The global pandemic underlined the fragility of ageing labour markets in the world’s largest economies. This occurred at a time when their governments had already been running out of options to offset adverse demographics. Support for populist parties in post-Brexit UK, post-Trump US and increasingly in Europe, had often countered any improvement in migration trends which might have relieved the pressure on labour markets.
By the end of this decade the working age population of the world’s largest economies will be shrinking. A diminishing pool of labour is likely to mean companies will have to compete with each other to secure the staff they need. This is expected to drive up the labour share of GDP – the cost of employee compensation as a percent of GDP – and consequently drive down labour productivity, and the profit share for companies.
Fewer available workers will exacerbate inflationary trends in the global economy already resulting from other major macro-economic trends such as the challenges to globalisation and the transition to renewable energy (see investing into the new era).
Companies will have little choice but to respond with investment in technology focussed on automation, robotics and artificial intelligence (AI) as they strive to counter rising labour costs by increasing productivity. The promise of a “fourth industrial revolution”, enabled by smart robotics, could have profound implications for productivity and global economic growth. These benefits may persuade governments that the cost of some displaced workers as a result of these new technologies is a worthwhile trade-off. Against the backdrop of high indebtedness and higher inflation further stretching public finances, they offer some relief.
Any response by companies and governments to deteriorating labour markets, however, is only likely to help partially offset the inflation pressures resulting from the new regime, and not alleviate them entirely.
Severe cyclical shortages of workers
Interest rates have risen sharply in the past year to combat very high inflation in the wake of several supply side shocks that have hit the global economy. The pandemic was a key factor in disrupting supply chains and has also contributed to the current imbalance in labour markets. For example, the US unemployment rate fell to 3.4% in January and April 2023, which was the lowest level since May 1969, a period in which high and persistent inflation followed for years (chart 1).
The rise in interest rates should slow demand and domestic inflation pressures in time, but the shortage of workers appears to be more than a cyclical phenomenon. Arguably, unemployment rates were very low even before the pandemic, but the recent falls have been larger, for most countries, than would have been implied by the recovery in GDP. This suggests a more structural fall in unemployment rates.
In the US, the extent of the shortage of workers is underscored by the unemployment to (unfilled) job vacancies ratio (chart 2). The ratio remains below one and close to record lows. This means that even if authorities could match every available unemployed individual with a job in the same location, with the same desired terms and skillsets, there would still be a shortage of workers.
This is not unique to the US either. Looking at available data for Germany, France, the UK and Japan (chart 3), we find that their respective unemployment to job vacancy ratios may be higher than that of the US, but compared to their own history, all are close to if not at their record lows. There is a severe shortage of workers appearing in many developed economies.
Some of the staff shortages can be explained by a growing preference from employees to work fewer hours, which has lowered average hours worked compared to the period before the pandemic. This is forcing companies to hire more individuals to cover the same amount of work, reducing the available pool of labour.
In addition to fewer hours being worked, reduced absolute participation in the labour market has also contributed to shortages. Understandably, the pandemic appears to have made some people reluctant to return to work. This is noticeable amongst older age groups. Moreover, fewer individuals are choosing to work beyond retirement age, and more workers are choosing to retire early.
In the UK, there has been a remarkable rise in non-participants due to long-term sickness, possibly due to the poor state of the National Health Service (NHS), with long waits to see general practitioners and having procedures done. However, labour participation rates have recently started to recover as the rising cost of living has encouraged reluctant and retired workers to return. The UK reflects the cyclical trends, that are likely to lift participation in the near-term, and some of the structural changes that are hampering the recovery.
In the eurozone, while there has also been a preference for fewer hours worked on average, the participation story is very different. Thanks to significant and well targeted policy interventions in labour markets during the pandemic, many workers remained attached to their employers, reducing hours and receiving monetary support. The result has been a continued high participation rate, and a larger recovery in employment.
The “missing workers” in the US and UK are often referred to as “hidden slack” in the labour market, with optimists believing that these workers could return to employment and the labour market return to pre-pandemic trends. We are less convinced based on developments over the last year, and where the eurozone differs is that the hidden slack does not exist. This means that the eurozone is more prone to an inflationary shock if there is to be a spike up in demand, although for now at least, the labour market is not as stretched there as in the US and UK.
In the near-term, the reduced availability of staff is driving negotiated wage inflation higher. Companies have been forced to compete with each other to hire the staff they need. Higher average wages, combined with a recovery in jobs growth has pushed up the labour share of national income (GDP), to the detriment to the share for corporate profit. This is shown for the US in chart 4.
The split of national income between workers and companies has been relatively high in favour of firms since about 2003. The period since and up until the pandemic was also characterised by low inflation. Companies were successful in driving down costs, and expanding the profit share at the same time. However looking ahead, profit margins are coming under pressure from higher costs including wages, along with other structural inflationary pressures such as the challenges to world trade (see new world order will challenge globalisation) and the transition to renewable energy (see the accelerating response to climate change).
In the near-term, high inflation and the aggressive actions by central banks suggest that growth will slow to sub-trend levels, with a risk of turning negative and triggering recessions. This is a feature of the Schroders baseline forecast for the US, which should put the profit share under pressure. The fall in profits is one of the key motivators for companies to cut back expenditure and hiring, which in turn reduces aggregate income for households, and in turn slows demand and inflation.
Beyond the near-term cyclical pressures to the profit share, there are also structural factors that companies need to consider and start to prepare for. Poor demographics and ageing populations across most of the world are likely to slowly but surely worsen labour shortages, as greater shares of populations move into retirement age.
Ageing populations set to limit growth
The age dependency ratio is defined as the ratio of the total number of people aged below and above the working age (15-64) to the total number that are of working age. The ratio is a useful indicator of a country’s demographics, highlighting the degree of dependence of young and old to those that are working, especially where most countries still operate a pay-as-you-go system, where current pension liabilities are paid for by today’s taxpayers.
As chart 5 shows, age dependency ratios across most of the large economies are forecast by the World Bank to rise sharply in the medium-to-long term. Japan is famous for its elderly population, but Germany, France and Italy could look very similar to Japan by end of this decade. India and Indonesia are the only two countries that see an improvement (fall) in their dependency ratios this decade. The rest all see a deterioration, which continues into the following two decades as well.
Ageing demographics is not a new idea, and has been anticipated by investors for some time. However, companies are likely to see a more meaningful impact due to a step change in the number of workers coming up for retirement. This is clearer to see when looking the total population of those of working age (chart 6). The total working-age population of the largest economies is likely to start to contract from the end of this decade, and continue to do so from there onwards. However, the numbers are boosted by better demographics in India and Indonesia. When the pair are excluded, working-age populations actually started to contract in 2017.
For companies and investors, a shrinking workforce is a major problem. The vast majority of companies achieve growth in their revenues and profits by expanding their markets. This also requires greater output produced, or services delivered, often requiring a combination of capital, labour and land. Continuing to achieve growth while the workforce is shrinking is very difficult. It effectively means that profits growth will be reliant on productivity growth, rather than simply growth in demand and supply.
Could higher migration be the solution to labour shortages? In theory, it is possible for migrant workers to fill the gaps left by those indigenous retirees. However, the politics of migration has been very negative for some time, and this is unlikely to change anytime soon. Indeed, inward migration rates have been falling in recent years (chart 7), and where there have been spikes in the number of migrants, populist parties have been able to capitalise on xenophobia.
The number of migrants required is also very high compared to recent trends. For example, Germany currently has a crude net inward migration rate of 3.8 per 1,000 of its population. The World Bank forecasts Germany’s working age population to fall by 70.8 per 1,000 of its population between 2022 and 2030. This means that Germany would have to more than double its current rate of net inward migration in order to maintain a stable working age population over the eight-year period. France would have to see a five-fold increase!
Given the political backdrop, we highly doubt that migration will sufficiently offset the labour shortages expected from demographics in the coming years. A shrinking pool of labour is likely to mean that companies will have to compete with each other to secure the staff they need, driving up the labour share of GDP, and in the absence of stronger labour productivity, lowering the profit share for companies.
As mentioned earlier, firms are also facing other structural inflationary pressures. The disruption caused by the pandemic to global manufacturing has prompted companies to examine the resilience of their supply chains, and re-consider the lack of diversification. The emergence of a new world order is also testing existing trading ties, with politicians encouraging companies to reshore as much economic activity as possible.
Coping with these pressures and the expected worsening shortages of staff will be the biggest challenge for firms in the new regime. We believe that the only real solution available to firms is to substitute labour with capital. The use of automation, robotics and AI to reduce the reliance on workers is the only way most companies can continue to grow against this challenging backdrop.
Rise of the robots
The introduction of robotics, automation, and AI (artificial intelligence) is becoming increasingly popular among businesses in various industries. The reasons behind this trend can be traced to a variety of factors, including the need for increased efficiency, cost reduction, and improved productivity.
One key advantage of robotics and automation is their ability to perform repetitive tasks at a much faster rate and with greater accuracy than humans. This can lead to improved productivity and reduced costs, as businesses are able to produce more goods or services in less time. In industries such as manufacturing and logistics, robotics and automation are being used to assemble products, package goods, and transport materials.
Similarly, AI is being used to improve efficiency and productivity by analysing large amounts of data and providing insights that were previously impossible to obtain. In finance, for example, AI is being used to detect fraud, predict market trends, and improve investment strategies. In healthcare, AI is being used to assist with diagnosis and treatment, as well as to monitor patient health and wellness.
Another key advantage of robotics, automation, and AI is the ability to reduce human contact and limit the risk of transmission in the wake of the COVID-19 pandemic. In healthcare, robots are being used to disinfect surfaces and deliver supplies, while in retail, automation is being used to enable contactless payments and reduce the need for human interaction.
In addition to the benefits mentioned above, the adoption of robotics, automation, and AI is also being driven by the need to remain competitive in an increasingly globalized and rapidly changing business environment. Companies that fail to adapt to new technologies risk falling behind their competitors and losing market share.
Despite the many benefits of these technologies, there are concerns about the impact on employment. Some fear that the introduction of robotics and automation will lead to job losses, particularly in industries such as manufacturing. However, others argue that these technologies will lead to the creation of new jobs in areas such as design, engineering, and maintenance.
Furthermore, the adoption of robotics, automation, and AI can also have a positive impact on the environment. By automating processes and reducing waste, these technologies can help businesses to reduce their carbon footprint and contribute to a more sustainable future.
In conclusion, the introduction of robotics, automation, and AI is being driven by a range of factors, including the need for increased efficiency, cost reduction, and improved productivity. These technologies are also being adopted to reduce human contact and limit the risk of transmission in the wake of the COVID-19 pandemic, as well as to remain competitive in an increasingly globalized and rapidly changing business environment. While there are concerns about the impact on employment, the growth of the robotics and automation industry is creating new job opportunities in areas such as design, engineering, and maintenance. Additionally, these technologies can also have a positive impact on the environment, making them an attractive option for businesses seeking to become more sustainable.
The last 500 words in italics were written by a generative AI, after we asked it to give us an overview of why robotics, automation and AI are set to grow. The use of natural language algorithms combined with probabilistic models help these programmes deliver almost natural text and responses, allowing them to provide a more realistic service for companies. The explosion of generative AI has excited investors since the start of this year, and it is a fantastic technological leap for companies that can use it to deliver better and more cost effective goods and services.
Of course, the use of machines to aid and even replace workers is nothing new. The history of economic development has had three significant leaps in the introduction of new technology. The first, frequently referred to as the “Industrial Revolution” began in the 18th century, when water power was introduced to mills and factories, before steam power was combined with machines to increase the volume and speed of production. This development significantly boosted human productivity, and eventually led to the creation of the first industrial cities and factories.
In the 19th century, the second industrial revolution occurred as the invention of electricity was combined with machinery, and the assembly production line came into existence. These developments not only greatly expanded production capacity, but also led to specialisation in more complex manufacturing.
The third industrial revolution began in the 1970s with the introduction of memory-programmable controls, now referred to as computers, which helped introduce partial automation. Pre-programmed actions and decision trees led to the creation of robots, particularly useful when applied to assembly lines – in many cases aiding, but in some instances replacing workers in manufacturing processes.
Academics argue that the fourth industrial revolution is underway and it will be driven by information and communication. These changes will help harness network effects to improve the efficiencies of factories, by introducing smart robots that can communicate with each other, and also predict failures ahead of time, and thereby correct for errors.
Our view is that the new regime will be less about the creation of new technologies (although that may be possible), and more about the adoption and faster expansion in the use of current technologies. The motivator will be labour shortages which are likely to drive up the relative cost of labour to technology. However, the next few charts will highlight the scope for greater use of technology and automation, starting with industrial robots used generally in the manufacturing of goods and products. Chart 8 shows strong growth since 2011 in installations of such robots in the major regions. Growth was briefly interrupted by the pandemic, but promptly returned to a strong trajectory, with installations up 31.2% year-on-year (y/y) in 2021.
When looking at which industries have installed the most robots, we can see clear disparities. The automotive and electronics manufacturing sub-sectors, for instance, have consistently installed more robots than the food production, or the plastics and chemical production sectors, for example (chart 9). Although this is far from being a perfect analysis, it does suggest that there is scope for greater investment amongst some sectors.
Analysis on the density of robots installed by sector (or how many robots installed per 10,000 employees) would be ideal, but unfortunately, such data is not readily available. However, robot density in manufacturing by country is available, and is very revealing.
Data for 2021 shows that amongst US manufacturers, there were 274 industrial robots installed per 10,000 employees (chart 10). This is considerably above the world average of 141, but is significantly below two of the biggest manufacturing powerhouses of Japan and Germany, which have robot densities of 399 and 397 respectively. But even Japan and Germany fall short of Singapore and world leader South Korea, with densities of 670 and 1,000 respectively. The huge differences between even some of the most advanced economies show that there is huge scope for further investment and integration of robotics technology.
As the relative cost of labour continues to rise, we would expect more adoption of robots. One reason why some countries have lagged behind could be that the robots available are not capable of performing more complex tasks for a given cost. Thankfully, costs are falling, as smart robots are quickly being introduced. This new generation of robot represents the fourth industrial revolution, and what makes them special is that they have the ability to communicate with one another. Amazon is famous for having warehouses that are almost entirely run by robots. They can zip around carrying stock, avoiding each other and work together. As this technology trickles down, the design of not only stock management, but also manufacturing processes are likely to undergo a revolution.
In 2021, only 8.2% of all industrial robots installed were smart or “collaborative” robots. Though this represents a steady rise in the total share from 2.8% in 2017, and in itself is a 50% increase compared to 2020 (chart 11). Smart robot technology may help introduce robots into factories and production facilities that were previously thought to be too complicated. These could help lower costs for businesses and boost productivity.
Are you being served…by a robot?
The use of robots has generally been more accepted in manufacturing and production rather than in services, and so the latter has lagged behind in the use of such technology. The “human touch” is seen as important for many customers, and a preference for dealing with individuals has restricted investment in this area. However, the pandemic was a major disrupter to this trend, helping robots gain greater acceptance amongst customers. Suddenly in 2020, customers did not want to, or were not allowed to deal with other people for fear of spreading the virus. And so, technological solutions were introduced to help minimise customer/server contacts in various service provision scenarios.
In truth, the introduction of these technologies began many years ago. A great example is the self-checkout system at most small convenience stores and even larger supermarkets. This is a common case where a robot which enables the customer to complete his or her purchase without the need to interact with an employee. But there are other examples appearing where robots are reaching new heights in the level of skill in replacing human workers. For example, there are several fast food restaurants where the entire kitchen is run by robots. All meals are ordered through computer terminals, then prepared and served by machines. Given the size of the global restaurant industry, this has potential for huge productivity gains.
Ultimately, two pre-conditions will dictate the pace of rollouts of such technology in service sectors. The first is the willingness of customers to be served by robots and machines, and the second is whether these technologies are economically viable, or even advantageous once set-up and maintenance costs are taken into account. Many service industries will take time before both pre-conditions are met. But consumers are certainly becoming used to dealing with these technologies on a more regular basis. Consumer robots, or those that are owned and used in homes, are growing rapidly. Global consumer spending on robots for entertainment use (such as toys) reached $3.2 billion in 2021 – up 26% compared to five years earlier (chart 13). However, purchases of robots for domestic use, such as robot vacuum cleaners, or robot lawn mowers, have jumped 169% over the same period, reaching $5.8 billion.
Importantly, the strong growth seen in sales of robots for both domestic and entertainment use suggests that acceptance by households is increasing, which should allow service providers to slowly introduce more technology to the customer experience.
Learning to learn
Much of the technology discussed so far has been designed to carry out some form of repetitive task. More advanced robots can co-operate with one another to allow for specialisation, but so far, these have all relied on a set amount of inputs and parameters to be able to operate. AI technology attempts to go a step further and aims to reduce the amount of inputs required. AI still operates within a set of parameters, but it is designed to identify problems and solve them independently of human inputs. Working through an iterative process – or trial and error – it can learn how best to proceed.
Early examples of AI focused on solving mathematical problems, with one of the most famous being Deep Blue, a chess-playing system that defeated world champion Garry Kasparov in 1997. Since then, focus has turned to developing the use of “natural language” – or the ability to understand language and respond in the same form.
“Logic will get you from A to B. Imagination will take you everywhere.”
Albert Einstein – Physicist (1879-1955).
The evolution eventually reached what is now referred to as “large language models” (LLMs), examples of which includes ChatGPT – a popular chat bot based on the Generative Pre-trained Transformers-3.5 and 4. These models are designed for conversational applications, such as generating a designed length, format or level of detail of language being requested – such as the text highlighted above.
The release of ChatGPT by OpenAI in November 2022 along with others such as Google’s Bard in March 2023 created a tremendous amount of excitement around the applications of this technology. Not only could it be utilised in language processing, for example, translation work, but could also help creative industries tremendously when combined with other technologies. There are now many examples of AI programmes that can generate imagery, produce graphic designs, and even generate computer code, potentially vastly reducing the time required to carry out these complex tasks.
Investors are therefore understandably already mobilising capital and are looking for opportunities in this space. Investment opportunities range from firms directly involved in creating and advancing AI technologies, to those producing the hardware and infrastructure required to support these systems. This is very visible in the performance of US large technology companies over the first half of 2023 compared to the rest of the US equity index.
While this technology continues to advance, investors will continue to focus on the inventors and innovators. Over the medium term, the diffusion of this technology and its application will likely attract investors to those companies that seek to gain the greatest returns as they apply AI to enhance their productivity.
So far, data management, processing and cloud solutions, followed by medical and healthcare firms, and thirdly financial technology firms (including investment banks) have invested most in AI. Data to 2021 shows rapid growth, which is only expected to accelerate further with the introduction of LLMs and generative AI (chart 14).
Where can AI be useful?
Unlike most robots that are used in the physical production of goods, or the physical delivery of services, AI is far more likely to have an impact on the knowledge economy. A detailed study by Goldman Sachs looking at the degree of automation that can be introduced by tasks in the workforce found that around two-thirds of current workers are exposed to some degree to automation, and around a quarter of current work tasks (not jobs) could be substituted by AI.
The study looks at the share of time spent on tasks that could be automated by AI, and makes a comparison by sectors (chart 15). The study classifies jobs as likely to be replaced by AI where the proportion of associated tasks that can be automated is greater than 50%. Where the share of tasks that can be automated falls between 10-49%, it is assumed that AI will complement and support the worker. Lastly, where automation is less than 10% of the tasks, then no automation is introduced (not worthwhile).
Overall, the study suggests that 7% of US employees would be substituted, 63% would be complemented, and 30% would mostly be unaffected. The two sectors that could benefit the most from AI are legal and the office and admin support sectors, where both could reduce staffing by around a third. There are sectors that could replace about a 10th of their staffing, but overall, the majority of service sectors are more likely to use AI to complement existing workers, and in the process raise productivity.
Professions that are least likely to use or be replaced are those with more manual tasks such as building and ground cleaning and maintenance; installation, maintenance and repair; construction and extraction; and production roles. These are more likely to be aided by normal machines, or even robots, but not AI according to the study.
How much could technology boost productivity growth?
Estimates on the potential benefits from technology vary hugely. Many studies focus on savings to companies and sectors brought about by time saved, using workers’ wages as a unit of measurement to translate the benefits into a monetary figure.
For example, a Mckinsey & Co report (June 2023) estimates that generative AI could add $2.6 trillion - $4.4 trillion annually to the global economy across 63 use cases analysed. It found that labour productivity could rise by between 0.1% and 0.6% annually through to 2040. However, if generative AI is combined with other technology such as robots, then that could rise to between 0.2% and 3.3% annually. Also interestingly, it found that 75% of the monetary benefits are likely to come from four areas: customer operations task; marketing and sales; software engineering; and lastly research and development.
PWC (2020) estimates the potential boost to global GDP of $15.7 trillion by 2030 (or 14% of GDP), with China seeing a 26% boost to its GDP growth, and the US seeing a potential 14% boost. It suggests that the biggest sector gains will be in retail, financial services and healthcare.
These estimates along with the studies discussed earlier are largely based on the capabilities of current technology. As these technologies develop further, so will their useability and adoption rates. But importantly, and as mentioned above in the section discussing robots, the availability of the technology is not enough. The two pre-conditions for adoption are required, being the acceptance by customers, and the adoption being economically advantageous. At this early stage in the development of AI, there is little consideration of the adoption costs for firms, or for wider society.
The acceptance could be difficult. The famous Skynet and the T-800 – fictional depictions of how AI evolves – do no favours. The more meaningful resistance, however, is likely to come from the fear of job losses, especially of those that have invested in education and therefore moved up the value chain. Political blockages could slow progress.
Growing labour shortages, high indebtedness and higher inflation are likely to exacerbate the pressure on public finances (see the return of “fiscal activism”). The promise of a productivity and GDP growth boost could persuade governments that the cost of some displaced workers is a worthwhile trade-off.
Meanwhile, economic viability will also be a challenge, at least at first. As the cost of these technologies fall, the benefits relative to the expected rise in labour costs will start to make sense. The use of technology could enable greater onshoring of production and the provision of services. If the creation of new high-skilled domestic jobs also follow, then the investment may even attract fiscal support.
While the development of these technologies is largely independent of the current economic climate, we believe that the shortage of workers will force companies to begin to make greater use of robotics, automation and AI in order to manage their rising costs. Higher cost inflation will be the catalyst for investment, but given the estimated benefits outlined earlier, may only slightly offset the higher inflation pressures coming from the new regime.
Summary and conclusions
- A shrinking pool of labour is likely to mean that companies will have to compete with each other to secure the staff they need. This is expected to drive up the labour share of GDP, and consequently drive down labour productivity, and the profit share for companies.
- Given the political backdrop, we highly doubt that migration will sufficiently offset the labour shortages expected from demographics in the coming years. Inward migration rates have been falling, and where there have been spikes in the number of migrants, populist parties have countered these trends.
- The use of automation, robotics and AI to reduce the reliance on workers is the only way most companies can continue to grow against this long-term challenging backdrop.
- We expect labour shortages are likely to drive up the relative cost of labour to technology and drive the adoption of existing productivity enhancing technologies and automation. The huge differences in adoption of such technologies between even some of the most advanced economies, however, show that there is huge scope for further investment and integration of smart robotics technology.
- By harnessing network effects that help to improve the efficiencies of factories, the use of smart robots is being hailed by some as the fourth industrial revolution. The promise of a productivity and GDP growth boost could persuade governments that the cost of some displaced workers is a worthwhile trade-off for adopting smart robotic technologies more widely.
- Services sectors have lagged behind in the use of such technology compared to the manufacturing and production sectors. While consumers are certainly becoming used to dealing with these technologies on a more regular basis, acceptance and economic viability of such technology in the service sectors will take time.
- The shortage of workers will force companies to begin to make greater use of robotics, automation and AI in order to manage rising cost, but such actions may only slightly offset the higher inflation pressures coming from the new regime.
Subscribe to our Insights
Visit our preference center, where you can choose which Schroders Insights you would like to receive
The views and opinions contained herein are those of Schroders’ investment teams and/or Economics Group, and do not necessarily represent Schroder Investment Management North America Inc.’s house views. These views are subject to change. This information is intended to be for information purposes only and it is not intended as promotional material in any respect.