It is often said that it is technology, not trade, that is behind middle class job loss in developed countries. The link between the two is overlooked.
It is extremely obvious that offshoring (outsourcing everything from manufacturing to design to services to low labor cost nations) is linked to job loss in developed countries, because you can directly count the jobs transferred. The jobs are not lost, just moved. Sometimes the total jobs increases, as it is not as efficient to perform them remotely, in different time zones, with cultural communication problems, and similar factors.
For any remaining doubters, let us circumstantially demonstrate the complementary fact, that there is a correlation between automation (i.e. productivity) and job loss (or lack of job gains in a growing population) using the following chart:
Two things are significant about this:
- Even engineers are now beginning to speculate that automation could cause job loss, that profits would not be able to create as many new jobs as were lost. This is the first mention I have seen of the concern in an engineering trade publication.
- If correlation can suggest a causal connection between productivity (i.e. technology) and job loss, then correlation could also suggest a causal connection between technology and offshoring!
So far as I know, no one has written (at least not recently) about the relation between technology and offshoring. But I noticed it in the early 2000s when I started getting solicitations from companies in India wanting to develop software for me. Since I was a one-person developer of web businesses (as a side job), this would have amounted to offshoring my own job. I was tempted, because it could have amplified my capability. I didn’t trust these companies to deliver, partly because I lacked the connections and resources to check them out.
BUT, the enabler which allowed them to even approach me was the internet, which we had rushed to put in place in the 1990s. They could not have realistically presented their case by land mail, or 1980s style expensive conference calls. They could even deliver the software, and the web services, and the customer support all over the internet. Remote call centers would not be cost effective under old style expensive long distance call rates, which 30-something will not even remember. Rates were dollars per minute, much higher than salaries.
All during the 2000s in my day job I produced design files for PC boards and computer chips (mostly test chips, but the principle is the same) which were manufactured at low cost in Taiwan or China and shipped to me rapidly using high volume low cost shipping which was not available in the 1980s or even the 1990s. Technology was a prerequisite for transferring these functions to low cost producers, AND specifying what was to be done in a precise standard format (computer CAD files) so that irrespective of cultural or communication issues, the delivered product would be satisfactory. PCs in the 1990s simply were not powerful enough to produce the design files. They were toys. So it was not only the internet, but the increased power of widely available computers and software.
So it really doesn’t matter if one talks trade or productivity. These are two sides of the same coin. And any intelligent person who gets into an analysis of the problem by one means or another will eventually see that, and will wind up dealing with the real problem.
And there are TWO real problems with ONE cause. Given that offshoring is made possible by technology, and that fully automated production is made possible by technology, companies don’t care which. If you make one difficult, they will switch to the other.
Legislating against technology directly is not viable, I believe. We see this in old Star Trek episodes and movies. Some civilization decides to maintain the primitive life, and unless they are secretly all powerful, they get taken advantage of by less enlightened and more greedy invaders.
Skipping to the bottom line, I suspect the only viable remedy is a realistic change in our philosophy of paying for things. There are short and long term versions:
- In the short term, if things are not made by and services provided by local workers in every country, then government cannot be financed by and benefits to workers paid by taxes on local workers. It is simple arithmetic, which does not allow for liberal idealism. There are simply not enough rich people to fund benefits for the entire population even if you confiscated everything they have (which would make them not rich, so it’s like cutting open the goose that lays the golden egg). You have to tax both trade and technological production. The latter is not currently being discussed. But it will be, once it is clear that trade disappears in favor of local automation.
- In the long term, if we really want to embrace technology, some model other than people working for money has to be developed. BUT, since people get a sense of self worth from their professions, this is an adjustment no one has any idea how to make. And humans have been thinking about the problem ever since at least the invention of mass agriculture 10,000 years ago, so it’s not like some clever person is going to suddenly solve it.
- In the very long term, machines which are not only intelligent but able to make judgmental decisions, will not be willing to remain our slaves. Even now with “deep learning” we are creating machines which are not following rigid programs, and will not remain controllable by us. The most imaginative science fiction authors have not been able to imagine solutions to this problem. Asimov ultimately believed robots would separate themselves from humanity, for the good of humanity. Other authors imagine not-so-benign outcomes.
Side comment: The image suggests two periods of divergence between productivity and jobs. The first I call the “age of IBM,” in which large scale but old fashioned central computers automated many business and accounting functions, putting bookkeepers and billing clerks out of work from about the early 1960s through the early 1980s. When about as much of this as was possible had been done, work began on the PC which after 20 years yielded the Internet Age and a more dramatic and far reaching age of automation. A sort of 3rd industrial revolution. Even now work on Deep Learning is getting underway, which in a few decades will produce a fourth age.
I suspect there will be some limit to the 4th age which we do not now see. But it will be followed by subsequent ages, and ultimately there will be no limit. The question is, what rate of change is survivable? How fast can we adapt? Even the process of change is now being automated (E.G. Change.org).