February 7, 2011

The great stagnation of inventions, in two charts

Tyler Cowen wrote an article about stagnant growth in the standard of living for people in the developed world, which he develops at greater length in an e-book that I have not read, called The Great Stagnation (link in the article). If you search the Marginal Revolution website for "TGS," you'll find lots of links to responses to his overall thesis, namely that sometime in the mid-1970s the dizzying growth in our standard of living that began with the industrial revolution had started to plateau.

I have not read all of them, but have browsed them, and read a fair number of comments in those posts. A lot of the focus is on technological innovation, admittedly just one piece of how well we have it, but still an important one -- and certainly a more easily quantifiable one. Cowen's critics grasp for recent examples of new gadgets that are surely just as dazzling and enriching as yesterday's doodads were.

However, nothing that I've read has made that explicit comparison -- is the iPhone really as revolutionary, against the recent background, as the telephone was, against its recent background? No way, Jose. Most techie geeks, like geeks in general, are completely uninterested in history, and so lack any frame of reference for today's stream of novel doohickies. It should be obvious, and Cowen in that NYT article even points out several such improvements in the standard of living that totally swamp the adoption of the internet or Skype or containerization -- electricity, the automobile (the railroad for that matter), and so on. Always seeking shelter from the real world, techie geeks are incapable of appreciating how much greater life became after the introduction during the 19th C. of something as seemingly mundane as matches.

To quantify the obvious, I went to the book 1001 Inventions That Changed the World (see the link for details on who wrote it and their methods, or pick up a copy). I tallied the number of inventions important enough to make it into this book and smoothed these data with a 10-year moving average. For most of recorded history before the industrial revolution, there's a very low and steady rate of inventions made each year, so I only plot from 1700 onward (click to enlarge):

You can clearly see how runaway the rate of invention was when the industrial revolution got going. However, even by 1900 that rate had leveled off -- the peaks after 1900 are not taller than the one around 1880, and the troughs are not consitently or substantially higher than the one around 1890. Still, you can see a mini-boom in inventions from about 1920 to 1970 -- the age of the transistor, the computer, and most of the things we consider hi-tech. There is also a very clear steady downturn starting around 1985 -- probably not coincidentally after Bell Labs, a powerhouse of 20th C. invention, had to close down when AT&T was busted up.

As Cowen mentions, even during periods of long-term increase in inventions, there are still cycles up and down around that trend. Nevertheless, it looks like even that secular trend upward has been reached by now, and we will only be cycling around that plateau level.

By the way, the recent decline cannot be blamed on it taking awhile to recognize monumental inventions. First, there are decline periods far into the past as well, so we don't need to appeal to a "blindness to recent inventions" bias. Second, the bias obviously goes the other way -- techie geeks who compile lists of the Top 100 Awesomest Inventions always over-hype the very recent past. Therefore, such lists will always give a fair number of false positives for the recent past -- declaring something important when it will turn out to be trivial -- but never a false negative -- overlooking something that is earth-shaking.

Another way to understand it is to ask, How far would we have to zoom out in time in order for The Great Stagnation to disappear? Instead of lumping inventions in year-by-year bins, I lumped them into decades -- still the recent decline shows up. How about half-century bins? Now there is no recent downturn, but a plateau effect still remains. 75-year bins? No again. Only when we look at how many inventions were made per *century* does the growth look as strong during "recent" times compared to the next-most-recent times (i.e. growth during the 20th compared to the 19th C.):

If you squint, you can see that even here the slope is a little shallower for the most recent period than the next-most-recent period, but it's close enough to declare no plateau, let alone a decline, in recent times. However, this is not what the critics are arguing -- that 20th C. growth was just as impressive as 19th C. growth. They're trying, and failing, to show how frenetic growth has been during the past several decades compared to the earlier decades.

A final, non-quantitative way to think about how little things have changed in the big picture since the 1970s or '80s is to do a thought experiment: think of some invention after the cutoff year and ask how radically it could have changed history if it had been introduced 100 or however many years ago. Then do that for an invention a bit earlier than the cutoff date. Better yet, take a handful of inventions on either side, then compare which set could have more dramatically altered the course of history. Sending cell phones into a world with only landline phones (not even cordless ones, without answering machines, and without the ubiquity of pay phones for on-the-go communication) would not change things as radically as sending the telephone into a world with only letter writing and no high-powered means of transportation to deliver them.

Or imagine traveling back in time and showing the average person what shiny new toys you've got in the future, and ask how amazed they would be. Even as recently as 1985, when Back to the Future was made, someone living 30 years earlier would have been astounded by a portable music player (or "personal stereo," as Walkmans used to be called) and a camcorder, which Doc Brown calls something like a portable television studio. (This would be even more dramatic by sending 1955's technology to a 1925 consumer, of course.)

But then take something fairly common from 2010 like a phone that can take pictures on-the-go and show it to someone from 1980 -- meh, that's kind of neat, but we can already make phone calls just about anywhere and anytime we want, and we already have affordable cameras and even instant cameras. It's a cool improvement, but we were hoping for flying cars, X-ray specs, virtual reality -- something in the realm of "just imagine the possibilities."

There's a lot more to say about the topic of stagnation since the '70s or '80s, and you can search this blog for "Malthusian" to find some relevant earlier posts. But the rate of trailblazing invention isn't too hard to quantify, making that prong of Cowen's argument a rather open-and-shut case. Still it's worth illustrating with graphs and qualitatively since most people who get worked into a tizzy over the stagnation idea have no conception whatsoever of what life used to be like much earlier than when they were kids.


  1. Just shooting from the hip here, but it seems as though the last few years (let's put that around 10-20) have been not so much about making new things, but making old things better. For example, Apple's iAnything. Great devices and all, but not so much new territory is being explored, just variations/expansions on a pre-existing theme.

  2. sometime in the mid-1970s the dizzying growth in our standard of living that began with the industrial revolution had started to plateau.

    Right around 1973.

  3. You've mentioned the closing of Bell Labs before. I wonder if that's just a data point (an important one) in a general decline of the science/engineering research culture in the West, especially the U.S.

    I see two things happening in 1970s and 80s.

    First, the rise of the "megaversity" - the massive expansion of the universities and change in the ethic by which they were governed. They started being by run by "suits", not scholars, who obsessed over funding and "efficiency" (e.g., the grad-school proletariat), and were happy to act as intellectual whores (Pomo, identity-group studies, corporate-or foundation-funded directed research) for publicity, to court politicians and the bien-pensants, and for a little extra cash.

    Research projects became extremely large and expensive, and therefore tended to be run by "courtier" types, adept a making connections and securing grants. The granting system itself: committee-based... need I say more?

    There's demographic side to the megaversity too. It's reasonable to assume that staff hired during the massive expansion of the 1960s and 70s were lower in quality than those hired in a more selective era. (And, culturally, they brought with them all those horrible 68-er ideas.)

    By the late 80s, early 90s, this clump was well past its peak research years and was blocking the entrance of new talent into the system. A researcher's peak years in a lot of fields are age 25 to 40, yet a lot scholars were unable to a proper, secure position until they were 35 or so.

    Outside the university, a lot of economic gains we've had have come about through rationalization. I think there's been a shift to managerialism in a lot of formerly engineer-driven companies. If you follow Schumpeter, you'd expect this drive for efficiency to squeeze out the space needed to incubate longer-term, serendipitous innovation.

    (Google, apparently, lets employees spend 20% on side-of-the-desk projects. Most companies would see that as irrational. Most companies are like Walmart: squeeze out every inefficiency, transfer all creativity and initiative to head office.)


  4. I run in some fairly geeky circles, on occasion, and I'm well known for my pessimistic view of how un-life-changing their latest gizmos are. When I say that it saddens me that I've never lived through any revolutionary changes, as my parents or grandparents did, the geeks all look aghast and quickly try to run through the tired litany of what I consider improvements, not true inventions.

    I had access to a Commodore 64 and then we later owned a Tandy 1000. Today I'm typing on a ThinkPad. The disparity in computing power is obvious, but the actual function of the computer is the same.

    That's the big one. A quick list of things you'll read as the "Top 100" inventions/innovations of the last 30 years, and their equivalent:

    DVD/CD: VHS/Cassette
    ipod: Walkman
    Satellite TV: Satellite TV (remember the giant ones?)
    4G smart phones: land-lines and 1G bag and brick phones
    LCD and Plasma TV: CRT
    Digital Photos: Polaroids and disposable cameras
    Game consoles: Atari
    (and the list could go on)

    My point isn't that these new things aren't awesome, it's just that they haven't fundamentally changed how I live my life now from when I was a kid. They have been more of a shift from black and white television to color, rather than from no television to black and white.

  5. Agnostic, I have an unrelated question/comment.

    I'd like to collect data similar to what you did to investigate academic fads back in 2008 on gnxp. Did you essentially count the number of hits for each year to get the data or did you have a script or some other clever solution?

  6. Chief Seattle2/7/11, 3:07 PM

    Most of the economic energy of the last 30 years or so has been directed towards making things cheaper and more accessible to non-experts - not making them function better. Whether that's airplane deregulation (a plane flight is almost exactly the same as in 1980, but it's far cheaper in real dollars) or electronics (I just paid $650 for a mac mini, almost exactly the same as my Dad paid in 1984 dollars for a Commodore 64 with tape drive). Until recently food prices hadn't increased in nominal dollars for 30 years - and I can still get a gallon of milk for $2.49 without looking hard, not much more than the $1.75 that was common 30 years ago. Part of those cost decreases is productivity, but a big part is also lower labor costs, either immigrant or overseas. So the big economic change of the last 30 years has been a flattening of wages and product availability across the world.

    This efficiency and lower labor costs has been in a sort of feedback loop with technological advance. When it's easier to make a buck by moving a factory overseas than by investing in new product or processes, that's where capital and entrepreneurial energy will go. That reduces technological advance, which reduces the availability of new products and processes and keeps the focus on efficiency.

    But it feels like we're nearing the end of that cycle. There's only so much efficiency to squeeze out without breakthroughs. The former third world is seeing inflation and demands by labor for wage increases.

    One possible other reason for stagnation - there hasn't been a good war or even a good proxy war since the Soviets collapsed in '89. War drives a lot of spending and innovation and removes a lot of bureaucracy. Too bad it's gotten so darn destructive.

  7. The PLATO Network was started in 1960. The PLATO network was a network of connected computer terminals similar to the network of PCs today with the internet.

    A lot of the computer stuff and things people do online today that are considered novel and revolutionary like plasma displays, touchscreens, forums, message boards, education, online testing, e-mail, chat rooms, instant messaging, remote screen sharing, multi-player games, virtual economies, "social media," etc. have been around since the 60s and 70s when they were developed under the PLATO system.

  8. Underachiever2/7/11, 10:51 PM

    The weakness with your argument is that great inventions can sometimes only be judged after the fact.

    Some of your argument (and the preceding anonymous') is the idea that we can only credit the original invention of a concept and ignore the technology which builds it up. Therefore, anonymous takes the existence of the PLATO network to dismiss the internet. That is a fine argument as far as it goes, but you have to follow it all the way.

    If the internet (because of its predecessor PLATO) counts as a great invention of the 1960s, then our own time gets several great inventions. We must give ourselves credit for every development which ever comes from stem cell research, gene sequencing, and genetic engineering because they are all being done to a limited extent today and were invented by us.

  9. "Did you essentially count the number of hits for each year to get the data or did you have a script or some other clever solution?"

    I should've written a script but never spent much time on a given chart, so just punched them in by hand in JSTOR.

  10. If you think the cell phone is not revolutionary, travel to a developing country. The land line phone is terribly ill suited for much of the world and for more than a century failed to reach potential consumers. Today there are millions of people in developing countries that have never used a land line but find cell phone economical, useful, practical, and essential. Tyler Cowan's notion that innovation is slowing down is myopic and wildly off the mark.

  11. lemmy caution2/9/11, 3:07 PM

    You are probably right about the decline of big inventions in the last few decades.

    Here is a book from 1918 on great inventions:


    It includes all of the then recent great inventions, movies, radio and airplanes. These were all invented around 1910 and by 1918 were pretty popular. The book does include a lot of inventions of the 1800s that don't really seem that big of a deal now.

  12. I think the invention of the World Wide Web and easy to use browsers, such as mosaic and the much more widely popular Netscape have been pretty damn transformative.

    Yeah they sat on the back of the internet which was invented through DARPA funding in the 1960s, but until the WWW and browsers it wasn't a very widely used thing. It was used by our military and university computing departments and research labs mostly in the 60s and 70s. Oh there were bulletin board services like Compuserve in the 1980s, but that was nothing like the huge importance the web has today for most college educated people at least, and many who aren't.

  13. So you used some author's subjective evaluation of whether or not an invention was sufficiently important to be "big" invention to make a chart, and we are to treat said chart as empirical evidence of anything?

  14. You failed to follow my instructions to see the book for info on who compiled them and on what basis. They are authors, not "some author" in the singular, dumb shit, and we are to treat your comment as a refutation of anything?

  15. in anon's defense, amazon fails to provide much relevant information about the book. the editorial notes some of the content included and some of the omissions. a quick google search yields that 50 historians, designers, scientists and anthropologists worked on the book, and that they had difficulty narrowing the list to 1001. It still doesn't say anything about their methods for identifying the inventions or for cutting that list to 1001. But that’s ok; the editor doesn’t seem to claim that these are the 1001 MOST IMPORTANT inventions ever. If we’re simply looking at invention quantity, the US patent office provides good data indicating that patent application/issued has increased 5-10x since 1963. (http://www.uspto.gov/web/offices/ac/ido/oeip/taf/us_stat.htm) I admit that most of these parents aren’t useful, and aren’t physical objects. but it’s still a testament to real development that our society produced.

    I can’t find the authors’ Materials and Methods section or list of inventions, so I can only speculate that the authors’ bias: it probably doesn’t include many industrial/biochemical/engineering/statistical/political techniques. Things that you can’t see. Haber Bosch process, polymerase chain reaction, tandem mass spec. Or are they mere improvements?

    I think that differences in degree, when compounded, become a difference in kind. Compare a modern suspension bridge to a fallen log. I can’t imagine any bridge being fundamentally different from the version that directly precedes it, but what we have now is completely different from version 1. Which invention do we credit? Version 1? The first one that doesn’t suck? The one most useful by 2011 standards? We can’t credit all of them; there must be at least 1001 bridge technologies alone.

    I don’t agree with your argument that the improvements we’ve made to today’s gadgets are inconsequential. “But then take something fairly common from 2010 like a phone that can take pictures on-the-go and show it to someone from 1980 -- meh, that's kind of neat, but we can already make phone calls just about anywhere and anytime we want.”

    Agnostic and Camel, the new “improvements” have fundamentally changed the way people live their life. Cell phones allow people moving around to communicate with other people moving around. Whereas pay phones allow people moving around to communicate with people who are stationary. This drastically affects the way people plan (don’t plan) a night out ahead of time. You previously described the effect of mobile broadband on sports trivia impressiveness. Though I suppose this is more a testament to technology deployment than the invention. It also affects the way people do business. I can check prices of things while I’m interacting with the other partner.

    Dynamite is not just a better gunpowder; Computerized Axial Tomography is not just a better xray; total parenternal nutrition is not just food. Being able to tunnel through mountains or detect more diseases is a new function, a new invention. The functionality differences between my computer today and my computer in 1995 is the difference between that and an etch-a-sketch. The processing speed and networking allows coordination and information distribution previously not possible. I can be connected to my friends whenever I want. I no longer have to listen to what content is being broadcasted to me by a central media; I can listen to whomever I want. I ask you a question: could the revolutionaries have overthrown Mubarak using 2000 level information technology?

  16. Since our perception of time and its reach are so dwarfed by the accumulated relics of history, can we ever know what is truly historical?

    The internet (or whatever electronic method you argue shrunk mankind's reach) is unquestionably revolutionary.

    The question is what dictates what.

    Do our inventions follow a pre-assembled mental evolution or do our inventions steer evolution?

    I think this is the question we need to ask.


You MUST enter a nickname with the "Name/URL" option if you're not signed in. We can't follow who is saying what if everyone is "Anonymous."