Scott Sumner  

Beckworth interviews Summers

Private Foundations and their ... Does Prosperity Make Us Utilit...

David Beckworth's recent interview of Larry Summers was a treat for two reasons; there was lots of thought-provoking discussion, and I found a written transcript of the interview. Here's an excerpt (discussing secular stagnation):

You have a demassification of the economy. Think about Amazon rather than malls. Think about the fact that an office building for lawyers now require 600 square feet of space per lawyer where it used to require 1200 square feet of space per lawyer because they no longer need filing cabinets or paralegals to deal with the content of those filing cabinets because of the cloud. And think about the fact that the country's leading technology companies and most valuable companies in the world, Google and Apple, have as their major business problem a surfeit of cash flow and how to manage that cash flow. All of these things taken together suggests ample reason for believing that real interest rates would have trended downwards, and that's in fact what we have seen. And while many at the Fed were very quick to attribute low interest rates to so-called head winds, I thought by 2013 that the head winds theory was implausible and by 2017 that the head winds theory was ludicrous, that it was hard to see what head wind there was by 2017 that one wouldn't expect to be semi-permanent.

And in fact, if you fit a trend from the late '90s through the period before the crisis, it more or less tracks the current level of real interest rates. So I think we're living in a world, David, where the neutral real interest rate is close to zero, where that low neutral real interest rate means a couple things. It means that we're likely to live in a more levered, more bubble-prone economy than we have historically. Some ways the way to understand where the puzzle before the crisis was, here it was, we had the mother of all housing bubbles, we had vast erosion of credit standards and all of that was only sufficient to produce adequate growth, not overheating. And in the same way, we're gonna have to keep interest rates low enough that we're gonna promote high leverage, problematic credit in order to generate growth. And when, and if, and it will happen sometime, growth slows or collapses, historically the Fed has cut rates by 400 basis points in recessions and there isn't gonna be that kind of room next time around. So that's the kind of secular stagnation theory as I see it.

I think he's right about the "demassification" of the economy (is that a word?), but I have a slightly different view of bubbles. Summers is probably right that low rates are the new normal, but this should lead to higher than normal asset prices, according to most standard metrics. I've argued that in the 21st century there will be many more examples of phenomenon that people call "bubbles", but that bubbles don't actually exist.

Interestingly, later in the interview Summers expresses views that are closer to my own:

I think the Fed in general has an urge to normalization, but I think it's better for the policies to follow from the arguments, rather than for the arguments to follow from the desired policies. And I don't see the financial stability rationale either. I'm not sure that markets are extraordinarily overvalued. If I believed with confidence that markets were a bubble, I'm not sure that would be a reason to tighten policy. It might be a reason to ease policy. 'Cause when the bubble bursts, we're gonna have a real problem. And so, I might as well get some stimulus into... Behind the economy. So, I don't see the case for tightening. Things could happen in the data in the next several months that change my mind, but on the current evidence, it seems to me that you have to work a bit too hard to manufacture a case for tightening.
The Fed tightened policy in 1928-29 to address a perceived stock market bubble, and it led to exactly the sort of policy mistake that Summers worries about here.

Summers says some positive things about NGDP targeting, but at times he seems to lose the thread of the argument:

Second, I think that if you're serious about a symmetric target, you have to be prepared to exceed the target sometimes. If now, after nine years of recovery, with years more of recovery forecast, and with an unemployment rate below 4.5%, if that's not the moment to exceed the 2% target, I don't know when that moment would ever come.
Actually, the time to exceed the target is during recessions. The time to fall short is during booms.

And this:

I think all those things make an attractive case for nominal GDP. On the other hand, it's the sum of a good and a bad, and that seems a slightly awkward thing to target, and so I'm not absolutely certain that I would prefer it to a price level path target of some kind.
Maybe I'm missing something here, but I'd say the mixture of good and bad is precisely what makes NGDP targeting desirable. If you targeted something good you'd want it to be as high as possible and if targeting something bad you'd want it to be as low as possible. The fact that NGDP contains both "good" and "bad" components helps to explain why we don't want the NGDP growth rate to be either too high or too low.

Screen Shot 2017-09-18 at 11.46.23 PM.png

Comments and Sharing

COMMENTS (12 to date)
Garrett writes:

A lot of the arguments I hear that the stock market is overvalued focus on the Shiller PE Ratio, which takes current prices and divides by the average of inflation-adjusted earnings over the last 10 years. This ratio is at around 30 right now, which is around where it was prior to the Great Depression, but not quite where it was at the peak of the "tech bubble."

The thing is, the ratio still includes 2008 and 2009 (see this chart). If we have a couple years of flat earnings, these years will roll off, the denominator with increase significantly, and the ratio will fall. Any amount of earnings growth will accelerate the fall.

A related argument I hear is that profit margins are elevated (see page 3), and therefore once wage growth picks up they'll erode. My issue with that is the upward trend in margins and downward trend in wage growth are decades old, so there needs to be a very good reason to predict a reversal of either.

Of course, if there's a recession then stocks will fall and they would have been overvalued now. I'm sure lots of people will pat themselves on the back if that happens.

Michael Byrnes writes:

The other riff on that "good and bad" comment is that inflation itself can be either good or bad depending on the context, but a strict price level target would not discriminate.

bill writes:

Beckworth is a great interviewer and this is already on my list. I hope he asked Summers something like, "so if inflation is a "bad", isn't it better that it be higher when RGDP growth is low and vice versa? Isn't counter-cyclical inflation better than pro-cyclical inflation?"

Scott Sumner writes:

Everyone, All very good comments---I agree.

Mark Bahner writes:
I think he's right about the "demassification" of the economy (is that a word?),

It may be a word, but it's not the proper word. ;-) The proper word is "dematerialization":

From Wikipedia (a great example of dematerialization!)

Jesse Ausubel and others have been writing about it for ~30 years:


Here's a good summary article, with lots of examples:

Ron Bailey on dematerilization

Kevin Erdmann writes:

The comments on interest rates and leverage seem like an especially bad example of reasoning from a price change. His stated reason for having low interest rates is that firms are flush with cash. He begins by explicitly giving a supply cause for low rates, yet in the next paragraph, he says that low rates will lead to higher demand for debt. Google and Apple won't have demand for debt.

Maybe the leverage he is talking about is outside of the corporate sector? Maybe housing? In 1986, median Price/Income for homes in LA was 3.9x. In Dallas it was 3.3x. By 2005, in LA it was 11.3x and in Dallas it had declined to 2.8x. The rise in leverage in the household sector was not significantly due to interest rates.

Of course, the solution to that problem kind of solves some of the problem Summers is talking about. A few trillion dollars worth of residential investment in cities where the value of the investment is worth 3 times the cost of building would do wonders for growth and for those low interest rates.

Scott Sumner writes:

Thanks Mark.

Kevin, Yes, the lack of investment opportunities is partly caused by restrictive housing policies.

Rajat writes:

We've discussed this topic before, many times. I hate to say this, Scott, but looks like Summers is another proponent of the 'misery index' rationale for the dual mandate. Like Ryan Avent for example: here.

[html fixed--Econlib Ed.]

Thaomas writes:

I do not understand the supposed link between "demassification" (Amazon?) and a decline in long term real interest rates.

More to the point, what is the policy implication? Monetary policy is not supposed to have an interest rate goal, rather a real growth ("full employment") and a stable prices objective. [Neither is "financial stability" an objective although in a particular context, restoring financial stability may be necessary to restore growth and prevent deflation. Whatever interest rate results (or may be necessary when ST interest rates are being used as instruments to influence growth and prices) ought not be a fundamental issue.

Mark Bahner writes:
I do not understand the supposed link between "demassification" (Amazon?) and a decline in long term real interest rates.

Yes, I'd also like Scott to explain this. (I have my own theory, but it's probably better to get the information from someone who knows what he's writing about. :-))

Mark Bahner writes:


I do not understand the supposed link between "demassification" (Amazon?) and a decline in long term real interest rates.

No word yet from Scott on this, so I'll take a stab.

First, a little background. My only knowledge of economics comes from a few college courses taken back when malaise was king. But I am on record as betting that a Nobel laureate in economics will be more than a factor of 100 too low in his prediction that per capita gross world product (GWP) will be “only” $90,000 (year 2000 dollars) in the year 2100:

Long Bet prediction #194

My reasoning was this: Right now we have about 7.5 billion people, with an average per-capita income of $10,000, for a GWP of about $75 trillion. Now suppose computers/robots add the equivalent of 1000 times the current population of humans? (That could be done, because computers/robots don’t need houses or food beyond electricity.) If the GWP per computer/robot-capita is the same as the GWP per human-capita, the GWP (combined computer/robots plus humans) would go up by a factor of 1000, to $75 quadrillion, or $10,000,000 per human-capita.

BUT…will that be what actually happens? For instance, the median per-capita human income in the U.S. is about $30,000 per year. Why would anyone pay a human that kind of money if a computer/robot could do the same amount of work for

THAT’S where I think the link between “demassification” (“dematerialization”) /Amazon and long term interest rates come in. For example, as I’ve written previously, on-line shopping with delivery by autonomous vehicles is going to obliterate brick and mortar retail. That will save everyone tons of money--and greatly dematerialize the economy--but it will show up as deflation that we’re not good at measuring. The deflation will be there, but we won’t measure it accurately. Therefore, interest rates will stay low...unless the government prints massive amounts of money to keep the not-properly-measured deflation from happening.

That’s my theory, anyway.


Mark Bahner writes:


A second a massive unmeasured (or at least not completely measured) deflationary period, the real rate of interest would actually be huge, even at zero percent.

(And here I'd been thinking that my mortgage rate was a good deal. :-))


P.S. Mostly kidding about the mortgage. I don't think we're in a massively deflationary period now, but in 1-3 decades, as computers/robots become ascendant, I think we could get into a massively deflationary situation...unless the U.S. government starts creating a whole lot of money.

Comments for this entry have been closed
Return to top