Do robots reduce wages?

Will Merritt

Since the 1950s, when the first industrial robots were used in manufacturing processes, robotics and automation in the workplace has continued to increase across almost every industry sector. Such advancements have changed, and continue to change, the employment landscape, improving efficiency, safety, productivity and quality. However, while such advancements have beneficially impacted global economic growth, allowing for cost reduction, there have also been potentially detrimental ramifications stimulated by the shift to a robotics based economy. Such consequences have come in the form of unemployment, and wage reductions, especially for some lower socio-economic employment groups.

Keynes, as far back as 1931, acknowledged the seemingly looming threat of technological advancements, predicting a model of unemployment resulting from an excessive supply of labour in proportion to its demand. Keynes’ technological concerns essentially stressed automated labour as ‘outrunning the pace as which we can find new uses for labour’ (Keynes, 1931). In certain circumstances, Keynes' prediction has materialised: as an estimated 3.3 million jobs in the United States’ manufacturing industry has been replaced by the invention of one robot between 1980-2016 (Acemoglu, 2020). However, it is important to distinguish that it is smaller towns - heavily reliant on low-skilled employment in large factories – which are forced to face the brunt impact of job displacement as a result of automation. Thus, creating a surplus of unskilled workers, which instigates a decrease in price, or in this case, wage rates. Unskilled workers are forced to compete against each other when robots can simply be substituted to provide consistent and accurate performance for increasingly low initial investment and running costs.

If these workers are unable to retrain in new skills, as Keynes makes comment of, which would provide access to ‘better’ jobs which cannot be fulfilled by robots, they risk having their wages cut even suffering unemployment. However, the reduction of wages at the hands of robots can appear in more subtle ways. Even when hard automation (when technology fully substitutes for labour) is not applicable, job roles may be subject to soft automation or task displacement (certain aspects of the job are automated). In this case the worker retains the position, but their value is minimised and consequently wages are driven down. Platforms such as Uber and other GPS dominated industries (taxi services) are such professions where the effects of the advent can be demonstrated. The knowledge of the streets of London by cab drivers has been largely devalued by automation – resulting in a 10% reduction in wages, not fewer drivers (Frey, 2023). The profession became accessible to more people and excess supply led to a price decrease.

In an attempt to protect the earnings level of the low-paid workforce at risk, governments could introduce a National Minimum Wage. Nevertheless, such an approach does not solve the issue as unskilled workers, who are at risk of hard automation, would still need to be paid at an increased rate, driving employers to invest in robots, decreasing demand for unskilled workers and decreasing wages (Lordan, 2019). A Universal Basic Income (UBI) (Kelly, 2021) is one idea which has been suggested, and is currently being trialled in Finland, in order to mitigate income inequality and also protect those workers unable to progress in a technologically advancing world.

Furthermore, for those workers successful in retaining their automatable job at their existing hourly rate, there remains a risk of wage disinflation (restricted wage growth). A recent study reported an inverse relationship between robot density and growth of hourly rate, predicting that ‘one additional robot per thousand workers will reduce wage growth by as much as 0.63%’ (Pilcher, 2018). An example of the reality of wage disinflation was highlighted in an article by Barclays Bank, in relation to average wages of US lorry drivers which has remained at a level of around $38,000 between 1980 to 2018. According to the report, soft automation is the causal factor of the stagnation, as technology such as ‘power steering, cruise control and automatic braking’ has made lorry-driving a less skilled profession which in turn limits the wages commanded (Barclays, 2018).

Having determined that the general category of workers, most at risk of reduced wages, are low skilled, low paid workers whose jobs can be substituted by robots; research has identified further socio-economic inferences as to the specific groups most vulnerable within the category. The Lordan and Newmark analysis in 2018 highlighted an increased impact of job automation on workers who were above 40 years old, those with low levels of education, and ethnic minority groups. Whilst researchers at LSE and King’s College London, found that utilisation of robots in the workplace resulted in a “sizeable” increase in the gender pay gap in Europe (Oppenheim, 2020). Their analysis found that a 10% increase in the number of robots used by a company, resulted in a 1.8% rise in the pay gap between male and female workers. This is compounded by the fact that men still occupy a greater percentage of higher level positions, therefore women are at greater risk of reduced wages.

There are, however, circumstances when the incorporation of robot technology and automation into the workplace can present workers with new opportunities and ultimately lead to an increase in wages. Leading robot manufacturer, Aethon, claims that robots can actually “boost wages'' (ST Engineering Aethon Inc, n.d.). In a “Smart Warehouse” environment, soft or complementary automation enables robots to be used in a stock retrieval function freeing up workers to undertake the more ‘valuable’, quality assurance based, final order checking and packing tasks, which in turn elevate the skill level of the job and commands higher rates of pay for the worker.

Furthermore, historically, new technologies have actually created jobs which did not previously exist. At the start of the 20th Century over 40% of the US population was employed in agriculture, over 100 years later, this figure is reduced to below 2% - primarily due to the use of increasingly effective machinery in farming – yet unemployment levels are relatively low; possibly suggesting that new and different jobs have been created to occupy employment demand (Rajadhyaksha, 2018). This is further exemplified in the invention of the motor car.  Short term, the invention was the cause of unemployment for those skilled in horse-related jobs who would have endured a loss or reduction in earnings. However, from the creation of the motor car blossomed a new automotive industry, necessitating new skilled roles and providing better paid jobs; ultimately contributing trillions of pounds annually to the current global economy. The invention, and subsequent industry, around motorised vehicles helps to possibly challenge the notion that there is a fixed amount of work available in the economy to be shared among workers, playing into what economists acknowledge as the ‘lump of jobs fallacy’. Instead, new opportunities, and the wages attached to them, are to introduced alongside technological advancements rather than limit them.

In addition to the creation of new industries and related jobs, robots bring with them benefits including increased productivity, cost efficiency and quality assurance; improving the overall performance of a company. When automation reduces the cost of producing goods, and in turn the price of those goods to the customer, economic theory would state that demand for those goods would typically increase and consequently demand for the non- automated workers, and the wages paid to them, would also increase (Costa, 2022). If robotics positively impacts company profits, such financial gains may be shared throughout the company in terms of wage increases and most certainly in performance related pay rates and bonuses. It is, however, evident that those benefiting most significantly and frequently in this scenario would be those at middle and senior level management.

In summary, despite the general offset of displaced jobs by the creation of new jobs (Briggs, 2023), the consensus of current economic analysis is that the effect of robots reducing wages is largely restricted to the socio economic group of workers who are less educated and who perform unskilled work which is most easily substituted with automated technology. To date, this has been most prevalent in manufacturing industries, however recent advances in technology, and specifically in Artificial Intelligence, has led to a belief that automation will become integral across all sectors, from retail to healthcare, and finance to journalism. Headline hitting predictions for the future effects of AI such as the exposure of 300 million full-time jobs to automation (Briggs, 2023) with 1.7 million of those in the finance and insurance sector (Rajadhyaksha, 2018) could appear to forecast an unprecedented level of unemployment. Upon closer examination, however, the anticipated global productivity boom which could accompany such a seismic change in working practices could result in an upward shift in wages following a demand-driven improvement in education and training worldwide.

References:

Acemoglu, D. a. R. P., 2020. Robots and Jobs: Evidence from US Labor Markets, s.l.: National Bureau of Economic Research.

Barclays, 2018. Delayed expectations: Automation, productivity and wages. [Online]

Available at: https://www.cib.barclays/our-insights/robots-at-the-gate/delayed-expectations-automation- productivity-and-wages.html

Briggs, J. a. K. D., 2023. Goldman Sachs: 'The Potentialy Large Effects of Artificial Intelligence on Economic Growth'. [Online]

Available at: https://www.ansa.it/documents/1680080409454_ert.pdf

Costa, R. a. Y. Y., 2022. Adopt, adapt, improve: AS brief look at the interplay between labour markets and technological change in the UK. [Online]

Available at: https://economy2030.resolutionfoundation.org/wp-content/uploads/2022/11/Adopt-adapt-and- improve.pdf

Frey, C. B., 2023. BBC News Interview - AI could replace equivalent of 300 million jobs - report [Interview] (28 March 2023).

Kelly, J., 2021. Forbes: 'Artificial Intelligence has caused a 50% to 70% decrease in wages - Creating income inequality and threatening millions of jobs. [Online].

Keynes, J. M., 1931. Economic Possibilities for our Grandchildren. s.l.:s.n.

Lordan, G., 2019. People versus machines in the UK: Minimum ages, labour reallocation and automatable jobs.

[Online]

Available at: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6886789/

Oppenheim, M., 2020. The Independent: 'How robots are increasing the gender pay gap". [Online]

Available at: https://www.independent.co.uk/news/uk/home-news/robots-gender-pay-gap-uk-automation- work-a9622511.html

Pilcher, C. a. P. a., 2018. The Impact of Industrial Robots on EU Employment and Wages: A Local Labour Market Approach, s.l.: s.n.

Rajadhyaksha, A. a. C., 2018. 'Robots at the gate: Humans and technology and work', s.l.: Barclay's CIB Imapct Series.

ST Engineering Aethon Inc, n.d. Three Ways Robots Boost Wages. [Online] Available at: https://aethon.com/three-ways-robotics-boost-wages/

Wallace, T., 2018. The Guardian: 'Robots will steal your wages but not your job'. [Online].