New results on total factor productivity and the import of creative destruction

by on October 13, 2017 at 1:45 am in Data Source, Economics, History, Science | Permalink

I am not sure I trust any TFP measures (what if innovation is simply embodied in investment?), but this paper by Gerben Bakker, Nicholas Crafts, and Pieter Woltjer is worth a ponder:

We develop new aggregate and sectoral Total Factor Productivity (TFP) estimates for the United States between 1899 and 1941 through better coverage of sectors and better-measured labor quality, and find TFP-growth was lower than previously thought, broadly based across sectors, and strongly variant intertemporally. We then test and reject three prominent claims. First, the 1930s did not have the highest TFP-growth of the twentieth century. Second, TFP-growth was not predominantly caused by four ‘great inventions’. Third, TFP-growth was not driven indirectly by spillovers from great inventions such as electricity. Instead, the creative-destruction -friendly American innovation system was the main productivity driver.

For the pointer I thank David Levey.

1 Steve Sailer October 13, 2017 at 2:19 am

My working memory is too old and worn out to comprehend this abstract’s double negative writing style. Are they saying “First, the 1930s _did_ have the highest TFP-growth of the twentieth century”?


2 Ray Lopez October 13, 2017 at 4:48 am

Yes SS, that’s what they are saying. It’s not even controversial, well known that the 1930s was a great period of rapid productivity, mainly because deadbeat “Oakie” type workers were fired, and electricity replaced steam, gasoline replaced oats.

The most controversial part of this report might be the implication that Gordon’s “four great inventions” (not defined, but probably electricity, airplanes, oil (replacing coal) and cars) was significant, “Third, we estimate that the technology clusters associated with Gordon’s
‘great inventions’ accounted for just under 40 percent of TFP growth … during 1899 – 1941”, which ignores the other 60%.

Bonus trivia: could the Great Depression simply be a routine Kondratiev wave that happened to coincide when electricity replaced steam, oil replaced coal, and the auto replaced the horse? And it lasted about four years, the bulk of bad things ended in 1934. Hardly anything to be traumatized about?


3 Todd K October 13, 2017 at 7:27 am

“… could the Great Depression simply be a routine Kondratiev wave that happened to coincide when electricity replaced steam, oil replaced coal, and the auto replaced the horse? And it lasted about four years,”

Look at the diffusion graph toward the top of this Atlantic article. The transitions were far longer than four years and your timing is off:

70% of households had electricity by 1929 and the number went down a bit but overall flat until 1935 when it started to climb higher at a slower rate than before..

60% of houesholds had a car by 1929 and then dropped to around 50% until 1955 when the percentage of households with a car was back to the 1929 level before growing to 70% by the late 1950s.


4 Anonymous October 13, 2017 at 9:15 am

All wave theories are false. Reductions for a small data world.


5 James D October 15, 2017 at 7:38 am

No, they’re saying that Alex Field’s famous claim that the 1930s was the most technologically progressive decade of the twentieth century is wrong.

See Field 2003 & 2012:


6 anon October 13, 2017 at 9:02 am

The “Not” lines are supposed to stand by themselves. From the Conclusions:

“We provide a detailed account of sectoral contributions to overall TFP growth which shows that TFP growth was broadly based during most of the period 1899 to 1941. It appears that TFP growth accrued across the economy from multiple disparate sources. This means that although the sectors which Gordon (2016) identified as comprising the ‘great inventions’ made a substantial direct contribution. averaging just under 40 percent of TFP growth they did not represent a dominant component of TFP growth. This conclusion survives when we take account of TFP spillovers across sectors.”

Similar for the other “Not”s


7 Andrew Smith October 13, 2017 at 3:32 am

It would be interesting to see Robert Gordon’s take on this wide-ranging paper


8 dearieme October 13, 2017 at 6:39 am

I did enjoy “strongly variant intertemporally”, a masterpiece of bad English.

I also laugh at the idea that they can confidently infer cause from a bunch of observations – and presumably noisy and inaccurate observations at that – on what is presumably a fiendishly complicated system.


9 JWatts October 13, 2017 at 8:27 am

I hope the writing is of higher quality in the paper than this snippet.


10 Normal October 13, 2017 at 8:50 am

Do economists really believe that progress came from “four great inventions”? I certainly hope they’re not teaching this kind of BS.

Electricity was not an invention, it was a discovery. It is Science, not Entrepreneurship.

Batteries, generators, transformers and motors were inventions. The “great invention” Electricity is in reality a scientific discovery, followed by thousands of inventions made by many people across the globe and over time.


11 kevin October 13, 2017 at 1:15 pm

I’m not sure what you mean by “progress”? The paper is about TFP, “mana from heaven”, which scientific discoveries seem a pretty good explanation for. Inventions are largely accounted for already in investment/capital deepening.


12 catter October 13, 2017 at 2:37 pm

Not simply electrification, but the full integration of individually-powered machinery into plant design. Similarly, not just cars, but the growth of supply chains based on trucking, made possible by better roads and vehicles.


Leave a Comment

Previous post:

Next post: