Will the Fourth Industrial Revolution lead to large-scale unemployment?
When the impact of modern technologies like artificial intelligence, machine learning and robotics are discussed in a business or social context the conversation inevitably turns to the impact on employment. There are justifiable concerns that technology will lead to widespread unemployment or, at the very least, a workforce not ready for the types of opportunities provided. When technology changes the structure of economic activities some employment opportunities will be lost. This raises the question: Will this only be those jobs that are repetitive and rather simplistic, or will it have a zero net effect on formal employment?
If anything, the Fourth Industrial Revolution’s significant impact on policy and strategy is intensifying the level of conversation that, up to now, has mostly been based on anecdotal evidence or historical reference. However, there are always lessons to learn from the past. In addition, we need explorative technological forecasting based on extrapolations from the past to make sense of the future.
There are justifiable concerns that technology will lead to widespread unemployment or, at the very least, a workforce not ready for the types of opportunities provided.
Therefore, this study looked at the impact of technological innovation on employment over the long term to obtain a better picture of what the future might hold. The technological unemployment debate has split the academic world into technology pessimists and technology optimists. The pessimistic view is that innovation destroys jobs; it creates structural changes in the economy that drive unemployment in the unskilled labour market, while increasing employment in the skilled market. The optimistic view is that innovation creates a myriad of possibilities in the form of new employment, and also positively influences multiple support industries and sectors that already exist, with a net positive effect.
Technological innovation and unemployment
The media and management literature warn about the perils of technological unemployment. To move beyond the sensation-seeking statistics, it is necessary to obtain a better understanding of the types of jobs gained and lost as a result of technological innovation.
Four aspects are important when looking at the effect of technological innovation on employment. Firstly, any study about economic impact could be done at product/process level, firm level and industry level. It is of course also possible to extrapolate industry level to country or larger geographical levels. Secondly, we need to take cognisance of the multiple dimensions of innovation. Thirdly, we need a long-term lens to look at high-level correlations between technological developments and economic cycles. Finally, the types of unemployment need to be analysed to try and isolate technology induced unemployment from other large employment trends that could influence the data.
The type of unemployment depends on where the economic system finds itself on the economic supercycle, also called the K-wave pattern or Kondratieff long wave.
The literature indicates four main types of unemployment:
- Structural unemployment: This is defined as the disparity between the jobs available in the market and the skills of the workforce.
- Frictional unemployment: This is defined as the short-term unemployment experienced while people are looking for jobs.
- Cyclical unemployment: This is defined as loss of work due to economic downturns.
- Seasonal unemployment: This refers to unemployment as a result of seasonal productive activities – like those typically found in the agricultural industry.
According to early 20th century Austrian political economist Joseph Schumpeter, all types of unemployment could be ascribed to the creative destruction process. He explained that unemployment is mainly frictional; unemployment as a result of technology adoption was temporary and of a cyclical nature. In contrast, the Italian economist Pasinetti believed that structural changes in the economic system generated a technology-induced loss of employment. A global analysis of the 2008 recession indicated that job polarisation – referring to a structural change from low-skilled or unskilled labour to skilled labour – was typically concentrated during recessions which coincided with technological changes.
One of the mostly widely used theories to investigate the interdependency between technological transformation and economic activity is the Kondratieff waves introduced by Russian economist Nikolai Kondratieff in his 1925 book The major economic cycles. The type of unemployment depends on where the economic system finds itself on the economic supercycle, also called the K-wave pattern or Kondratieff long wave.
During the initiation of the upswing (recovery phase), the type of unemployment should be deemed structural as there was a mismatch between the skills required and the proficiency of the labour force. When the economic system entered the prosperity phase, unemployment was reduced, leaving predominantly frictional and voluntary unemployment. As the prosperity phase progressed, the type of unemployment became technological owing to the development of process improvement innovations. At the peak of the wave, the technological innovations reached their full maturity and, in conjunction with the start of the development of new inventions, initiated the recession. In the recession phase and throughout the depression phase, the main type of unemployment is cyclical.
Technologically induced loss of employment usually arose in the later stages of recessions. Fears of such losses stemmed from the labour-saving goal of most technological innovations and were typically sparked in periods characterised by radical technological innovation.
Both pessimists and optimists agree that short-term unemployment follows in the wake of technological innovation. However, the debate about the long-term impact of technological innovation on employment continues.
Empirical studies have been conducted to determine the effects of technological innovation on unemployment. Unfortunately, most studies were undertaken at individual firm level and do not take into consideration the creation of new industries and markets brought about by radical technological innovation.
Both pessimists and optimists agree that short-term unemployment follows in the wake of technological innovation. However, the debate about the long-term impact of technological innovation on employment continues.
Taking a closer look at technological innovation
Technological innovations have an impact on society, the economy and unemployment. The innovation continuum ranges from radical to incremental. Radical technological innovation typically creates new industries and markets while incremental technological innovation improves what already exists and is usually labour-saving in nature. Radical or revolutionary technological innovation has a far more significant impact on economic activity, and therefore employment, than its incremental counterpart.
Radical or revolutionary technological innovation has a far more significant impact on economic activity, and therefore employment, than its incremental counterpart.
We also need to distinguish between product innovation and process innovation. Process innovation improves the efficiency of the production process, or processes supporting the production process. Product innovation improves existing products or creates new products for the market. In general, process innovation is deemed a driver of unemployment because it can replace human workers in the course of optimisation practices. Product innovation, on the other hand, leads to employment growth owing to a growth in the market, at the firm level.
Most studies on technological unemployment have used a microeconomic framework to determine its impact on firm level. The firm-level research and development expenditure, which serves as a proxy for technological innovation, can easily be related to a firm’s employment trends. However, a firm-level analysis limits the ability to determine the net effect of radical technological innovation on, for example, another firm in another industry, since it affects multiple levels of the economy.
Seeing the bigger picture
Part of the reason why the impact of technological innovation on employment is still unclear is the fragmentation of the studies that have been undertaken. These studies cover different geographies, levels of investigation (firm, industry, sector, or country) and time frames.
In general, process innovation is deemed a driver of unemployment because it can replace human workers in the course of optimisation practices. Product innovation, on the other hand, leads to employment growth owing to a growth in the market, at the firm level.
Very few studies are performed on a macroeconomic level, partly because it is challenging to find a suitable proxy for technological innovation and to control for co-deterministic macroeconomic factors.
This is where a systematic literature review, as an academic research method, is well suited because it can integrate the results from a large number of empirical studies. This study used a longitudinal dataset – 213 data sets from 24 primary studies – to analyse the effect of technological innovation on employment. The analysis is thus done on a larger dataset gathered by multiple researchers, across various firms and industries, and in different sections of the K-wave.
While the meta-analysis did not render significant results at a macro level, analyses at firm, process, and product levels delivered significant effect sizes with interesting results.
What did the study find?
Technological innovation creates employment at firm level. This study supports a positive correlation between technological innovation and employment, but only at firm level. It provides robust scientific evidence to counter some of the negative narratives about technological unemployment. The data required to explore the question on a macro level, and over a significantly longer period, is unfortunately still lacking.
Technological innovation creates employment at firm level.
It was also found that product innovation has a generally higher positive impact on employment than process innovation. Both product and process technology innovation lead to increased employment at firm level. Despite process innovation having a small effect size, this counters the prevailing belief that technological innovation destroys jobs.
Interestingly, the negative impact of process innovation prevalent in the current discourse is not supported by the data. Arguments that this could be unique to the Fourth Industrial Revolution due to automation do not hold true. Previous large-scale technological innovations have also led to process automation, just using different technological innovations.
At a firm level, product-based technological innovation enables not only additional employment capabilities but also strategic market benefits. Having the ability to keep up, or even outpace competitors, allows for better overall performance.
It is important to invest in the training and education of employees in order for the firm to leverage new product development without having to employ external people.
It is imperative to manage the transition of the workforce. It is important to invest in the training and education of employees in order for the firm to leverage new product development without having to employ external people. It is recommended that policymakers in countries where unemployment is a concern should consider developing supportive initiatives that could help firms to expand or establish long-term product and process innovation abilities. This includes initiatives to facilitate the transition of the workforce – from an educational and skills development perspective – to coincide with shifts in technology. The combination of these two initiatives would require the collective effort of policymakers, major industry players, and academia.
Overall, the key take-outs from this study are:
- Technology does not destroy jobs; it shifts employment opportunities.
- Technological development leads to various types of innovation during macro-economic cycles.
- Process innovation does NOT lead to job losses as generally believed. In fact, the data shows it small positive effect size on job creation.
- Unemployment is also the result of the failure to upskill people for new types of employment opportunities.
Will the robots take over your job? Not likely. But you will need to acquire new skills to stay in demand in a digital world.
- This article is based on the MBA research assignment of Bertus Buys, with Prof Martin Butler as his research supervisor. The title of the research assignment is The robot ate my job: A longitudinal literature review of the impact of technology on employment in the last 200 years.
- Prof Butler lectures in Digital Enterprise Management and Technology Futures at USB. He is also head of Teaching and Learning at USB.