Christina Romer defends NGDP targeting
I found this excellent Chistina Romer interview at Yahoo.com.
I’d like to briefly discuss the extent to which we should try to return to the previous trend line, and also the location of that trend line.
In the comment section you will occasionally see people criticize Romer for starting the trend line from 2007, when according to people of the Austrian persuasion (and many others as well) overheated economy was experiencing unsustainable levels of real estate speculation. In fact, by 2007 the real estate industry was already sharply depressed from 2005-06 levels, but let’s accept that critique for the sake of argument.
I decided to see what would happen if we pushed the trendline back to the 3th quarter of 2001, which was the trough of the previous recession. I don’t think anyone would claim the economy was overheated at that time. In the following table I list the current NGDP gap (2011:Q3) assuming 4 possibilities:
Starting date 5% trend 4.5% trend
2001:4 -10% -5%
2007:4 -12.4% -10.4%
Romer used a 4.5% trend and estimated a roughly 10% NGDP gap. She probably used calendar 2007 as a starting point, which may explain the slight discrepancy.
There are good arguments for both 4.5% and 5%. The long term trend for RGDP growth has been about 3%. So given the Fed’s 2% implicit inflation target, 5% seems about right. On the other hand many people (including me) think trend growth has now slowed to about 2.5%, or even a bit less. That makes 4.5% seem more appropriate.
[BTW, I don’t think this is the right way to think about the target, as I believe problems generally attributed to inflation are actually due to changes in either employment or NGDP. But I will play along since everyone else views low and stable inflation as one of the policy goals.]
I favor level targeting, which means returning to the previous trend line. But which trend line? In 2009 I favored returning to the trend line up through mid-2008. But so much time has now gone by that it’s no longer clear where the relevant trend line is. It would be different if the Fed had already announced a NGDP target, then we could draw a 5% trend line from that point forward. But they didn’t. If you don’t follow my argument, take this reductio ad absurdum example:
Assume that economic historians determined that the Roman Empire’s NGDP grew at a 5% annual rate during the reign of the “five good emperors.” They also showed that Italian NGDP is currently 23% above a trend line extrapolated from 180AD, the date Marcus Aurelius died. Should the Bank of Italy exit the euro and deflate Italy’s NGDP by 23%?
Each day that goes by makes the 1990-2007 trend slightly less relevant. Don’t get me wrong, I strongly support Christina’s Romer’s policy if the alternative is current policy. But at this late date it’s perhaps too much to expect the Fed to return all the way to the previous trend line. And I’m not sure it’s still optimal; many labor and debt contracts have been made post-2008, and much lower expected levels of future NGDP. So a good compromise might be to use the 2001:4 cyclical trough as the starting point, and 4.5% as the trend rate of NGDP growth. Those are both highly conservative assumptions, and yet allow us 5% worth of NGDP “catch-up,” before returning to the trend line. That’s 7% NGDP growth for 2 years, far higher than currently expected. If the Keynesians are right (and I think they are 80% right) the vast majority of that extra growth would be real. That would bring unemployment down quite rapidly, facilitating pro-supply side adjustments to the UI program. A virtuous circle would ensue.
PS. Don’t expect consistency in my blog posts; I’ll probably continue to throw out 5% numbers for simplicity. I also favor targeting per capita NGDP, not total NGDP, but usually don’t bother mentioning that fact. Let’s first get the Fed on board for the general concept, then we can work out the details.
PPS. Speaking of emperors, did Berlusconi remind you more of Caligula (orgies) or Nero (fiddling while Rome burned?)
Tags: Italy
14. November 2011 at 12:44
Berlusconi? Tiberius, sort of; not Caligula or Nero.
14. November 2011 at 12:55
I should add that my idea of these guys comes from Robert Graves, not primary sources.
14. November 2011 at 13:08
MMT shows that NGDP targeting has to be done with fiscal policy to be effective.
http://moslereconomics.com/2011/11/14/it-must-be-impossible-for-the-fed-to-create-inflation/
For all the QMs constantly confused what is fiscal and what is monetary: The distinction is very easy: fiscal policy creates/destroys state money, monetary policy determines the price of state money by conducting *swaps* of different forms of money (interest bearing and not) to set the rate.
Fiscal policy are the vertical operations, as simple as that:
http://bilbo.economicoutlook.net/blog/?p=332
14. November 2011 at 13:13
Nero – he fiddled and diddled plenty too.
But here is another contest:
Who was worse – Nero or Bush ?
14. November 2011 at 13:15
I mean – after all – Nero was at least amusing.
14. November 2011 at 13:18
NGDP targeting is really a back-door way of targeting wages. If the productivity trend has slowed, such that 2.5% rather than 3% potential RGDP growth is expected, the NGDP growth rate consistent with a given wage growth rate will not have changed. If we thought, before the slowdown, that 5% was a good NGDP growth rate, there is no reason to revise that downward to 4.5% (although you might argue for revising it downward a little due to slower population growth). The reduction in the target rate from 5% to 4.5% is a dangerous concession to those who believe in the I-word.
14. November 2011 at 13:36
Andy,
That’s an interesting viewpoint, one I hadn’t considered. Targeting NGDP essentially is targeting wages (well, income). But I’m not sure I understand why it is the case that “If the productivity trend has slowed, such that 2.5% rather than 3% potential RGDP growth is expected, the NGDP growth rate consistent with a given wage growth rate will not have changed.”
My guess is you mean that demand-side policy should not be responding to supply shocks (i.e. changes in RGDP growth). But I think Scott has been explicit before in that whatever trend NGDP follows, the monetary authority (if it exists) should stick to trend, no matter what RGDP does. Scott’s arguing now that the trend has changed, relative to what it was before 2007.
14. November 2011 at 13:37
I hope you read the comments section for that Yahoo article. People are a little skeptical of “the printing money”.
14. November 2011 at 13:41
Silas,
Is there a particular reason to lend those comments any more weight than the comments on a pro-free trade article complaining “they took our jerbs”?
14. November 2011 at 13:46
Re: Berlusconi
I have to go with Caligula, but Nero is disqualified. Per Wikipedia’s List of Common Misconceptions:
“In fact, upon hearing news of the fire, Nero rushed back to Rome to organize a relief effort, which he paid for from his own funds. He also opened his palaces to provide shelter for the homeless, and arranged for food supplies to be delivered in order to prevent starvation among the survivors.[3] Finally, he made a new urban development plan that attempted to make it more difficult for fires to spread.[4] However, it is true that he blamed the fire on Christians”
http://en.wikipedia.org/wiki/List_of_common_misconceptions
14. November 2011 at 13:52
@johnleemk: No, but in fairness, the same could be said for “experts” like Scott “just make people spend more and everything’ll be okay again!” Sumner.
Also, I think it’s a good reminder of how far advocacy of NGDP targeting has to go.
14. November 2011 at 13:52
How should Wikipedia know? That’s just some internet scribbler after all.
14. November 2011 at 13:59
Andy and John:
One the growth path is set, you stick to it. But why 5% to start? Sumner chose 5% because it sums the growth rate of potential output and the Fed’s inflation target of 2%. But that’s it.
Scott:
You are looking at this wrong. You mindset is too tied to growth rate targeting. It is like, you advocate targeting the growth rate, and then it is like, oh yeah, it is level targeting so we should “catch up” if growth is too low.
So, your notion that it is adding 5% to what it was in 2008 is wrong. Where you should be going 5% beyond isn’t where nominal GDP was at the beginning of 2008, (as would be correct with growth rate targeting.) It is 5% from what the growth path was in 2008.
Find the trend from the Great Moderation. If you want to pick some date and slow it, from its actual 5.4% value to 5% or 4.5% or 3%, then that is fine.
Those “Austrians,” who say we shouldn’t be increasing from 5% in 2008 because real estate prices are high are dead wrong. Look at the growth paths from the entire Great Moderation.
Anyway, using the bottom of the last recession. It is no more arbitrary than copying the growth rate from the Reagan/Volcker recovery from the trough of this recession.
But I think you need to stop thinking in terms of growth rate targeting. _A Growth Path_. It is a series of levels traced out into the future.
I guess this is 7%, 7%, 5% makes sense to you. At bottom, your intuition is 5% from where we are now. Growth rate targeting.
14. November 2011 at 14:09
Romer was great.
14. November 2011 at 14:12
i dunno about Berlusconi, but the FOMC reminds me of Nero. maybe its all the lead in the DC water system.
14. November 2011 at 16:09
We inch ever closer!
So we’re at 2001 at 4.5%. – you are pitching the Tea Party on 5% of catch-up growth / money printing.
Run the numbers again at 4%, and you got yourself a deal.
Oh what? At 4%, we’re on target?
So maybe you get make up of 5%, and 4% going forward, starting November 2012.
That my boy I think you can sell!
14. November 2011 at 16:43
I liked the 2.5 %, close to my estimate of 2.6% RGDP reasonable growth.
I have been looking at the extensive borrowing by the public, over the last two decades, that must have greatly contributed to RGDP, and cannot at present be sustained. This has led me to believe that it is too much to expect much more than that 2.5% figure for many years.
Hopefully, we can bring in high productivity for the future and achieve better numbers. But to catch up to that original trend line looks very difficult, and long term at best.
14. November 2011 at 17:08
Kevin, I actually read one of the primary sources (Suetonius?) but I can’t recall much of what I read. Didn’t Tiberius have that love pad on Capri?
OhMy, Central banks have been targeting inflation for decades with zero help from fiscal authorities. MMT is simply not relevant.
JimP, It would be fun to match each president with an emperor.
Andy, I completely agree, as I tried to indicated on my aside. I am willing to take any starting number they choose. But once set, the NGDP trajectory should never be changed—not in 2000 years. I was trying to consider the issue from the Fed’s perspective, not mine.
Johnleemk, See previous answer.
Silas, The comment sections of Yahoo articles? Yeah, I learn a lot from reading those.
Cthrom, I knew it was a myth, as all good stories are.
Bill, I am 100% for level targeting, but there is an issue of where to start, and what rate to impose.
Morgan, I’m not going that low.
RTB, In a sense it doesn’t matter, as we’ll never catch up to the RGDP trend. The key is NGDP.
14. November 2011 at 17:35
JimP:
“I mean – after all – Nero was at least amusing.”
And Dubya wasn’t? 🙂
14. November 2011 at 18:17
Scott, now that we know what kind of woman you are…
Now looky, the operable bit is it begins Nov 2012, you get a ton of traction, if you just wash your hands of Obama and put out a world of woo to the Tea Party.
14. November 2011 at 20:13
Scott – Wouldn’t it be better to measure NGDP trends from cycle peak to cycle peak when resources/capacity are closer to fully employed/engaged? If we do that we get just a touch over 5% p.a. from the cycle peak in 1990 to the cycle peak in 2007. If we lop off the entire 2002-07 expansion due to the housing bubble, the NGDP trend during 1990-2001 (cycle peak to cycle peak) is actually a bit faster, which makes Bill’s point about the Austrians being dead wrong about returning to a proper level of aggregate nominal spending/income.
14. November 2011 at 20:37
Scott,
Imagine an NGDP futures targeting regime where the Fed made the same arguments:
“Well, we could reduce NGDP back to trend levels after the overshoot, but by now people have made contracts with higher NGDP in mind, so we’ll abandon that old target for a higher one. Plus now trend growth is higher (we think).”
Windfall for someone long NGDP! And of course, speculators, realizing that the Fed will abandon their target after a miss, would challenge the Fed to do just that by bidding up the futures price, making the target even less achievable.
Of course, you could argue, “under my regime, the Fed will never miss.” Of course it will. As long as speculators think that the Fed will abandon its target, they will force it to abandon its target.
14. November 2011 at 22:39
I think where we start the trend line is pointless argument. Especially when we start looking back 10 years back. The fact is the economy cannot handle large nominal shocks.
I always like to try and think about things from a supply-side perspective even when it comes to monetary issues. Nominal shocks to the upside through irresponsibly expansionary monetary policy cause high inflation which makes it hard for businesses to set prices and write contacts. All this increases the cost of doing business and hence the economy suffers.
Nominal shocks to the downside are even more devastating. Reason being that wages are sticky and labor markets do not adjust. This causes unemployment to rise rapidly causing obvious problems to the economy.
So, Morgan’s argument that we are basically almost at trend if you look back 10 years ago is pointless. Imagine the Fed set a 5% NGDP growth rate starting in 2001. But instead NGDP continued to grow at a 7% rate. So now it’s 10 years later and your NGDP is 20% higher than you wanted it to be at this time. I am pretty sure that if the Fed suddenly did get serious about hitting its target and brought NGDP down 20% in one year, we would have a VERY BAD recession. Even if the market believed that all the Fed was doing was bringing NGDP back into line with its desired trend line that it had wanted all along. Wages and contracts have already been set at a much higher NGDP level. The shock would be too much to bear for the economy and it would go into dysfunction.
15. November 2011 at 01:28
“If the Keynesians are right (and I think they are 80% right) the vast majority of that extra growth would be real.”
“But once set, the NGDP trajectory should never be changed””not in 2000 years”
Suppose you’re wrong about the first part, and there’s very little real growth. Do you still refuse to change the trajectory?
15. November 2011 at 05:52
Some perspective on historical growth rates of RGDP, NGDP and per capita RGDP, NGDP:
http://www.measuringworth.com/growth/growth_resultf.php?US%5B%5D=DOLLAR&begin%5B%5D=1913&end%5B%5D=2010&begin%5B%5D=1929&end%5B%5D=2010&begin%5B%5D=1947&end%5B%5D=2010&begin%5B%5D=1790&end%5B%5D=2010&beginP%5B%5D=&endP%5B%5D=&US%5B%5D=NOMINALGDP&US%5B%5D=REALGDP&US%5B%5D=GDPDEFLATION&US%5B%5D=NOMGDPCP&US%5B%5D=GDPCP
Data Source: http://www.measuringworth.com/uscompare/sourcegdp.php
15. November 2011 at 05:55
Some perspective on historical growth rates of RGDP, NGDP and per capita RGDP, NGDP:
{Sorry: the link in my first attempt was bad}
http://www.measuringworth.com/growth/growth_resultf.php?US%5B%5D=DOLLAR&begin%5B%5D=1913&end%5B%5D=2010&begin%5B%5D=1929&end%5B%5D=2010&begin%5B%5D=1947&end%5B%5D=2010&begin%5B%5D=1790&end%5B%5D=2010&beginP%5B%5D=&endP%5B%5D=&US%5B%5D=NOMINALGDP&US%5B%5D=REALGDP&US%5B%5D=GDPDEFLATION&US%5B%5D=NOMGDPCP&US%5B%5D=GDPCP
Data Source: http://www.measuringworth.com/uscompare/sourcegdp.php
15. November 2011 at 06:25
Liberal Roman,
“Imagine the Fed set a 5% NGDP growth rate starting in 2001. But instead NGDP continued to grow at a 7% rate. So now it’s 10 years later and your NGDP is 20% higher than you wanted it to be at this time.”
You just described 9 years of failed policy of targeting the NGDP level. If 2001 = 100, and you were targeting 2002=105 and 2003=110.25, and actual NGDP 2002=107, then your 2003 target is still 110.25 (5% growth path) not 114.49 (7% growth path) nor is it 112.35 (5% growth from 107).
Obviously, you would not want to correct for 10 years of compounded failures in one year.
15. November 2011 at 09:03
NGDP per capita is a good start, but shouldn’t you really target NGDP per working age person? Look at this chart from the Economist comparing the US and Japan:
http://www.economist.com/blogs/freeexchange/2011/11/america-following-japan
…and I’ve been wondering, once that’s factored in, does it change any of the conclusions in any of your Japan posts?
15. November 2011 at 10:24
“Suppose you’re wrong about the first part, and there’s very little real growth. Do you still refuse to change the trajectory?”
See that’s the trade Salem, we do a bit of make-up, we lock down at 4%, and no matter what – unemployment is now from bad fiscal policy, time to end minimum wage, davis-bacon, etc.
15. November 2011 at 12:29
Tommy, I agree that however you cut it the Austrian argument is wrong (about the wrong trend being used.) I picked a conservative assumption to drive home that point.
David, You said;
“Well, we could reduce NGDP back to trend levels after the overshoot, but by now people have made contracts with higher NGDP in mind, so we’ll abandon that old target for a higher one. Plus now trend growth is higher (we think).”
You misinterpreted my point, I am completely against doing that sort of thing with a NGDP targeting regime.
Liberal Roman, Good point.
Salem, No, because the problems that are generally associated with inflation are actually associated with NGDP growth. So as long as NGDP is on target I don’t worry about inflation.
Eric, Thanks for that data.
jj, NGDP per working age person is probably better, I just refer to NGDP for simplicity. I’ve actually argued that nominal hourly wage rates are ideal.
15. November 2011 at 15:02
It seems as though commenters are dancing around the idea of a “natural” rate of RGDP growth by looking at historical trends. This is something I’ve been wondering about too, with relation to NGDP targeting. I don’t have an economics background, so please forgive me if these are stupid questions:
Suppose that the Fed sets a 5% NGDP growth target. Then suppose that RGDP growth over the next ten years averages 0%. What would this look like? Is this even possible?
Suppose that the Fed sets a 5% NGDP growth target. Then suppose that RGDP growth over the next ten years averages 5%. What would this look like? Is this even possible?
Suppose that the Fed, for no apparent reason, sets a 15% NGDP growth target. Would this target be as easy for the Fed to defend as a 5% target?
15. November 2011 at 18:13
Andrew, I use the same logic to reach my thinking.
Basically, everybody is looking back at 3% RGDP and assuming it happens going forward… or at least 2.5%
They are confident enough that they want to slap another 2% of inflation on there.
I think we should use past RDGP to define NGDP.
So see set course for 3%, and now hating inflation is baked into our very being.
ANY inflation except when we grow below historical levels is DESPISED because causes rates to go up.
So our government is suddenly forces to find every single policy possible that makes prices sticky and END it with extreme prejudice.
And yes, under those circumstances I think it is wroth betting on whether we get 3% continued as pure RGDP or not.
15. November 2011 at 18:26
Scott,
Sorry, I believed you just made that same argument: abandon the original NGDP level target because contracts are being written with a lower growth rate in mind, and because trend growth is now slower. I must have misinterpreted it.
16. November 2011 at 16:35
Andrew, Yes, all of those are possible.
David, No, I probably said don’t try to regain the 1990-2007 trend, as contracts have been re-written at lower NGDP levels. But if we had actually been targeting NGDP during 1990-2007, I certainly would not have made that suggestion.
17. November 2011 at 12:35
Andrew,
Assuming that by “average growth of x%” you mean “at the end of the time period in question, the metric has grown at a x% annualized rate” and NGDP targets are met:
“Suppose that the Fed sets a 5% NGDP growth target. Then suppose that RGDP growth over the next ten years averages 0%. What would this look like? Is this even possible?”
Here, the price level would have risen from 100 to ~162.89. (things that used to cost $100 will cost $162.89). On average, people are no better economically after 10 years.
“Suppose that the Fed sets a 5% NGDP growth target. Then suppose that RGDP growth over the next ten years averages 5%. What would this look like? Is this even possible?”
Here, the price level would have stayed flat from 100 to 100. However, people on average are ~63% better economically (there stuff is better and/or they have more of what they want).
“Suppose that the Fed, for no apparent reason, sets a 15% NGDP growth target. Would this target be as easy for the Fed to defend as a 5% target?”
I would expect that defending policy to sustain 15% NGDP annualized growth over ten years would be difficult; using empirical studies of historical data would be impossible.
5. January 2016 at 22:45
[…] Even Tyler Cowen, a big detractor of Keynesian ideas, has cautiously endorsed the idea. (Update: Christina Romer and a number of Fed officials have also come out in favor of more easing.) The forces arrayed […]