In this article, the author criticizes the anti-Luddite argument — that increased productivity increases capital, which increases the number of new jobs and kinds of jobs — by saying that this process favors skilled workers more than unskilled workers. He uses this figure in support:
How do the Austrians explain the raise in skilled workers' wages accompanies by a drop in the unskilled workers' ones?
To me it seems natural that with progress of society, there is an increase in demand for more skilled people and decrease in demand for less skilled ones (just like there is increasing demand for flashlights and decreasing demand for oil lamps). One would be interested to see an accompanying graph that shows the change in the number of college-educated people and high school drop-outs.
there is an increase in demand for more skilled people and decrease in demand for less skilled ones
There is never a decrease in demand for less skilled or unskilled people, but this may depend on what one means by "skilled", though it seems more plausible that there are more low level/min wage jobs than skilled labor jobs.
That argument is stupid; it ignores changes in the composition of the groups in question. College graduates are a far larger percentage of the total population now than they were in 1970, and high school dropouts a much smaller portion of the population than they used to be. It's reasonable to assume that productivity and education are caused by a common factor or two (like intelligence, aptitude for hard work, etc; lets call it productive potential), rather than that productivity is caused by education. So, in 1970 the category of 'high school dropout' would have included more people with reasonable productive potential than it does now, as since it became easier over time to finish high school and college, the most productive people in the 'high school dropout' category would have moved into 'high school graduate', leaving only the people with the least productivity potential remaining in 'high school graduate', and similarly the most productive high school graduates would move into the category of college dropouts or graduates. To put it mathematically, whereas maybe in 1970 the category of 'high school dropout' represented the worst 20% of the population in terms of productivity potential, now it represents the worst 10% of the population, so the average productivity of the sample would be lower. (The least productive 11-20% of the population would still be more productive than the least productive 10%, making the average productivity of the lowest 20% higher than that of the lowest 10%. To help understand, think that the average wage/productivity of the least productive 80% of the population would obviously be much higher than that of the least productive 10%, as the least productive 80% of the population still includes many highly productive people, such as people in the top quartile of productivity potential). Hence, the average wage could easily have fallen solely due such changes in the composition of the samples studied.
MadMiser: I think that you missed the vertical axis, which indicates that the graph is composition-adjusted.
The author of The Atlantic article misses two very important points. The first of these is touched on but not explored. The authors comment on minimum wage laws, unemployment insurance, health benefits, prevailing wage laws, and long-term contracts. Add to these the myriad of other additional regulations that make it extremely more difficult to hire someone today than it was in 1963 and you’ll get some idea of why labor is being replaced by automation. A business owner doesn’t have to pay for a machine three years after he stops using it, as is currently the case with employees.
The second error involves their superstar argument. This argument might have been valid a decade ago but clearly is not today. The book The Long Tail clearly explains how recent technology has made it possible for small producers to connect with their small audiences.
Overall, this article fails to make the case for the ancient and obsolete narrative that innovation harms anyone except temporarily for some displaced workers. At every level, the deployment of more capital in a job results in higher incomes for the workers who use the capital.
Haha, you're right; oops! Well, even assuming the data is correct, the article mentions Samuelson, but completely ignores the theory to which his name's attached: the Stolper-Samuelson theorem. Considering said theory predicts and explains exactly what the author's describing, falling unskilled wages, in terms of international trade theory, the author really needs to make the case as to why his luddite theory is a better explanation than the Stolper-Samuelson theorem - something he doesn't even attempt to do. From the wikipedia article on the Stolper-Samuelson theorem: "Unskilled workers producing traded goods in a high-skill country will be worse off as international trade increases, because, relative to the world market in the good they produce, an unskilled first world production-line worker is a less abundant factor of production than capital." So if falling wages for unskilled workers can be explained entirely by increased international trade (an increase in the unskilled labour supply relative to capital), then he has no justification in saying such falling wages must have been caused by technological progress.