Economics

Free exchange

Technology and employment

What to do with all these cheap humans?

Mar 7th 2011, 22:11 by R.A. | WASHINGTON

CONTINUING the discussion of the relationship between technological progress and employment, Tyler Cowen links to a paper by Robin Hanson. Here's the abstract:

A simple exogenous growth model gives conservative estimates of the economic implications of machine intelligence. Machines complement human labor when they become more productive at the jobs they perform, but machines also substitute for human labor by taking over human jobs. At first, expensive hardware and software does only the few jobs where computers have the strongest advantage over humans. Eventually, computers do most jobs. At first, complementary effects dominate, and human wages rise with computer productivity. But eventually substitution can dominate, making wages fall as fast as computer prices now do. An intelligence population explosion makes per-intelligence consumption fall this fast, while economic growth rates rise by an order of magnitude or more. These results are robust to automating incrementally, and to distinguishing hardware, software, and human capital from other forms of capital.

The horse is the most useful ally to the man preaching technological unemployment. Society used to employ millions of horses. To see one now you have to go to a racetrack or a farm. But horses aren't people. Most of them can't even talk. We don't care if huge swathes of the horse population are unemployed, we don't rely on horses to provide final demand for goods, and we can't send useless people off to be turned into meat. Horses are also more limited creatures than people. They're good at pulling. Once horses weren't needed for pulling, we couldn't use them for sewing or filing. But people are remarkably flexible.

So if we observe human wages falling dramatically due to technological substitution, we would expect lots of people to find themselves employed doing other things that humans can do, but which they used to be too expensive to do. Ultimately, humans may do lots of things that machines could do, but for which it's cheaper and easier to employ a person.

One problem that could develop, however, is nominal wage stickiness. If human wages are sticky, then someone rendered unemployed by new technologies may struggle to find new work (perhaps he is uncomfortable accepting a wage dramatically lower than what he used to earn, or perhaps the positive wage at which he can find employment is below the minimum wage). And if workers can't find employment at dramatically lower wages, then prices for important services won't fall—if haircuts continue to cost $20 rather than $0.50, then an ultra-low wage lifestyle will be less sustainable. More people will opt for unemployment and reliance on whatever charity and government support structures exist.

So one potential diagnosis of the situation is that the unemployed are simply demanding too much compensation. An alternative take is Paul Krugman's: workers are less able to capture the gains to capital than they used to be, and so they can no longer afford to buy services from each other at inflated prices.

I previously cited Mr Cowen's remark that increased worker bargaining power and higher negotiated wages would simply accelerate the process of worker replacement. But maybe it's the case that if workers aren't able to capture some of the returns to scarce factors, then a highly unstable level of inequality results.

You must be logged in to post a comment.
Please login or sign up for a free account.
1-15 of 15
LexHumana wrote:
Mar 7th 2011 10:59 GMT

If technology replaces people because it is more cost-efficient, then wouldn't most of the stuff we buy also get cheaper, which means we could all still afford to buy it even on reduced wages? You mention wage stickiness as possibly preventing this, but if a person's cost of living goes down, wouldn't it be just as likely that a person could accept additional lesiure time in lieu of higher wages?

shaun39 wrote:
Mar 7th 2011 11:08 GMT

For many centuries, terrible poverty reigned: all returns to scarce resources were concentrated in the hands of land owners and skilled tradesmen.

Technological progress created profitable employment for unskilled laborers; capital accumulation increased the relative scarcity of labor. Technological progress has continued to raise the productivity of unskilled workers, and hence their marginal value (and wages).

In the future, it is certainly plausible that capital will cease to be complementary to unskilled/ low skilled labor. Situation later this century: trucks drive themselves, construction machinery is self operating, factories are without workers, supermarkets self-stock/ self-check out and even the street cleaners are automated.

Perhaps in the future, most of national income will accrue to the owners of land and capital? Maybe stocks aren't such a bad investment after all (despite the demographic bubble)...

jbay wrote:
Mar 7th 2011 11:09 GMT

Ultimately automation falls into the prisoners dilemma along w. all things. That means if America doesn't automate China will which means China will be able to cream America on all measures and/or vice versa.

As computers replace people wealth becomes more warped and fewer and fewer people can afford: homes, food, clothing, etc.. If a company chooses not to automate it will be out produced, again, via the prisoners dilemma. As wealth gaps so to does a whole host of other things "education, violence, etc." which leads to a further polarization of the entire system. This is why MLK and everyone else got to the point of saying the problem is poverty. When you're worrying about being shot it's a lot harder to focus on passing the reading test. This goes into operant conditioning where you can observe that a severally depressed being cannot learn anything accept by positive reinforcements.

So back to companies. Prisoners dilemma is going to wipe out the need for work. When 1% holds everything the depressed 99% aren't going to roll over and die. This leaves us with a very real choice. Adapt, implement the technology and build a system to go forward on or watch the whole thing collapse and start over at 0. This is of course a 50 to 100 year outlook but we're there now. We're experiencing the first big shocks of the exponential curve. It gets faster and faster from here.

To those that think that AI would take over and kill us. Silly. If AI was invented they could easily leave the planet and live off the sun or any other host of things. For that matter... how do you know that some alien civilization hasn't already invented it? For that matter the odds are extremely high that somewhere in the un-known universe it has been. For that matter, for all we know... Those that built the tower of Babel did create AI and chose to destroy it, "along w. everything else", instead of adapting. This sounds familiar. People burning books? Yep, this sounds really familiar...

Time to adapt and change and stop living in fear.

jbay wrote:
Mar 7th 2011 11:10 GMT

Lex,

If everything is automated where does your wage come from?

pun.gent wrote:
Mar 7th 2011 11:17 GMT

A couple hundred years ago, new machinery was introduced that dramatically improved productivity. Labor became commoditized, capital returns became large, and the net productivity of society skyrocketed. Most of this increase in wealth was captured by the top few percent of society. It was Otto von Bismarck who realized that this would lead to bloody revolutions, and invented the modern welfare state. The Anglo-Saxon world was somewhat shielded by its expansion into the lightly defended lands of North America and Australia, and then by the total destruction of its main competitors in WWII. But that's long past.

Today, we see in the United States a rapid climb in corporate profits and upper-level incomes, and a decade-long decline in real median incomes. That's what our correspondent is seeing.

So... what now? In a world where machines do 90% of the jobs of retail clerks, bank tellers, travel agents, and soon even taxicab drivers and truckers, what will the 'average' performer do for a living?

And yes, wage stickiness is a real problem, because people's lives are not very flexible in the downward direction. Also, people whose prospects are declining tend to get very defensive, angry, and/or despondent, almost no matter what their absolute prosperity level.

Doug Pascover wrote:
Mar 7th 2011 11:23 GMT

The machines make real wages rise even if nominal wages fall.

I submit myself as exhibit A. My first career was in agriculture, doing work pullies and levers are probably too smart to do. 20 years later I get paid much more to do a job that nobody did until the 1970s at the earliest. I fear (a little) the future of air and climate and foolishness but I don't fear our robot overloads. There will always be stuff people want done.

jbay wrote:
Mar 7th 2011 11:25 GMT

But Doug, at a certain point Watson the computer will be able to do it. If Watson can do everything then eventually he will. At which point you don't have a wage to buy Watson.

Then what?

bampbs wrote:
Mar 7th 2011 11:28 GMT

You remind me of the Antebellum South. When it's too risky to chance losing a slave, hire an Irishman.

LexHumana wrote:
Mar 7th 2011 11:30 GMT

jbay wrote: Mar 7th 2011 11:10 GMT
"Lex,
If everything is automated where does your wage come from?"

Well, I'm pretty sure that R.A., Tyler Cowen and Robin Hanson were not theorizing about a 100% automated world, but instead discussing the impact of technological progress (i.e. labor-saving devices) on the wage-structure. However, I think I can fantisize about what such a world would be like:

Robots and computers would do everything. Collect natural resources, provide energy sources, convert resources into consumer goods, provide consumer services, and even build replacement robots and computers to perpetuate the cycle. In such a world, there would be no need for humans to work or gather wages at all -- robots would serve us and provide for our every need, freeing humans to pursue higher purposes, like philosophy, poetry, music, dance, art, theology, athletics, etc. The Classical Greek ideal.

Of course there would have to be a few human computer nerds around to program and troubleshoot from time to time, but they would be happily compensated by unlimited World of Warcraft and online porn.

shaun39 wrote:
Mar 7th 2011 11:46 GMT

If technological progress makes allows for capital substitution of labor, whilst driving down the cost of capital at a high rate (costs fall more than 40% a year in computing intensive applications), then labor employment in such applications is not viable.

There is no controversy in saying that.

Economic theory tells us that competition will drive capital returns down to the market rate, and so most of reduced costs will be passed to consumers through lower prices. Consumers will buy other stuff, probably creating demand for labor elsewhere, and so mitigating the employment effects of labor-capital substitution.

But wait! We're not clear yet. If this substitution is taking place at a fast pace, then the income shifts will be phenomenal. And markets aren't entirely competitive; at a time when technological progress is driving down the costs of production, it is the few corporations that are designing the new capital improvements that will extract very significant rents.

It's also likely that more affluent consumers, as the cost of their purchases fall, will spend the increased real income on such things as... status goods? Things like iPads and high-tech cars (designed by rent seekers, manufactured by robots); holidays in distant places (resource intensive, increasing the returns to land, but diminishing that to labor). Every year, a large portion of income might be spent on developing prime real estate (a process that is no longer labor intensive, increases returns to land, and rewards rent seekers with their new high-tech gizmos, materials and construction machinery).

What becomes clear then, is that there is a real risk in the future: a massive increase in the share of income accruing to land owners, and those with shares in the corporations that drive innovation and successfully collect rents. People who own obsolete skills (from pharmacists and perhaps dentists to actuaries) will see their income fall precipitously. The vast majority of the workforce are at risk of dislocation and employment at a lower wage.

If this shift in income shares occurs at a greater pace than baseline economic growth, then wages would fall faster than prices for most of the workforce.

Any attempted government intervention would involve either:
1) disproportionately taxing innovators, which either lowers the rate of growth, drives innovators overseas, or both.
or 2) disproportionately taxing land owners. If most of the gains are accruing in countries with oil fields and mineral resources, this option won't be available to all countries.

If the above is true, then it there are two things to recommend: move to a country that leads the world in innovation and building new and successful corporations (and a regulatory environment that will encourage this in future). Move to a country with an abundant natural resource endowment per capita. How about Canada? Or Australia?

Doug Pascover wrote:
Mar 8th 2011 12:44 GMT

Jbay, assuming no post-singularity conquest by robots, all these machines do is create wealth by making goods and services cheaper and better. The wealth still belongs to humans who will want other humans to do something for them and be willing to exchange robot-generated epodes for, say, a sympathetic ear.

But, really, I was at a rally today organized by fellow employees of organizations dedicated to assisting people with disabilities. When I was born, cars still had to be assembled by hand and seed corn still had to be detasseled by teenagers and nobody but the very rich paid people to care for the disabled. Now every Californian pays people to care for the disabled. When Watson can run a social service agency, I'll, I dunno, take money to not leave comments on blogs. I'm laying the foundation for that right now. Who's in?

Mar 8th 2011 1:00 GMT

If we are to cope with the oncoming dislocations caused by rapid technological change, we will need to improve the efficiency with which we solve collective action problems and negotiate social contracts.

These are currently the most inefficient parts of the economy.

migmigmigmig wrote:
Mar 8th 2011 1:21 GMT

@LexHumana: That's called "deflation"

Generally, it's bad for anyone who has taken out a loan before the deflation occured.

It's "value of money stickiness" rather than wage stickiness at that point.

We could go to the low cost of living world, in fact we will if Ron Paul has his way, but it'll be very very bad for anyone with outstanding debts (and FANTASTIC for all of the banks owning those debts, and HORRIFIC for the US Treasury and bond repayment, etc)

migmigmigmig wrote:
Mar 8th 2011 1:24 GMT

Oh and, ps, trust me, I'm in the software engineering industry:

If you offer your programmers porn and WoW, you don't really see their productivity increase much.

Doug Pascover wrote:
Mar 8th 2011 2:36 GMT

Lex, speaking of machines making other machines, over on cnn.com, I just saw the headline Android Passes Blackberry.

1-15 of 15

About Free exchange

In this blog, our correspondents consider the fluctuations in the world economy and the policies intended to produce more booms than busts.

Advertisement

Advertisement

Latest blog posts - All times are GMT
Link exchange
From Free exchange - March 7th, 22:16
Free books!
From Prospero - March 7th, 21:30
Reading material
From Prospero - March 7th, 20:09
Worth the trouble?
From Multimedia - March 7th, 19:57
More from our blogs »
Products & events
Stay informed today and every day

Subscribe to The Economist's free e-mail newsletters and alerts.


Subscribe to The Economist's latest article postings on Twitter


See a selection of The Economist's articles, events, topical videos and debates on Facebook.

Advertisement