Wednesday, February 6, 2013

I'm not feeling reassured here

Derek Thompson, one of my favorite journalists, has a post worth reading that pushes back against the job loss from automation-- but also read the comments, which make a lot of good rebuttal's to what Thompson is saying. Part of what I like best about Thompson is that he is a remarkably even-keeled guy, a rarity in a newsmedia that gravitates towards the euphoric and the apocalyptic. It's especially unusual in a guy who writes often about tech, and a guy with a generally neoliberal flavor, as they tend to be the ones in a rush to ignore human suffering. Optimism that doesn't degrade into absurd technoutopianism, or that doesn't discount the incredible hardships of the poor, is a rarity these days. I appreciate his sobriety, and I don't know, maybe he's right. But there seems to be a lot of wishful "hey, it's probably not gonna happen soon" going on there. Worth saying, too, is that Thompson focuses exclusively on employment rate, and doesn't consider job quality, which to my lights is equally important. As his colleague Matt O'Brien points out, automation doesn't have to lead to mass joblessness to severely undermine the conditions for workers.

Seriously, read the comments of Thompson's post. Tons of insight there.

One thing I'd like to point out is that even if we merely have a change in the skills or temperament needed for some jobs, rather than just the elimination of jobs, that essentially renders large swaths of people unemployable. Jobs that interface significantly with technology (which is to say, jobs) typically require significant retraining for people who have been laid off from other careers. People who can't adjust-- or, even worse, people who aren't given the opportunity to adjust because employers assume they can't-- are left behind, and fall into the precarious lives that so many are suffering through right now. As one of the commenters on the post points out, economists tend to think of jobs in net; if 5,000 people are laid off, and then 5,000 jobs added, hey, that's balance. But the original 5,000 people are very unlikely to be significantly represented within the 5,000 newly employed, and the longer a person is unemployed, the harder and harder it is for them to eventually find work.

Matt Yglesias continues a recent trend in writing a post on this issue that is, well, bizarre. His Slate blog has taken a turn for the odd in general lately. In the post-- which, in keeping with another recent trend of moving towards a positively Mickey Kaus-like brevity, runs all of two paragraphs-- simply says that technology that eliminates jobs has always been with us and that it's strange to talk about it now. Well, yeah-- it's a long term trend. But as is the case with global warming, the fact that the trend is old says nothing about whether it is accelerating or whether we are reaching an inflection point that will undermine our basic way of life. I'm not sure what Yglesias thought is saying here. The fact that the trend has been going on for a long time doesn't mean it isn't causing massive human suffering. But suffering hasn't really been Yglesias's beat lately.

There's plenty to read. If it isn't clear, I'm mostly interested in these issues because they seem like a useful frame to discuss the long-term trend of less and less income going to labor, and in broader strokes, the way in which our society has become a massive machine for generating wealth for those at the top. I find the "income inequality" conversation to be frustrating on a variety of levels, in large part because it is so well worn. I'm trying to address how we think our society is supposed to function, and what happens when one of the basic planks starts to degrade. Contrary to Thompson and Yglesias, I don't think it's too early to think things through, even if these problems don't start to affect us in mass in the near future. We have a habit of deciding problems are problems too late rather than too soon.

14 comments:

Brett said...

It's hard not to take the historical view, when you've read about previous waves of new technologies and industrialization. A lot of the exact same types of arguments showed up following the adoption of new technologies in the early twentieth century, late 19th century, and even early 19th century following successive waves of new labor-saving machinery being used in factories. In almost all of those cases, the first couple years tended to suck for people displaced (all the more so because social spending to ameliorate the pain was either non-existent, patchy, or weaker than now), but then things tended to shift, and society ended up better off in the end.

Everybody says, "This time it's different", too, whenever it comes up. People were saying that even before the late 1990s boomlet - it's just that they get more attention now because of the downturn.

The main difference is just that we now have better safety nets than the previous waves. If all else fails, workers who get displaced are often either older (and closer to retirement and social security, as is common in manufacturing where the average age is quite high) or younger (with more time to retrain).

I also think it's worth pointing out that if automation is supposed to be as good as it says it is, that could actually prove useful for the shrinking labor force down the line. All that artificial intelligence might let people do all manner of stuff that used to be only capable of being done by people with B.S. or even PhDs, meaning that a lot of the education expenses wither. That's something Paul Krugman predicted semi-seriously in an essay 17 years ago.

There's plenty to read. If it isn't clear, I'm mostly interested in these issues because they seem like a useful frame to discuss the long-term trend of less and less income going to labor

Is it actually a decline in the amount of income going to labor, or just a greater share of new income going to the richer folks?

Anonymous said...

Wait, I'm not sure if I understand your objection here Freddie:

As one of the commenters on the post points out, economists tend to think of jobs in net; if 5,000 people are laid off, and then 5,000 jobs added, hey, that's balance. But the original 5,000 people are very unlikely to be significantly represented within the 5,000 newly employed, and the longer a person is unemployed, the harder and harder it is for them to eventually find work.

Assuming we break even on net job creation/loss, do you actually think it matters that the exact new jobs that were created aren't directly filled by people who lost the exact old jobs that were lost? I mean, even assuming that all the people taking the new jobs were already employed, that's still 5000 new vacancies in whatever in their old jobs, which will themselves be filled by people (who in so doing are creating vacancies of their own, etc) and so on.

The math doesn't work unless at some point someone who was previously unemployed is now employed. That's what net means. Of course, you could be implying that the created vacancies are actually just lost/downsized away, but if that were the case then the economy wouldn't be showing net balance in terms of job creation, but a net loss.

So... what's wrong with net balance job creation, exactly?

Squarely Rooted said...

To be fair to Yglesias, he's made this point before -

"the tendency to want to explain the current period of prolonged labor market weakness—which really is new and unusual—in terms of the long-term trend toward the invention and deployment of labor-saving machines is strange."

He's saying it's strange to cite new technologies as the cause of the recession, not that it's strange to ever be concerned about their effects on anyone, ever.

Ethan Gach said...

I'll have to double check, but to Yglesias' point, I'm pretty sure wage stagnation started in the 1980s, around the same time that manufacturing went belly up in the face of foreign competition.

Manufacturing is a thriving industry now, but does not employ as many people per inflation adjusted dollar's worth of widgets produced.

Yglesias is aware of this of course, which is why he's always pushing "construction" as the panacea for all displaced blue collar labor.

Ethan Gach said...

On another note, why do you keep getting deleting posts Freddie?

Freddie said...

? The only post I deleted was one where I wrote a title and accidentally hit post. There was never any content in it.

tongorad said...

Freddie, on a related note, did you see William Black's pieces about Yglesias?

Yglesias pours the Geithner, Holder, Breuer (GHB) banksters immunity doctrine in our drinks

Yglesias mimics “Mankiw Morality” and bashes Bastiat

Some pretty fine writing that challenges Yglesias's competency in commenting on economic and ethical matters IMO.

Ethan Gach said...

I spoke too soon (now that the empiricism one is up).

Saw a couple ghost posts in my RRS feed and was disappointed. Mystery solved.

Brett said...

@Ethan Gach
I'll have to double check, but to Yglesias' point, I'm pretty sure wage stagnation started in the 1980s, around the same time that manufacturing went belly up in the face of foreign competition.

It actually started earlier, in the mid-1970s. There was a brief period in the late 1990s when real wages went up across the board (including at the very bottom), but now we're back to the old trend.

If you look at household income, though, that's gone up in real terms over that period. It even went up slightly in the past decade, with rapidly rising health care expenses devouring most of it.


Manufacturing is a thriving industry now, but does not employ as many people per inflation adjusted dollar's worth of widgets produced.


It seems to be going the same way that agriculture is in the US. Unless something really drastically screws with the economies-of-scale with most manufacturing, it'll eventually just be a highly productive sector with low employment and a slender fraction of GDP.


Yglesias is aware of this of course, which is why he's always pushing "construction" as the panacea for all displaced blue collar labor.


It's not just construction, although I wouldn't dismiss that (pay in construction isn't bad - I did it a few years back). It's why Yglesias tends to be big on examining stuff that makes housing too expensive in booming areas for poorer people, and also occupational licensing - both of them tend to weaken job growth for lower-skilled workers.

@SquarelyRooted
He's saying it's strange to cite new technologies as the cause of the recession, not that it's strange to ever be concerned about their effects on anyone, ever.

It's strange, but not surprising. There have been people saying that this wave of automation was going to be the one that kills off jobs for good for at least two decades now (Jeremy Rifkin was saying it in 1994), but they tend to get a bigger audience when economic and employment growth is weak.

@tongorad
Some pretty fine writing that challenges Yglesias's competency in commenting on economic and ethical matters IMO.

I thought it was weird to go off on a Mankiw-related tangent, based just off of Yglesias saying that if it was better for all of us to bail out the banks rather than let them go bankrupt, we ought to bail them out.

I mean, we basically do that already with FDIC insurance, just on a lower level.

matt said...

I took Yglesias to be making 2 perfectly good points.

1) Explicitly, that current unemployment isn't the result of technological progress.

2) Implicitly, that technological progress is as such a friend of the worker, to the extent that the increase in productivity can be kept out of the hands of capital.

Together, these imply that the suffering of unemployment is a political problem. I'm not sure what you would object to in this.

Steve D. said...

Historically it has been true that the implementation of labor-saving machinery hasn't caused long term widespread unemployment. I wonder if there are any good models out there for the dynamics of the situation or a way to predict what sectors the displaced workers would enter.

I'm guessing that we'll see a transition to more jobs providing leisure consumption. Just as fewer people are necessary to grow our food in comparison to the past, fewer people will be necessary to make things. This frees up people to become actors, artists, pure scientists, authors, etc.

As long as people value things done by people, people will have jobs.

If we do end up hitting the singularity and machines are better at every conceivable task, then that's a whole nother ballgame.

tongorad said...

Brett "I mean, we basically do that already with FDIC insurance, just on a lower level."

Wiki: The FDIC receives no Congressional appropriations – it is funded by premiums that banks and thrift institutions pay for deposit insurance coverage and from earnings on investments in U.S. Treasury securities.

Furthermore, deposit insurance for ordinary savers up to $250K doesn't seem like an apt comparison to the massive control fraud perpetrated by the banks, now does it?

shawn said...

I'm not so sure I'm willing to put my faith in what's historically been true. The industrial revolution replaced a lot of human labor with machines, but it was also accompanied by an absolutely huge increase in productivity. But a huge increase in productivity is really only useful if there's a demand for all that new productivity.

If a steam powered underwear making machine operated by one human could do the work that used to require 20 humans, that's not an employment problem if there's a demand for 20 times as much underwear as was being produced previously. Prior to the industrial revolution, there was an absolutely enormous pent-up demand for goods, because the average person didn't have much stuff. There just wasn't much stuff available, because making things took so much labor.

I don't think we have anywhere near the same situation today. Most people have more stuff than they know what to do with. With a few exceptions, the global production capacity for most goods and services already is more than enough to meet demand. Sure, there are many people in the world still living in poverty, but that's a result of social and political issues, not a result of a lack of stuff.

So we're potentially looking at significant numbers of employees becoming redundant across many existing industries, with very few new jobs being created in those industries, at any skill level.

So what about the less automation-ready industries, the creative fields? I don't think that's too promising. Our society already has way more artists/musicians/chefs/historians/philosophers/etc. than it's willing to financially support. There's a reason why there's so many jokes linking a liberal arts degree to a job flipping burgers. The idea that there's going to be a giant increase in demand for new creative work doesn't seem realistic to me, and even if there is, there's already a huge pile of educated people trying to find good work in those fields.

So that leaves new industries, stuff that doesn't even exist yet, maybe even fields that we can hardly imagine at this point. That's possible I guess, but I see no reason to expect new fields to be any less automated than what's happening to existing industries now. Arguably, it's realistic to expect them to be highly automated, and probably not even possible without significant automation. Asteroid mining might bring us huge amounts of resources sometime in the future, but I'm pretty sure the vast majority of the labor involved in it will be performed by machines.

I just don't see that many places where hundreds of millions of new good jobs will come from, and that's the kind of numbers we will need for our economy to continue to function on the same fundamentals that it's currently built on.

Anonymous said...

My father is a physicist and like me does not see a utopian future (all comes down to energy available to society in the end), but the physicists around him tend to be techno-cornocopians (technology/science will eventually solve all problems and we will live happily in a Jetsons future). They answer in a similar fashion to "what will people replaced by robots/computers do for income?" "they'll design things or do art or teach or all sorts of things" - ignoring a basic problem; does anybody really expect an assembly-line worker put out of work by a robot, or a para-legal put out of work by a legal precedent search-engine, actually to somehow just become programmers or robotics designers or television writers or successful artists?
This time DOES seem different as the breadth and rapidity of the change does not seem to allow for adjustment (Foxcon is already purchasing robots to replace pesky Chinese workers to build those nifty iPhones and iPads - what will they do now, go back to field work that has probably been replaced by machinery?). Technology will be very good at further concentrating wealth towards those who control capital, unless they eventually run out of enough people to buy the robot-made things. And that of course is the other side of this problem; as what were once middle-class jobs disappear, who will be left to buy the products available to keep the whole thing going?