Disguised Unemployment
Rapid advances in AI has raised alarms about possibilities of job losses and increase in unemployment. Such concerns have been raised every time a new technology arrives, but somehow humans have been able to create the jobs to keep everyone employed. The obvious question is how were these jobs created and can we replicate it this time as well. I argue that technology has already reduced the number of people required to run the world is below the existing population and substantial part of the jobs in the current world are nothing but disguised unemployment.
Jobs can be broadly categorized in 4categories
- jobs which are essential for running the world (operational jobs)
- jobs which are for the advancement of technology/science (research jobs) and
- jobs which are for human enrichment (arts/culture jobs)
- support jobs to manage/ administer first 3 categories (admin jobs)
The number of jobs in the first category reduce in number with technological advancement. From the days of hunter gatherers where entire population for needed for the food production, we have come to a stage where less than 1% of population in developed countries is employed in agriculture. But human needs evolved which required producing more goods/services which created additional employment in this category. This has been the reason for growth of economy and jobs, even though it results in wastage, as I had pointed in my last blog post. While the needs of the humans can increase over time. exponential technological growth means that this category of jobs have declined over time. In this category, if you have input x then you definitely get output y. y=x*multiplier. This multiplier would have increased over time.
The second category, research jobs, are considered pretty common in the technology companies. However, I would like to expand the definition of this category. These are essentially jobs where the probability of success is not 100%. In operational jobs, when one completes a set of actions, the output is guaranteed. But with research job, the probability of success is often low, though the output can be high. This is something like Private Equity players who invest in multiple companies hoping that at least few of them can be multi-baggers. By this definition, we can expand this category to employees which are working to ‘grow’ the company as well as to the professors of the universities which are involved in scientific research (I want to highlight scientific here to distinguish it from other research). In this category, the output y for input x is not deterministic. There is a probability factor, but the multiplier in this case can be much higher than first category. So y = x* probability*multiplier. There is net positive impact and this category takes humanity forward.
Third category is purely discretionary. Humans crave for life beyond mundane needs. They like to indulge in art, music, culture, sports — none of which is important for survival. People are willing to pay huge sums for these jobs — actor, sportsstars, perhaps as they meet their emotional needs. So then this be categorized in the first one. Possibly. But I would like this category for the jobs which are important for enrichment of human culture, but people are not necessarily willing to pay for them. Consider examples such as art cinema/theatre, history/social studies researchers in the universities. This is interesting category because the output is not tangible and jobs can theoretically be completely eliminated.
The fourth category is essentially to manage other people or to manage complexities introduced in the human life over the years. There can be genuine need for them or some are created due to unnecessary complexities or due to regulations/latest fads. This category is supposed to increase the output of the other categories.
Distribution of jobs in these 4 categories must have changed over time. When humans were hunter gatherers or had just started farming, each had job in the first category. When they started generating surplus, jobs from other categories started arising. But are these other categories jobs essential jobs? Can they be eliminated? Are they created because humans have too much surplus and it doesn’t matter if we waste part of it by giving jobs? If that is the case, these jobs are not really required and they are essentially disguised unemployment.
Let us first consider third category. There are multiple research jobs where people are researching history, literature or in short liberal arts. These are important for the society and I am all for retaining them (or I would love to work in one of such jobs). But they are essentially at the mercy of the funding authorities. Government/ private entities fund them through surplus (which is why there are more such jobs in developed countries where the governments and corporates have higher surplus). They are equivalent of the poets/ artistes that kings used to keep in their court in the past. They are important for the society, but society may not fund them. So these jobs are effectively ‘created’ jobs. Government rather than giving minimum basic income is saying that I will pay you for doing this research which can be useful for our society.
Fourth category is interesting because complexity creates jobs. Take an example of the US healthcare system. Because of the complexities involved, it requires multiple people to take care of insurance, billing, etc. So while people complain about healthcare in the US, it creates jobs. Efficiency is bad for creating jobs. Redundancy and wastage creates jobs. That’s why large corporate structures with multiple managers are good for creating jobs, though they may not exactly be required.
Second category is the trickiest. With automation and technology, the number of people required for operational work in corporates has gone down. But companies have surplus and want to grow. So they essentially hire more people than required in a hope that these employees will drive growth. Even though the growth may not materialize, during the interim period it provides employment. Apple recently laid off people working on its EV car team. While the project was scrapped, it did provide employment for __ people for __ years. Do tech companies or consumer goods companies really need thousands of people? I would argue that only part of it are required to keep the company running. The balance are hired in a hope that they may do a breakthrough which will propel sales and drive growth.
So even with AI, there could be multiple roles generated- from prompt engineers to AI content generators. But I think the important impact of AI will be that it will create sizable surplus to sustain more jobs in second and third categories. So yes, people will be employed, just that their jobs will not be contributing to creating the surplus, but rather consuming it!