Technological Unemployment - Summary of a summary

There is this approach to consuming information online. I don’t know what it is called or who came up with it. The approach is to look at a topic, write down the solid questions you would like to answer about the topic, consume the material about this topic that’s in front of you and after you are done, answer the questions that you had started out with. With several blogs (esp Slate Star Codex), this approach has helped me a lot. Especially because it gives me an idea about what I would like to know, what my presumptions / beliefs about the questions are before I read the article and am influenced by the author’s presumptions and beliefs.

Technological Unemployment is a post on Slate Star Codex. I have been following this blog, albeit without understanding much on several occasions, for the past month and I jumped right out of my seat when I realized what this post was about. I had recently had a discussion about this with my parents in the context of the (lately increasing) AI paranoia articles in mainstream online and offline media and on sources like LinkedIn where people seem to be posting stuff about this all the time. (Elon Musk’s rather arbitrary position on this doesn’t help)

Before I read the post, I wrote down some questions and after reading the post (and taking notes, yeah.), I answered them.


Prime Age Male Labor Force Participation Rate went from 97% in 1950 to 88% in \2014. Where did the ~10% of men go? Why aren’t they part of the workforce anymore? Were advances in technology the cause for their exit?

What sectors and income brackets were these 10% men a part of?


  1. This is third-hand information. Scott Alexander (the blog post’s author) read several papers and reports about the topic and composed his post. I have read his post and skimmed through the text and analyzed the graphs in the Brookings Institute Report.

  2. My questions will not be the same as the questions you would like to answer depending on your age, field of employment, childhood etc.

  3. The studies concentrate on the labor market in the US. So, it’s inherent in the blog post that everything is in the context of the American labor market. There will be wild differences between these numbers and the numbers for India or China, but the comparisons between the US and countries like Japan, Germany, Spain and other EU countries seems to indicate that the behaviour in these countries is more or less similar and the US labor market may be used as a proxy for these countries’ labor market.

  4. The numbers in the tables have been taken from graphs that are mentioned below the tables. The numbers are representative and as accurate as can possibly be taken from a graph which doesn’t have labels.

That said, here goes.

Goal: Are people losing jobs because of advances in technology?

What age-group, level of education and gender dominates this set?

Throughout the analysis, we concentrate on Prime Age Male Labor Force Participation Rates. i.e Men between the ages of 25 and 55 who have a job. Female LFPR are not considered here because the main jump for Female LFPR came around 1990-2000 and it seems rather premature to make anything of numbers that are only 20 years old.

Age Group - All age groups; But mainly advanced ages; where disability can be a major cause for retirement before the age of 55 or a change of role at home (stay-at-home dad); even apart from this, there seems to be some indication that younger people have a degree of robustness in the selection of their field of work and an option to go back to college and gain skills as compared to older people.

Level of education - People whose level of education is less than high-school are the worst affected. Men with college degrees show a rate of decline that has remained more-or-less unchanged since 1950. The following table shows the Prime Age Male Labor Force Participation Rate distribution for three years across level of education:

Level of education 1968 1988 2016
4 years of college and more 97% 95% 94%
Some college but no 4-year degree 96% 95% 87%
High school degree, no colege 95% 94% 84%
Less than high school 94% 85% 80%

Figure 4, Brookings Institute Report

Gender: Inconlusive

What kind of technology lead to the loss of jobs?

It’s better to rephrase this question as “Which sector lost the most amount of jobs because of advances in technology?”

Manufacturing and middle-skill level, middle-wage jobs were the worst affected. The conclusion on this is unanimous and there are no doubts about who was dealt the worst hand here.

Manufacturing jobs as percent of total:

1948 - 30% 2010 - 10%

Manufacturing jobs (absolute numbers):

Year Millions of jobs held in this sector
1970 18
1980 19.4
1990 17.9
2000 17.1
2010 11.5
2016 12.1

Figure 6, Brookings Institute Report

It’s worth noting that the share of manufacturing the in the GDP has remained fairly unchanged over this period of time.

Manufacturing has been losing jobs ever since World War 2. During the Industrial Revolution, there was an exodus of people from agriculture to manufacturing. Now, if there is to be an exodus, where would the people with the current manufacturing jobs go?

Surprisingly, it’s not the least-skilled jobs which lost out. In fact, they were the ones that gained because it’s fairly hard to automate these jobs away. These jobs are manually intensive or require some level of interpersonal communication and robustness (say against vandals or robberies in the case of truck drivers) which are not easy to solve in the real world.

A quote from the article that sums it up:

This is potentially consistent with a story where the jobs that have been easiest to automate are middle-class-ish. Some jobs require extremely basic human talents that machines can’t yet match – like a delivery person’s ability to climb stairs. Others require extremely arcane human talents likewise beyond machine abilities – like a scientist discovering new theories of physics. The stuff in between – proofreading, translating, records-keeping, metalworking, truck driving, welding – is more in danger.

There is one major caveat here though: As the number of jobs has always been declining, it’s hard to say without a shred of doubt that the jobs that were lost were indeed lost because of automation and not because people had a change of mind and decided to pursue other jobs or manufacturing jobs became less desired or some reason not related to automation at all.

Jobs were lost mainly in the manufacturing sector and in sectors that require medium skill level and are middle-wage

Was this technology cheaper over a period of time, compared to hiring people?

I couldn’t answer this question. Common sense seems to suggest that this should be the case or else why would businesses do something that costs more and employs fewer people? A quote from the blog post hints towards the same:

There may be some point at which we too stop being worth more than it costs to replace us. And the decline of manufacturing, the increase in labor force nonparticipation and despair in rural Rust Belt communities, etc, suggest that point is fast arriving.

This comment takes a stab at why even if one thing became cheaper, the inventors would be forced to move on to some other sector and if that continues, everything will become really really cheap but everyone will also be out of a job. Where did I end up?

Anyway, Couldn’t answer.

Did the jobs which were lost involve a hazardous work environment?

Jobs like those on the 24 hour assembly line that cause a lot of employee discomfort; toxic factories like the ones for fireworks, textile; works inside coal mines. In general, these are environments that adversely affect the employee just from working there. Dry cleaning is another example.

There’s no particular indication here that jobs in hazardous environment were more automated compared to non-hazardous environments. The Brookings Institute report does provide a good out for us though.

Table 1 reports the “Most Rapidly Declining Industries - by employment projections” as Apparel, Leather and allied manufacturing; Tobacco manufacturing; Postal service; Federal enterprises except Postal services; Manufacturing and reproducing magnetic and optical media. The first two sectors in this list are hazardous when the proper employee safety norms aren’t folloed, but that probably won’t be the case in the developed economies.

So, Not very conclusive; But some of the jobs which were lost did involve hazardous environments

Were the jobs that were lost monotonous?

The answer to this is a resounding yes. Just from the definition of middle-skill level, middle-wage jobs which have been the worst affected, this should be apparent:

Middle-paying occupations are stationary plant and related operators; metal, machinery and related trade work; drivers and mobile plant operators; office clerks; precision, handicraft, craft printing, and related trade workers; extraction and building trades workers; customer service clerks; machine operators and assemblers; and other craft and related trade workers

(The occupations in bold are monotonous, in my opinion. I am probably wrong, so please make your own judgement)

So, A resounding yes

Did the people who lost their job find another job?

If yes, was the new job higher or lower paying than the old one?

Once again, we are stuck without any data. But we have some strong indications about the trend here. The consensus seems to be that most of the workers who were in middle-skill, middle-income jobs who lost their jobs, slid into low-skill, low-wage jobs, because they couldn’t complete the requirements and acquire the skills to move to the high-skill, higher paying jobs.

This being the norm, there are certainly some people who do gain these skills and move into higher paying jobs.

Strong indications that most people who lost middle-income jobs, slipped into low-wage jobs. While the others either exit the workforce or gained the skills required to move to a high-skill, high-wage job

I conclude exactly as Scott Alexander does at the end of his post. I don’t think the people who aren’t part of the workforce anymore are not here because of the advancements in technology.

Technology has always been progressing forward and there doesn’t seem to be anything historical happening right now (and there aren’t any indications that something of the sort will happen in the next few decades)

Perhaps the rosy picture painted by more people leaving the workforce early to retire, take care of family at home, or because of disability is one worth considering; even if it seems too good to be true and can’t be accepted whole heartedly.

That’s the end of my third hand information dissemination effort. As mentioned earlier, definitely read the SlateStarCodex post. If you have the time, then read the Brookings Insitute report. It presents the data they have very well.