The Technium

The Post-Productive Economy

[Translations: Japanese]

Take a look at these farm houses which I saw under construction in remote areas of Yunnan province China. They were not unusual; farmsteads this size were everywhere in rural China. Note the scale of these massive buildings. Each support post is cut from a single huge tree. The massive earth walls are three stories high and taper toward the top. They are homes for a single extended family built in the traditional Tibetan farmhouse style. They are larger than most middle-class American homes. The extensive wood carvings inside and outside will be painted in garish colors, like this family room shown in a finished home. This area of Yunnan is consider one of the poorer areas in China, and the standard of living of the inhabitants here would be classified as “poor.”

Part of the reason is that these homes have no running water, no grid electricity, and no toilets. They don’t even have outhouses.

But the farmers and their children who live in these homes all have cell phones, and they have accounts on the Chinese versions of Twitter and Facebook, and recharge via solar panels.

This is important because a recent thought-provoking article by a renowned economist argues that the US economy has not been growing during the internet boom and probably will not grow any more than it has already because computers and the internet are not as productive as the last two industrial revolutions.


You can read the article here: Is U.S. Economic Growth Over? (PDF) by Robert Gordon.

Gordon answers his own question with: Yes, US economic growth is over for a while. I think Robert Gordon is wrong about his conclusion, but I wanted to start with one of the bits of evidence he offers for his view. He is trying to argue that the consequences of the 2nd Industrial Revolution, which brought to common people electricity and plumbing, was far more important than the computers and internet which the 3rd Industrial Revolution has brought us. (Gordon’s 1st Industrial revolution was steam and railroads.) As evidence of this claim he offers this hypothetical choice between option A and option B.

With option A you are allowed to keep 2002 electronic technology, including your Windows 98 laptop accessing Amazon, and you can keep running water and indoor toilets; but you can’t use anything invented since 2002. Option B is that you get everything invented in the past decade right up to Facebook, Twitter, and the iPad, but you have to give up running water and indoor toilets. You have to haul the water into your dwelling and carry out the waste. Even at 3am on a rainy night, your only toilet option is a wet and perhaps muddy walk to the outhouse. Which option do you choose?

Gordon then goes on to say:

I have posed this imaginary choice to several audiences in speeches, and the usual reaction is a guffaw, a chuckle, because the preference for Option A is so obvious.

But as I just recounted, Option A is not obvious at all.

The farmers in rural China have chosen cell phones and twitter over toilets and running water. To them, this is not a hypothetical choice at all, but a real one. and they have made their decision in massive numbers. Tens of millions, maybe hundreds of millions, if not billions of people in the rest of Asia, Africa and South America have chosen Option B. You can go to almost any African village to see this. And it is not because they are too poor to afford a toilet. As you can see from these farmers’ homes in Yunnan, they definitely could have at least built an outhouse if they found it valuable. (I know they don’t have a toilet because I’ve stayed in many of their homes.) But instead they found the intangible benefits of connection to be greater than the physical comforts of running water.

Most of the poor of the world don’t have such access to resources as these Yunnan farmers, but even in their poorer environment they still choose to use their meager cash to purchase the benefits of the 3rd revolution over the benefits of the 2nd revolution. Connection before plumbing. It is an almost universal choice.

This choice may seem difficult for someone who has little experience in the developing world, but in the places were most of the world lives we can plainly see that the fruits of the 3rd generation of automation are at least as, and perhaps more, valuable than some fruits of the 2nd wave of industrialization.

So if people value the benefits of computers and internet so much why don’t we see this value reflected in the growth of the US economy? According to Gordon growth has stalled in the internet age. This question was first asked by Robert Solow in 1987 and Gordon’s answer is that there are 6 “headwinds,” six negative, or contrary forces which deduct growth from the growth due to technology in the US (Gordon reiterates he is only speaking of he US). The six “headwinds” slowing down growth are the aging of the US population, stagnant levels of education, rising inequality, outsourcing and globalization, environmental constraints, and household and government debt. I agree with Gordon about these headwinds, particularly the first one, which he also sees as the most important.

Where Gordon is wrong is his misunderstanding and underestimating of the power of technological growth before it meets these headwinds.

First, as mentioned above, he underestimates the value of the innovations that the internet has brought us. They seem trivial compared to running water and electric lights, but in fact, as billions around the world show us, they are just as valuable.

So back to Solow; if digital innovations, millions of apps, the vast social networks that are being woven are increasing our living standards where is the evidence in the GDP?

I think the key sentence in Gordon’s paper is this:

“Both the first two revolutions required about 100 years for their full effects to percolate through the economy.”

Repeat: it took a century for the full benefits of the innovations to show up.

By my calculation we are into year 20 of this 3rd upheaval. Gordon wants to start the clock on the 3rd Industrial Revolution in 1960 at the start of commercial computers. That’s an arbitrary starting point; I would arbitrarily start it at the dawn of the commercial internet because I don’t think unconnected computers by themselves are revolutionary. Unconnected computers did not change much. Standalone personal computers hardly changed our lives at all. They sped up typing, altered publishing, and changed spreadsheet modeling forever, but these were minor blips in the economy and well-being of most people. Big mainframe computers helped the largest corporations manage financial assets or logistics, but a number of studies have shown that they did not elevate much growth.

Everything changed, however, when computers married the telephone. This is when ordinary people noticed computers. They could get online. Everything went online. Retail changed, production changed, occupations changed. This communication revolution accelerated change elsewhere. Processes and gizmos got smarter because they were connected. Now the advantages of personal computers made sense because in fact they were just local terminals in something bigger: the network. As the Sun Computer company famously put it: the network is the computer.

So the 3rd Industrial Revolution is not really computers and the internet, it is the networking of everything.
And in that regime we are just at the beginning of the beginning. We have only begun to connect everything to everything and to make little network minds everywhere. It may take another 80 years for the full affect of this revolution to be revealed.

In the year 2095 when economic grad students are asked to review this paper of Robert Gordon and write about why he was wrong back in 2012, they will say things like “Gordon missed the impact from the real inventions of this revolution: big data, ubiquitous mobile, quantified self, cheap AI, and personal work robots. All of these were far more consequential than stand alone computation, and yet all of them were embryonic and visible when he wrote his paper. He was looking backwards instead of forward.”

Finally, Gordon is focused, as most economists, on GDP which measures the amount of “labor saving” that has been accomplished. The more labor you save while making or serving something, the more productive you are. In the calculus of traditional economics productivity equals wealth. Gordon rightly points out that so far the internet has not saved a lot of labor. As I argue in my robot piece in Wired, Better Than Human (not my title), I think the real wealth in the future does not come from saving labor but in creating new kinds of things to do. In this sense long-term wealth depends on making new labor.

Civilization is not just about saving labor but also about “wasting” labor to make art, to make beautiful things, to “waste” time playing, like sports. Nobody ever suggested that Picasso should spend fewer hours painting per picture in order to boost his wealth or improve the economy. The value he added to the economy could not be optimized for productivity. It’s hard to shoehorn some of the most important things we do in life into the category of “being productive.” Generally any task that can be measured by the metrics of productivity — output per hour — is a task we want automation to do. In short, productivity is for robots. Humans excel at wasting time, experimenting, playing, creating, and exploring. None of these fare well under the scrutiny of productivity. That is why science and art are so hard to fund. But they are also the foundation of long-term growth. Yet our notions of jobs, of work, of the economy don’t include a lot of space for wasting time, experimenting, playing, creating, and exploring.

Long-term growth of that type that Robert Gordon studies is really weird if you think about it. As he notes, there wasn’t much of it in the world before 1750, before technological progress. Now several centuries later we have a thousand times as much wealth as before. Where does this extra good stuff come from? It is not moved from somewhere else, or borrowed. It is self-created. There’s a system which manufactures this wealth “out of nothing.” Much like life itself. There are certainly necessary conditions and ingredients, but it seems once you have those in place, the economy (the system) will self-generate this wealth.

A number of economists have wrestled with the origins of this self-generating wealth. Paul Romer and Brian Arthur both separately point to the recombining and re-mixing of existing ideas as the way economic growth occurs. This view focuses on knowledge as the prime motor in a self-renewing circle of increasing returns. Unlike say energy or matter, the more knowledge you spend, the more knowledge you earn, and the more breeds more in a never-ending virtuous spiral.

What is important is that this self-increasing cycle makes things that are new. New goods, new services, new dreams, new ambitions, even new needs. When things are new they are often not easy to measure, not easy to detect, nor easy to optimize. The 1st Industrial Revolution that introduced steam and railways also introduced new ideas about ownership, identity, privacy, and literacy. These ideas were not “productive” at first, but over time as they seeped into law, and culture, and became embedded into other existing technologies, they helped work to become more productive. For example ideas of ownership and capital became refined and unleashed new arrangements for funding large-scale projects in more efficient ways. In some cases these indirect ideas may have more long-term affect on growth than the immediate inventions of the time.

Likewise the grand shift our society is undergoing now, moving to a highly networked world in the third phase of industrialization, is producing many innovations that 1) are hard to perceive, 2) not really about optimizing labor, and 3) therefore hard to quantify in terms of productivity.

One has the sense that if we wait a while, the new things will trickled down and find places in the machinery of commerce where they can eventually boost the efficiency of work.

But it seems to me that there is second-order tilt in this shift to a networked world that says the real wealth in the long-term, or perhaps that should be the new wealth, will not be found merely in greater productivity, but in greater degrees of playing, creating, and exploring. We don’t have good metrics for new possibilities, for things that have never been seen before, because by definition, their boundaries, distinctions, and units are unknown. How does one measure “authenticity” or “hyperreality” or “stickiness”?

Productivity is the main accomplishment, and metric, of the two previous Industrial Revolutions. Productivity won’t go away; over the long term it will take fewer hours of human work to produce more of the goods and services those economies produce. Our system will do this primarily because most of this work will be done by bots.

The main accomplishment of this 3rd Industrialization, the networking of our brains, other brains and other things, is to add something onto the substrate of productivity. Call it consumptity, or generativity. By whatever name we settle on, this frontier expands the creative aspect of the whole system, increasing innovations, expanding possibilities, encouraging the inefficiencies of experiment and exploring, absorbing more of the qualities of play. We don’t have good measurements of these yet. Cynics will regard this as new age naiveté, or unadorned utopianism, or a blindness to the “realities” of real life of greedy corporations, or bad bosses, or the inevitable suffering of real work. It’s not.

The are two senses of growth: scale, that is, more, bigger, faster; and evolution. The linear progression of steam power, railways, electrification, and now computers and the internet is a type of the former; just more of the same, but only better. Therefore the productivity growth curve should continue up in a continuous linear fashion.

I suggest the growth of this 3rd regime is more like evolutionary growth, rather than developmental growth. The apparent stagnation we see in productivity, in real wages, in debt relief, is because we don’t reckon, and don’t perceive, the new directions of growth. It is not more of the same, but different.

Say we are watching an organism evolve. It might over time become extremely efficient in its energy use. Or it might become very large, cleverly optimizing its metabolic rate to manage its new girth. In either case, if we were biological economists watching it we would declare the organism to have grown in productivity.

But we can also imagine many other ways this organism might grow or evolve while keeping its metabolism steady. It could start complexifying, becoming multicellular. It might develop new sensors increasing its range of interaction. It might sift its reproductive strategy from making only one offspring garnering much attention, to making thousands of them with less care. It might evolve a tail and change its mode of locomotion entirely — all the time leaving its metabolism rate unaltered.

Our economy is moving into the latter mode of growth — an evolutionary uplift, which may or may not show an increase in productivity, particularly at the start of this phase. We see hints of this evolutionary growth already. The US economy shows:

Increased complexity — Derivatives, derivatives of the derivatives, flash crowds, dark pools of money, there are hundreds of new instruments and states of money.

Increased interdependency — National economies, particularly the US, are not longer independent cells, but part of a multi-organel system.

Increasing ubiquity of finance and monetization — More of our lives, from games to socializing to cooking to child caring, are now part of the greater economy.

Decreasing emphasis on ownership — In the parts of the economy run on information, data, and knowledge, these key ingredients can be used without owning them, and in fact often are more valuable when not “owned.”

There are many more, but these few demonstrate the way the economy is shifting rather than simply accelerating (although it is doing that too).

Technology will continue to increase productivity for the commodities of life, even if it takes another 80 years. But the next phase we are rushing into — the 3rd Industrial Revolution, the world of networks — the non-commodities of life will play a greater role in economic terms. When science fiction author Neal Stephenson laments: “I saw the best minds of my generation… writing spam filters” he should not give up. It’s not that different that the best minds of a former generation designing oil filters. These are the unglamorous but essential tasks in constructing a whole new infrastructure.

In his paper Robert Gordon talks about the huge value gained from “one-time” events, such as the one-time (first and last) move of a large proportion of women into the workforce. This new gain happens only once (assuming they remain). In this computer-internet economy we are experiencing a one-time gain from a huge one-time event. This is the first and only time a planet will get wired up into a global network. We are alive at this critical moment in history, and we are just at the beginning of the beginning of the many developments that will erupt because of this shift.

Happy new economy!


On January 23, 2013 Robert Gordon sent me a response to my critique of his paper after hearing about a podcast interview I did with Russ Roberts on EconTalk. I will post my reply to Gordon later; here is his letter in full:

Dear Russ (and hello Kevin aka KK)

While I don’t look at videos longer than two minutes (yours is listed at 58 minutes), I did download and print out KK’s essay in the Technium and have quite a few reactions.

The first 1/3 or so of KK’s essay features the Chinese with their well-built houses which are equipped with cell phones but no running water or toilets, indoor or outdoor. This is supposed to prove that, at least to some people, running water and indoor toilets are unimportant.

The commentators on KK’s essay immediately picked up the fact that he had not priced out the options. Cell phones are cheap and toilets are useless unless the neighborhood has been reached by urban sanitation infrastructure (pipes carrying fresh water and removing sewerage). This infrastructure was built in the U.S. between 1870 and 1930, and running water/ indoor bathrooms were almost universal in urban America by 1929. That is one of the greatest inventions of all times and overshadows today’s innovations.

But a more important objection to this irrelevant detour to China is that it has nothing to do with my paper. My choice of “Option A” vs. “Option B” explicitly states that option A allows you to keep everything invented up to 2002 and allows you to keep running water and toilets. All you have to give up is everything invented 2002-2012, including ipod, iphone, ipad, facebook, twitter, etc.

One of the things that existed in 2002, as it did in 1992, was the standard cell phone. This is what the poor people of China are using, not iphones. Just think of all the things you can enjoy with Option A – dumb cell phones, and on a desktop or laptop google, amazon, Wikipedia, full networked access to all the knowledge of the world.

The next section of KK’s essay claims that I “undervalue” the inventions of the third industrial revolution. Again, remember that I am giving a lot of credit to the computer for keeping the growth of productivity growing from 1960 to 2000 with the succession of inventions – computer printed bank statements and telephone bills (1960s), airline res systems and memory typewriters (1970s), personal computers, ATMs, and bar-code scanning (1980s), and then the marriage of the computer and communications in the 1990s. That is all given suitable appreciation in my story of what was invented before 2002.

Much of the rest of KK’s essay repeats the word “networking.” But this is really old stuff. I sent my first e-mail in 1993, 20 years ago, and was soon fully networked. My university web site was developed in the fall of 1998 and its design has not changed in 15 years (and it is quite unique, including 325 photos of economists, among other things).

But what KK misses is that the essence of networking was what the 2nd IR created. Virtually no house was connected to anything in 1870. By 1929, virtually every urban residence was connected to electricity, gas, phone, water, and sewer. The chapter in my book about the utter transformation of the living standard over 1870 to 1930 is subtitled “The Networked House.”

Then KK proceeds to criticize the concept of “productivity” by missing much of what is valuable to people. Let’s settle on several definitions. The standard of living is real GDP / population (Y/N). Productivity is real GDP / hours (Y/H). The growth of these can differ when H/N rises (female LFPR in the 1965-90 period) or falls (retirement of baby boomers and dropping out of lower income prime-age men now).

For KK to argue that real GDP does not capture the entire variety of human life, with his making art and playing sports, this is not a new idea but an ancient theme. The rising standard of living throughout 1870-2007 made possible a decline in H/N, much of which was taken as leisure time (far more in Europe than in the US). All this extra leisure time allowed human talents to flourish, and now we have “soccer moms” pushing their children into an endless round of activities from which my generation was exempt (I took piano lessons at the home of a teacher who lived in between my school and my house, which I walked every day. I was never chauffeured by my mother or father to any activity).

At a broader level, which KK does not consider, throughout history real GDP has understated the increase in the standard of living, because conventional price data do not give credit for the consumer surplus added by new inventions. KK would be right that smart phones are perceived as making the lives of their users better by more than the amount they have to pay for their data plans.

But the bias due to the undervaluation of the value of new goods goes way back to the beginning. The internal combustion engine created an entirely new “product” called “free personal travel.” Previously, ordinary people couldn’t afford a horse and its needs for stables and food and care in between uses. Suddenly there was this horseless carriage which did not use fuel when it was standing still. Suddenly ordinary families could venture outside of crowded cities and explore the country side. That was a “new product” as valuable as the iphone.

Some factoids are useful. The price of the Model T Ford declined so fast that in 1923 it cost only 15% of annual average personal income, and installment financing was available. Suddenly everyone could afford a motor vehicle, and the ratio of motor vehicles to households rose from 0 in 1900 to 85% in 1929.

The last section of KK’s essay is hopelessly vague. “this organism might grow or evolve.” If you want to talk about the value of new inventions, I highly recommend that you read Chapter 4 on the current state of medical care and technology in the following book:

Jan Vijg, The Technological Challenge: Stagnation and Decline in the 21st Century.

Vijg is the chair of the genetics department at Alfred Einstein medical school in NYC and is a distinguished scientist. His Chap 4 on medicine is worth reading. Much of the rest of his book is admittedly the contribution of an educated amateur, although I was impressed by his deep level of reading.

— Bob Gordon


© 2023