbies and speculators and government seemed to have rather manic/depressive behavior toward monetary units as a result, not to mention the Jay Gould gold shenanigans with the help of some in the Treasury Dept. Their idea of ‘sound money’ that they shill as money without politics or corruption involved in some way or another and includes the impossibility or extreme difficulty in robbing average Joe over night is nothing short of disingenuous or an ideological fantasy at best. And I haven’t ventured into the likely economic realities of them getting what they want, or how some of their supporting arguments are statist in nature and incoherent with entirety of conservative thought.
I thank you for your effort to persuade the conservative audience, but if you want to get their attention without being figuratively tarred and feathered in front of them by short-sighted anarcho-propagandists who have already sold their ideas, practically lock stock and barrel, their idiocy must be exposed and shut down. There are some diplomatic tacks that can be taken to accomplish it which would probably work better than cozy sugar coating on the truth. There is no need to contrast with liberals and Keynesians. Most already object to those ideas. They need to know why keeping the Fed or having something akin to it is important, and that your idea sets it on a market driven, forward-looking policy path rather than a finger in the air, discretionary debacle path, and what it means to their wallets. It is that discretionary part of the Fed that is at the center of the controversy, and at the epicenter of our current mess. That should be the focus while hitting multiple targets, the ‘can the Fed’ crowd and the gold bugs that have already nearly derailed the possibility of a rational and practical solution to the problem.
19. December 2010 at 09:09
Scott,
RGDP of 2% (not “significantly below 3%”) for the next five years would likely result in a 8-10% unemployment rate throughout the period.
Assume in “year 3” that RGDP is 2%, TIPS 5-yr inflation expectations reach 5%, and unemployment is above 8%. Also assume this implied NGDP growth rate would result in an overshoot of an NGDP expectations level target. What would the consequence be for Fed credibility and for inflation expectations if the Fed did not tighten in response? What would the consequence be for unemployment if it did?
If we have slow growth and high inflation, Fed policy will be asymmetric.
19. December 2010 at 11:25
Mark, Yes, in future decades we’ll look back on this Fed as “so 20th century.”
JonB, You said;
“My intuition as a noneconomist, informed in part by the errors of an insular global warming modeling community (I have no position pro or con. regarding the ultimate scientific facts), is that Kling’s concern about time nonlinearities in measurement and response leading to “high and variable” inflation requires more than “I don’t think” or “I believe” in terms of persuading your audience.”
Um, right after I said “I don’t think” I listed 4 reasons for not thinking so. Unfortunately I’m not knowledgeable enough about science to handle the rest of your question.
Andy, Two responses:
1. That’s why I favor targeting the forecast, I don’t trust the Fed.
2. Under level target the market helps the Fed, and gross errors are far less likely. Level targeting helps stabilize V.
The Fed MUST DO SOMETHING, the only question is what. Doesn’t it make more sense to aim for the target, than to intentionally try to miss the target? (Which is what they are doing now.)
You said;
“The other issue is that there may be no RE equilibrium that corresponds to the Fed’s target path for NGDP.”
I don’t agree, for all sorts of reasons. I’d add that if this were really a serious problem under fiat money regimes, it would be obvious by now.
And finally, this isn’t really an argument against NGDP targeting, it’s an argument against fiat money in general.
JimP, Thanks, That post is by frequent commenter 123, and is an excellent post.
Bill, I agree.
Will, You are partly right, it is an expected NGDP rule. But not NGDP – i, interest rates are not involved.
I agree about NGDP per capita
Benjamin, I don’t recall Bernanke ever mentioning NGDP targeting, which leads me to believe he is opposed.
Bonnie. Those are good observations.
David, You said:
“RGDP of 2% (not “significantly below 3%”) for the next five years would likely result in a 8-10% unemployment rate throughout the period.”
First of all this would be extremely unlikely under 5% NGDP growth. But let’s say you are right. In that case Keyneisan economics would basically go down the toilet, it would be massively repudiated. Krugman would be the laughingstock of the entire profession. (I would also be to a lesser extent.)
So the right would have free rein over monetary policy, and would certainly be able to tighten as required to keep NGDP growting at 5%.
You said;
“Assume in “year 3″³ that RGDP is 2%, TIPS 5-yr inflation expectations reach 5%, and unemployment is above 8%. Also assume this implied NGDP growth rate would result in an overshoot of an NGDP expectations level target. What would the consequence be for Fed credibility and for inflation expectations if the Fed did not tighten in response? What would the consequence be for unemployment if it did?”
I don’t think you’ve thought through the implications of your hypothetical. If that happened the economic profession would 100% buy the structuralist theories of what is wrong. The other side would be totally discredited. In that case no one would be asking the Fed to bring unemployment down. Instead they’d be clamoring for job retraining, lower minimum wages, less UI, etc. I think you are imposing a very different hypothetical on the current intellectual climate. It won’t work.
19. December 2010 at 11:51
Scott,
I don’t think you have to be an extreme structuralist to buy into the scenario. First, RGDP of 2% may produce only about 200k jobs/mo. If the increase in the labor force is 150k (taking into account the return of discouraged workers), then 2% growth would result over 8% unemployment in two years’ time.
Is your estimate of job growth with 2% RGDP much higher? It seems so from your comments.
Also, “the right” may want hard money policy even under 8% unemployment, but they certainly won’t want the Fed to tighten a tank the stock market!
19. December 2010 at 11:52
Sorry — that should have read “to tighten and tank the stock market”.
19. December 2010 at 14:43
Scott,
You have always given reasons for the statements prefaced by “I think” or “I believe.” My problem as an outsider to the economic vocation is that the explanatory jargon is very dense and appears be inconsistently used by the different theoretical schools. I don’t doubt your sincerity or knowledge but I don’ t intuitively grasp your answer to the unstable temporal feedback problem. Is Harless discussing this too?
My point about mathematical models is that they require dimensional rather than categorical thoughts, i.e. words in a blog. Their explanatory power is provided by pictures and/or mental visualization.
You are very frustrated by verbal gymnastics of all these theoretical schools and political groups which stand in the way of what to you is an obvious solution. Perhaps the answer isn’t a better sound bite, e.g. “growth in income” rather than inflation. Perhaps the answer is the cartoon.
Is there a group of anti-Asperger spectrum voters who marry young, work in nonverbal professions, and instinctively distrust macroeconomic/political solutions which appear to vest more centralized power in the hands of their Asperger counterparts? I’m not sure that I agree with Palin’s editorial in the WSJ for sound money but to dismiss her and her followers as stupid or uneducated is a grievous error. Would this group do better with charts and graphs than AS or AD?
I wish I was a better mathematician…Sigh.
20. December 2010 at 18:28
David, You said;
“Is your estimate of job growth with 2% RGDP much higher? It seems so from your comments.”
No, you misunderstood me. I can imagine 2% RGDP growth and slow job growth. What I find hard to imagine is 2% RGDP growth and 3% inflation, at a time of 10% unemployment. It’s not impossible. But I doubt it, and so far Krugman and I have been right about low inflation. Those numbers would make Krugman look foolish, but I don’t expect that to happen.
You said;
“Also, “the right” may want hard money policy even under 8% unemployment, but they certainly won’t want the Fed to tighten a tank the stock market!”
They do right now, even though it would tank the stock market, and they did in the 1930s, even though it did tank the stock market.
Jon, I don’t think I’ve ever called Palin stupid, nor do I believe she is stupid. I don’t know math like you do, but don’t consider myself stupid, and she doesn’t know money like I do, but that doesn’t make her stupid.
I don’t want to vest more power in the Fed. I want to take all discretion away from the Fed, and let the market set the money supply and interest rates. That’s probably less centralized than Palin’s preferred policy.
Nor am I bitter that people don’t listen to me. Indeed I’m flabbergasted that they do. I never in my life expected to be the subject of posts like this one at The Economist:
http://www.economist.com/blogs/freeexchange/2010/12/monetary_policy_2