No Great Technological Stagnation

By Artir

EDIT: I DO NOT SAY THAT ROBERT GORDON IS WRONG. Please read this post for reasons why you can have growth in individual technologies and productivity slowdown in general.

Some people many economists say we are living through a Great Stagnation. The term, was made famous by Tyler Cowen’s book of the same name and the latest iteration is, of course, Robert Gordon’s The Rise and Fall of American Growth.

But they usually look at economic factors like Total Factor Productivity (TFP) or GDP per hour worked. By these measures, we are living through a stagnant period. Some expected that the stagnation would go away once we properly accounted for the Internet and similar hard to account for innovations, but apparently this was done, and the picture doesn’t change much.

In this post, I take an engineering perspective and look directly at technology itself. For some reason, no one has done what I will do in this post. Surely productivity is important, but since technology is supposed to be a substantial component of TFP, someone should have looked into precisely that. Shame upon the World’s Blogosphere!

My point here is that, by the measures we have, there is no TECHNOLOGICAL stagnation (THIS IS COMPATIBLE WITH A SLOWDOWN IN PRODUCTIVITY GROWTH). I leave it to someone else to solve the puzzle of why TFP growth is low while technological growth is constant.

Let’s begin with some charts. Here (Nagy et al. 2011),  you have long run trends for several Information Technologies and different curve fittings for them. In the paper they explain which one is the best one. Hint: not the exponential!

Captura de pantalla de 2016-04-25 21-34-20.png

However, if you happen to be into pulverized coal-fired utility boilers, I also have a paper for you, Yeh & Rubin (2007)

Captura de pantalla de 2016-04-25 21-42-47.pngAnd what about energy technology more broadly? Yes, there are also papers for that. Koh & Magee (2008).

Captura de pantalla de 2016-04-25 21-45-09

Captura de pantalla de 2016-04-25 21-45-18.png

Captura de pantalla de 2016-04-25 21-45-35.png

Captura de pantalla de 2016-04-25 21-46-35.png

Captura de pantalla de 2016-04-25 21-47-23.png

Captura de pantalla de 2016-04-25 21-47-41.png

Some of these trends are plateauing. But there is no slowdown from the 70s.

And they even go on to say that

The consistency of the rates of progress in the different functional categories over time might make us hesitate in accepting theories that hypothesize clear breaks in technological eras associated with different kinds of operands. The continuous nature of the progress rate improvements over many decades in both energy and information technology appears inconsistent with the idea of distinct revolutionary periods for either or both types of technologies.

I also have plots for energy efficiency from Ayres et al. (2005)

Captura de pantalla de 2016-04-25 21-57-34.png

High temp heat are electric furnaces, low temp heat is heating, electronics is radio, TV and information processing (transmitting waves wastes energy). Average efficiency stays like that because the mix of appliances has changed, not because technological efficiency hasn’t, says Ayres.

Next, Moore’s Law, just because if you write about technological progress without mentioning Moore’s Law, Ray Kurzweil might get angry or something.  From Magee (2012)

Captura de pantalla de 2016-04-25 22-05-03.png

Speed for some transports, from Chang & Baek (2010)

Captura de pantalla de 2016-04-25 22-08-35

There seems to be a great stagnation in military aircraft speed(!). But not in trains or automobiles.

Now, some from Smil (2005)

Captura de pantalla de 2016-04-25 22-17-17.png

Okay, there is some thermal efficiency stagnation, but it may have recovered after the 90s. Data from TEPCO, later shows increasing improvements. It could be interesting to study just this technology to see what is really going on.

Captura de pantalla de 2016-04-25 22-15-15.png

Captura de pantalla de 2016-04-25 22-16-10.png

From Smil (2010) now, and dedicated to Anton Howes, steam engines!:

Captura de pantalla de 2016-04-25 22-22-10.png

Captura de pantalla de 2016-04-25 22-23-36

Captura de pantalla de 2016-04-25 22-24-03

Captura de pantalla de 2016-04-25 22-24-40.png

Solar panel efficiency, from NREL’s famous chart



Thermal efficiency for TEPCO power stations

Captura de pantalla de 2016-04-25 23-02-28

Aircraft fuel efficiency from here and here (The units in the first chart seem to be gallons per seat-hour, not gallons per seat-mile)


Fusion energy (a figure of merit)


And automobile fuel efficiency and emissions from here

Captura de pantalla de 2016-04-25 23-05-32.png

To finish, some plots from Lienhard (2006) and Lienhard (1985). Not all of them are post 1970, but they are interesting nonetheless

Captura de pantalla de 2016-04-25 22-31-44.png

Captura de pantalla de 2016-04-25 22-33-59.png

Captura de pantalla de 2016-04-25 22-34-18

Captura de pantalla de 2016-04-25 22-34-54.png

So there is no technological stagnation, at least in these technologies. From available timeseries one cannot see it. For a given technology, progress tends to follow a constant law until it’s not possible anymore for it to progress, in which case a switch to a different technology happens.

Even for breakthrough innovation, harder to measure, I would say that it has not slowed down, even when some argue the rate of innovation per capita has declined (But population has grown, so it’s not that bad). And even if that’s true, it didn’t began in 1970, but in 1905, so it’s not obvious that it would be the cause of the post-70s great stagnation.

Further Questions

  1. Are there less ‘really innovative’ innovations (Huebner’s argument)? (h/t Ben Southwood)
  2. Is there stagnation in technologies related to the service sector? (h/t Austan Goolsbee)
  3. Absolute technological progress vs per monetary unit technological progress.
  4. Eroom’s Law, aircraft SPEED stagnation, and other trends
This entry was posted in Blog. Bookmark the permalink.