What is automation?

The Cambridge Dictionary defines automation as “the use of machines and computers that can operate without needing human control”. Other definitions, as the one by the International Society of Automation, go one step further and define automation as “the creation and application of technology to monitor and control the production and delivery of products and services”. This means that automation is not only about machines or computers, but is more of a process, which involves everybody who forms part in the creation of a certain type of technology. Techopedia acknowledges that automation involves minimal human intervention, but that it also creates new jobs and increases productivity. So, are we becoming redundant, as machines and computers increasingly conduct human tasks?

Tasks versus skills - and why the conceptual difference is important when talking about automation

Influential work by Autor, Levy and Murnane (ALM) from MIT revolutionized the way economists look at skill demand (Autor, Levy and Murnane (2003)). Earlier work defined skills as the average education level of workers in a given occupation. ALM (2003) broke up with this definition when introducing a distinction between skills and tasks. In their definition, a task is “a unit of work activity that produces output” while a skill is “a workers’ stock of capabilities for performing various task”. The authors applied this new distinction to study the effects of computerization on the demand for skills as well as the distribution of tasks between humans and machines. To study this effect, they differentiate between three different types of tasks:

  1. Routine tasks. A routine tasks is a task that is sufficiently well defined to be carried out by a machine or a worker in a far-away country. Additionally, it is marked by repetitive, precise and well-understood procedures. An example would be bookkeeping, or working at an assembly line. These tasks are mostly carried out by the middle-skilled.

  2. Abstract tasks. An abstract task is a task that requires problem-solving, intuition, persuasion as well as creativity. An example would be managerial or analytical tasks, as well as working in engineering or science. These tasks are mostly carried out by the highly-educated.

  3. Manual tasks. A manual task is a task characterized by situational adaptability, visual and language recognition, as well as in-person interactions. Cooking food, driving a vehicle or cleaning are tasks that fall below this category.

Technological change and different tasks - are we becoming redundant?

automation

Autor, Levy and Murnane (2003) start to look at the effect of technological change via focusing on computerization. They find that it is mainly the routine-tasks that are increasingly being replaced by computers. In addition to that, these tasks are increasingly conducted by workers living in foreign countries. Moreover, workers conducting abstract as well as manual tasks benefit from computerization through an increase in productivity. As an example, managers become more efficient as computers help them to process, organize and filter information. Autor, Levy and Murnane (2003) also find an increase in the share of labor conducting manual tasks. This could be evidence for labor being displaced in routine-tasks moving towards manual tasks. In 2013, Autor and Price conduct an update of the ALM analysis and find that one of the previous findings seem not to hold for more recent periods. There is a moderate decline between 2000 and 2006 in both non-routine analytical and non-routine interpersonal tasks. (Picture by © Maxim Hopman)

Why is there job as well as wage polarization in the US?

In 2013, Autor and Dorn (2013) document that there is job as well as wage polarization between 1980 and 2005 in the US. This means that there is job growth among the low-skilled as well as high-skilled, but not the middle-skilled. Similarly, real-wages have grown for the low-skilled as well as high-skilled, but not middle-skilled. Autor and Dorn (2013) explain the observed polarization through the different exposure of local labor markets to routine tasks. While those labor markets marked by a high share of routine tasks experienced an increase in the adoption of information technology, others did not. This led to a reallocation of low-skilled labor into service occupations in these labor markets as well as an inflow of high-skilled labor, finally resulting in wage as well as job polarization. They then conclude that the observed polarization is mainly driven by this reallocation towards service occupations.

Autor and Dorn (2013) define service occupations as occupations in which one cares or assist another person (e.g. food service workers, gardeners, cleaners, or health aids as well as hairdressers, among others). These occupations are typically low-skilled occupations. The share of workers in service occupations have grown by over 30 percent between 1980 and 2005. The authors explain this by consumer preferences as well as computerization being able to substitute certain low-skilled tasks but not in-person service tasks. The researchers confirm this through applying a more sophisticated version of the ALM methodology. As they see service occupations forming part of the “manual task” category, this is in line with the predictions of the ALM model. Additionally, they show that this is indeed driven by computerization and not by deindustrialization, offshoring, rising demand for home production or increased low-skilled immigration. Moreover, the increased adoption of computer technology goes hand in hand with an increase in high-skilled labor.

Now, are we becoming redundant due to increased automation?

Well, the answer to this is yes and no. Similar to what has been observed for the effects of robotics (see my previous blogpost) there are losers and winners to technological change. While in the case of robotics it is mostly the low-skilled manufacturing workers being replaced, when focusing on computerization it seems to be the low-skilled non-service workers which are most affected. Not surprisingly, the high-skilled benefit from computerization through productivity increases. Still, there are certain groups among the low-skilled, which grow due to an increase in computer technology. Those are the service occupations, which experience not just an increase in employment but also earnings.