Tyler Cowen on the adjustment to growing automation

Despite the grim forecast, Tyler Cowen argues that western societies won’t collapse under the weight of future industrial change, but will eventually adjust to a new phase of ubiquitous automation—the period Brynjolfsson calls the ‘Second Machine Age’. However, he warns that evolution will take considerable time to play out.

‘If you look at the Industrial Revolution that starts in England say around 1780 and for a long time a lot of jobs do go away, wage gains are very slow, there’s a lot of volatility, it’s not really until the 1840s that real wages in England are going up significantly,’ he says. ‘So I think this time around it will actually be a lot like the last time. We will have a transition period of many decades. That will be tough for many people. In the very long run it will be splendid, but along the way it’s not always going to feel splendid. I think that is the historical pattern for a lot of these changes.’

There is more here, from Australian Radio National.


'Tyler Cowen on the adjustment to growing automation'

Is the royal we next?

For this New Dickensian Era, we'll just need some new Dickensian smartphone apps:


It's a bit strange for you to refer to yourself in the third-person like that, but that's just nitpicking.

Real wages didn't really go up in England overall because it wasn't until the 1830s and 1840s that you started to see industrialization and steam power penetrate all aspects of the British economy. Before then, they were impressive but mostly confined to a few particular sectors: textiles, steel production, mining, and so forth.

I'm not sure about the present wave. Theoretically they should go up if we get some strong growth and low unemployment, since that would lead to a tighter labor market (increasing wages and the demand for greater training of the workforce) and greater productivity (since the more expensive wages drive investments in labor-saving and complementary automation).

It seems somewhat obvious that the virtuous cycle you describe has broken down in the western world. It is at best stuck in "stagnation" mode.

One crucial issue is why Europe and Japan and now China have done so little to pick up the slack in US productivity growth and innovation. Despite predictions by Summers and others in the late 80s and early 90s of the US falling behind, Germany and Japan did not outpace the US in the leading edge industries of the 90s and since the crisis, no other major area has taken over from the US as a major technological innovator except at the margin. China has now had over three decades of catchup without successfully producing really important world beating companies. Recall that by the late 1970s -- less than three decades after Japan had been crushed in WW2 and derided for producing cheap junk -- Japanese autos and electronics were world famous and developing a reputation for high quality and reliability.

Well in the second quarter of 2013 Lenovo was the world's largest personal computer vendor by unit sales.

I'm not sure I'd call them a world-beating innovating company, but they are world-beating.

China has several problems - Chinese brand names don't sound good in the West (Lenovo is a better one, Huawei not so much). Also many large Chinese companies are SOE utilities and commodity companies. Finally, China is a world unto itself with a billion people. Sina Weibo has 61 million users, for instance.

One issue that's not mentioned a lot in these articles is that since the 1960s the thrust of regulation in the welfare states has been to protect the rights and privileges of the workers and to favor redistribution. Ironically this has had the long run effect of driving capitalists to substitute capital for labor and to focus on technologies that don't require so many hard to control, hard to train, and hard to fire human beings in less skilled positions. Just think of those Silicon Valley companies that Tyler likes to point out as unable to create jobs or the way that auto companies respond by a mix of outsourcing and robotics to labor unionization. Or even the way that airlines have tended to focus on better amenities and hardware in business and first class while avoiding big fights with the staff that is trying to hold onto contracts from the glory days of regulation. The more society tells employers that they "owe" the workers more, the more businesses will evolve to avoid workers altogether.

Ironically, or predictably?

At what stage, precisely, did capital owners not have an incentive to replace labour with capital? I'm not sure you could even ever remove that incentive in the first place.

I love how people feel the need to make everything into a morality play. So what you're saying is, because pro-worker reforms drove up labour costs they had a (apparently, much) greater incentive to automate the work away. Is the solution then to have worker costs so low there is never an incentive to increase automation? Is that socially optimal?

Seems to me that we're in a spot where returns on automation/capital are immensely outpacing returns on labour and that's ultimately a political problem.

I think you can do a broader analysis. All of the affluent market democracies faced a dilemma. Perhaps it was of "welfare, automation, or free trade" pick two. Picking three led to the imbalances we see today. One simply does not pick a costly welfare state and iPhones assembled for $132 a month in China (2010).

It depends on how much nonsense importers will put up with. Social Credit paid for with the printing press and modest inflation might work acceptably if everyone agrees to it. Its Keynes taken to the nth level but what can anyone do.

Thats partially true. Remember wages as GDP have declined by about half in the US (since peak America, 1973 or so) and in the minds of capital, lower costs and wages even when long term self destructive is always better. Forced redistribution to pay for the state incurred costs (education, infrastructure, public safety and national security) is often necessary since a corporation is inherently sociopathic by design

Also once facet of the mature industrial revolution that created the wage increases was pushing large groups out of the work force, 40 hour work weeks, women by social pressure and law to some degree leaving paid work (till recently) child labor and compulsory education laws (dried up cheap labor) and of course retirement. Oh yes and later immigration and trade controls. To get the same results you'd need state action along those lines

And note we do know what happens with high unemployment like we are going to see decades of population aging and decline. This is quite obvious, no income no babies but do to cognitive blindness and issues with cultural training people don't want to address it. Simply, many people in the world, enough that it is a passable majority in the industrialized and parts of the developing world have figured out that the best way out of poverty is smaller families. People don't want to live on beans and they don't wan their future offspring to either

The problem is that optimum number is well below human death rates. This is fine and even good for a while but how long does anyone want this to go on? You can't sell to people who aren't born and if corporations are unable or unwilling to participate that leaves either private autonomous enterprises like Robb's resilient communities that are not par tof the systems or some kind of state intervention such as social capital or a welfare state or worse a resurgence of fairly virulent illiberal ideologies.

And note yes there are more natal groups of immigrants , the US has its Hispanics (now shrinking though) and there are other groups such as Africans and the lot. The net result of wide scale immigration isn't an economic boon and it isn't vitality. Most of the groups either adapt to the lower fertility level or continue with low investment habits that limit the number of high climbers in that group. Neither solution fixes the underlying issue, a lack of investment in human capital.

So this graph http://research.stlouisfed.org/fred2/graph/?g=2Xa would trend up or stay flat were it not for the government making companies do what exactly?

American workers are also (mostly) American consumers.

If they cut their staffs down to nothing, how will these businesses survive?

Do they think the government can step in and supply the unemployed and underemployed with emough to live on -- and to continue buying?

We don't think that parallels with societies of the early 19th Century are valid. Today's spoiled children won't stand for decades of stagnation. We are of the opinion that troubles and instabilities will spoil the predicted techno-utopia.

The question is: will they stand for decades of 'perceived stagnation'?

The answer is yes, but this is why I'm less of a happiness optimist than Tyler.

How confident are we of data that show weak wage growth in the early 19th century, given that we can't agree on the last forty years?

The Tyler Cowen of George Mason University (Google "Tyler Cowan Masonic" and see what you get, or go here: http://www.urbandictionary.com/define.php?term=Cowan) model does not account for how intellectual and capital deepening makes science and technology move faster and take less time to achieve results, what is known as "catch-up" effect in convergence economics (by analogy).

Hence rather than 60 years to achieve a breakthrough and/or saturation, it will be 6 years.

Ray Kurzweil's singularity is coming and will be proved right, just like it was in computer chess, you watch.

Except most of Kurzweil’s predictions have been hilariously wrong so far. According to him, five years ago we were supposed to have: cybernetic musicians that played with human ones, cybernetic authors who were growing in popularity, autonomous nanobots, chat screens that would allow us to undress the other person on our screen without them knowing it, and the main entertainment for most people would be virtual reality environments. The fact that he stands by most of his predictions should be enough for most people to dismiss him as a charlatan.

The old gag was that "AI has been 20 years away for 20 years." Now it's been more like 40 years, and Kurzweil is still calling 20 years.

If we're willing to be a bit lax on what we consider "cybernetic", then he's not all that wrong. If content creation is algorithmicly driven these days, that surely sounds like a cyborg to me. And certainly the video game industry is far larger than all other entertainment industries.

Population grew considerably over that period though. If incomes and population don't grow, it would make this time worse.

Um...doesn't a growing population put downward pressure on wages?

You actually just made me more optimistic.

You know how people say, "some day, we'll look back on this and laugh"* in the midst of some misfortune? This snippet has a certain feel of "in the long run, everything'll work out," if we're just patient enough.

(* Actually, I personally don't say this: the only such instance in my own life was when my parents took the wrong turn on the highway in Italy and, instead of Trento, ended up at Trieste, and unknowingly nearly at the Slovenian border, before turning around.)

The problem is that we have to make it through the short- and medium-term before we can arrive at the long-term that Tyler says will be "splendid." There are real people who are impacted, in the same was as Luddites were not just irrational but reacting to their specific circumstances. And I think a lot of the potential impacts have been mitigated by the jobs the government has "created" (e.g., the growth in regulation requiring increasing number of workers dedicated to compliance) and by our trade and budget deficits, which are not sustainable in the long term. And I just haven't seen much in policy ideas or economic analysis that provides a way through this other than treating people as collateral damage. http://janetheactuary.blogspot.com/2014/02/piecing-together-minimum-wage-obamacare.html

Incidentally, Cowen pulls a quote from a source, without any introductory comments, rather than writing in the third person. So let's give him a pass on this.

The small problem is that as one can see via Ted Talks is that neither Cowen nor Gordon know the slightest thing about what has occurred technologically over the past 200 years. Like Krugman, they have not taken *one* rigorous science course in college. But hey, don't let that stop them from putting up pictures of toilets and airplanes as an argument for no change.

Is it valid to compare the industrial revolution with today's automation revolution?

In the industrial revolution the machine owners needed workers to operate the machines. But automated systems don't need workers to operate them.

When machines are smarter than humans why won't humans become like horses? The internal combustion engine built into vehicles replaced horses for most uses and horses did not eventually gain lots of new uses.

The designers and owners of robots and the owners of natural resources will have things to trade each other. Only taxes will generated the money to buy goods and services for most people.

Migrant workers are mostly farmers who shuttle between their rural homes and cities looking for work. They usually take the least-paid and most laborious jobs in cities. According to Liu, who herself is a migrant worker, older migrant workers are more likely to be victims to rights abuses due to age issues and poor educational background.

Comments for this post are closed