Robotics & Automation: The productivity paradox
News and Insights

Robotics & Automation: The productivity paradox

Advances in technology herald the dawn of a new era. Artificial Intelligence or “A.I.” enabled by super-fast computers and accessed over high-speed wireless internet connections, will revolutionize the way we work and live.

In a few years, intelligent systems and robots will perform physical and cognitive tasks more efficiently and accuracy than people, fully autonomous transport systems will operate safety and without traffic congestion, fossil fuel pollution will be cut, and diseases cured, or prevented before they even occur. These are just a few of the grand claims and predictions now common place in the media.

Skeptical onlookers are, however, quick to point out that these world-changing technologies so far appear to have failed to deliver any real benefit to the economy. If modern technologies are so powerful then why have we not seen more tangible impact on the productivity growth of countries deploying them? In this Thematic Insight, we explore how price deflation and something as simple as old habits may lie at the heart of this “productivity paradox”.

You can see the computer age everywhere but in the productivity statistics

Robert Solow, Nobel Prize winning economist quoted in “We’d better watch out,”
New York Times Book Review, 12 July1987, page 36

What is productivity?

In the field of economics, productivity is a gauge of competitiveness and efficiency. Specifically it is a measure of economic “output” per unit of “input”. When human labor over a given period of time is used as the “input”, it is known as “labor productivity”, and when capital is used as the input, it was known as “capital productivity”. Labor and capital are sometimes used in combination and this measure is known as “multi-factor productivity”. Also, economist Robert Solow described a number of other less tangible inputs beyond capital and labor, and these are known as “Solow’s Residual”.1 However, when looking at productivity in the economy as a whole, the standard is to look at “labor productivity”, and to use the value of gross domestic product (GDP) as the “output” and total hours worked by the labor force, as the “input”.

Growing productivity

Growth in labor productivity implies that the workforce have produced more in the same period of time, or that the same level of production can be achieved with a smaller workforce, fewer working hours, or both. Through productivity growth, businesses can “create value” and pass that value on to employees in the form of higher wages, to customers by way of lower prices, or use the excess profits for reinvestment in the business or as cash reserves.

Productivity growth can be achieved in a number of ways. The efficiency of the workforce might be improved through training and education, and of course hands-on practice and experience. Health and motivation also come into play here. Growth can also be gained through investment in tools and equipment, such as robotics and automation systems, as well spending to improve the fabric and organization of the factory itself. Henry Ford achieved a huge advantage over his competitiors by rethinking the entire production process. In his 1922 autobiography, “My Life & Work”, he describes breaking down tasks into individual steps to develop the most efficient way to complete each job:

“With one workman doing a complete job he could turn out from thirty-five to forty pieces in a nine-hour day, or about twenty minutes to an assembly. What he did alone was then spread into twenty-nine operations; that cut down the assembly time to thirteen minutes, ten seconds. We then raised the height of the line eight inches – this was in 1914 – and cut the time to seven minutes. Further experimenting with the speed that the work should move at cut the time down to five minutes.”2

A brief history of productivity growth

During the first industrial revolution, technological innovation delivered machines capable of large scale production and transport systems able to distribute goods great distances over land, more efficiently than anything that had come before. As a result, productivity, which had remained in “hibernation” for approximately three centuries, showed a sharp upturn in trajectory from 1820 to 1870.

Nobel Prize winning economist Edmund Phelps wrote in “Mass flourishing”3, that Britain was the first to hit this inflection point, followed by America, France and then Germany. Productivity (or, in Phelp’s words, “domestic output per capita”) in Western Europe grew 63% between 1820 and 1870, and by another 76% between 1870 and 1913.

Productivity then continued to grow through most of the 20th century and was particularly fast between 1925 and 1950, and even continued to grow through the Great Depression of the 1930s. The trend resumed from 1950 up to the early 1970s where it declined sharply for a whole year, and it is from this point on that the growth rate has settled at a level far below what the developed world had experienced for most of the previous 150 years. This is the subject of economist Bob Gordon’s 2011 book, “The great stagnation”. Productivity growth was so slow in the 1980s, just at the time when personal computers and “word processors” were being adopted more and more broadly in offices and homes, that Robert Solow was prompted to make his now famous quip which we used on the title page.

Although growth recovered somewhat in the late 1990s and early 2000s, it did not last long and since 2010 the US economy has seen the lowest growth in productivity over a sustained period in recent history. Can we simply dismiss this lack of growth in productivity as a short term anomaly in a long term trend, or has something changed fundamentally?

Automate to increase productivity

Robotics, automation and artificial intelligence are central to this dilemma, since increasing productivity is the primary reason why companies invest in robotics. As robotic and automation solutions become more affordable and smarter, the argument for investing in these systems becomes increasingly compelling, especially when faced more regulations and rising labor costs. Stripping out 2009, where demand dropped sharply following the financial crisis, growth in shipments of industrial robots (robots used in factories) has been approximately three times higher after 2009, than the growth rate seen in the years of economic boom prior to 2009.

Adapting processes

The flippant answer to the productivity paradox is that we spend so much time trying to remember passwords, sending one another messages or taking “selfies”, that the benefits of technology are quickly outweighed by the distractions. In fact many of current technologies are designed to entertain us, rather than to make us more productive. While there is an element of truth to this, it is unlikely to be the whole reason, otherwise businesses would have stopped investing in technology long ago.

One of the most compelling arguments, is also one of the simplest. When new technology arrives, it takes time to appreciate its full potential and to adapt processes to make best use of it. From Thomas Newcomen’s 1712 invention of a steam engine4 used to stop mines flooding, it then took another 70 years before steam power was combined with innovative mechanics to produce a continuous rotary motion critical to the steam powered machines used in the production of goods and in the locomotives used to transport the goods. In other words it took 100 years for the technology behind steam power to have its greatest impact on the economy, in the form of the industrial revolution.

In the second industrial revolution, we saw the same multi-decade lag in the use and impact of electricty. While the understanding of electricity made great advances in the early 19th century, it wasn’t until the second half of the century that pioneers such as Alexander Graham Bell, Thomas Edison, Nikola Tesla, George Westinghouse and others turned a scientific curiosity into the critical ingredient of modern manufacturing. In 1880, light bulbs were available and Edison had built electricity generating stations in Manhattan and London.5 Yet 20 years later only 5% of mechanical power in American factories were electric,6 and steam remained dominant. The slow pace of adoption is understandable. Steam power in factories was generated by a massive steam engine which would drive single shaft running the length of the factory, often overhead, and all the machines, presses and looms ran off this central shaft through a complex system of pulleys, belts and gears, sometimes through holes in the ceiling to give power to upper floors, or via subsidiary shafts to other buildings or even outside. Machines requiring more torque would be positioned closer to the central shaft and those less demanding could be further away.6

When factory owners first converted to electricity, they simply replaced the steam engine with an electric one and kept the old central drive architecture and network of belts and gears connected to it. As a result they saw only marginal benefit from the investment. The true value of electricity in manufacturing was only realized in the 1920s, when businesses began to understand that while small steam engines were hopelessly inefficient, small electric motors worked well, and electricity could be fed to each of these motors wherever it was needed. The real value here was indirect. Production lines could now be arranged to optimize the flow of work, rather than by the hierarchy of need for proximity to the central drive.

Even today the adoption of innovative technologies can take decades before old habits and processes are rethought and retaught enough to realize the full potential of the innovation. A current example is the internet. The first iterations were built in the 1960’s7, with the US military and academic centers connecting to it in the 1980s, and then in the 1990s it opened up to more commercial interests. But when consumers bought8 and rented9 music and movies online in 1998, we combined this new internet technology with old processes: music and movies ordered online but mailed out by post on optical discs (DVDs and CDs). This must seem a strange concept to the young “streaming media” generation. It is only more recently that the underlying infrastructure of high speed mobile internet connections, tens of billions of connected devices, cars and sensors, “hyper-scale” data centers storing application software, artificial intelligence algorithms and data, has developed to allow processes to really benefit from its potential.

With these examples, it is clear to see how technological innovation often requires a gestation period before it can be successfully and fully adopted into an economy. It also suggests that we may now be at the very start of another revolution in productivity, this time driven by intelligence machines and algorithms, and perhaps the slow down in productivity growth is simply the calm before the storm. Adding A.I. to the internet may become the greatest driver of productivity the modern world has seen, depending of course on the level of intelligence in the A.I., in other words the I.Q. of the A.I. The progress of software and A.I. is difficult to measure objectively, but there is reason for optimism. A 2017 survey of A.I. experts published by Oxford University Press, showed two-thirds of respondents agreed that progress in A.I. had accelerated in the second half of their careers, and gave a 50% probability that A.I. would be able to perform all human tasks by 2060 (Asian respondents suggested 2045).10

Cheaper machines building cheaper electronics

Since the 1970s the cost of technology has fallen at a steady pace, from mainframes, chips and sensors, to computers, mobile phones, TVs and robots. The cycle is self-perpetuating, since lower priced electronics are likely to find a broader customer base and greater volumes in production often drive economies of scale and scope, and therefore allow further price reductions. Lower cost component can then be used to reduce the cost of robotics and automation systems used in the manufacturing process. Of course in some cases prices do not appear to have fallen far. The latest phone announced by Apple CEO Tim Cook in September is said to be the most expensive iPhone yet. However, the technology inside the new phone is significantly more powerful, and therefore in a sense more “valuable”, than the previous one. So the new technology has much greater value for only a slight increase in price. Effectively this is still deflationary.

The calculation for productivity measures “output” in terms of GDP, and GDP is made up of both the price and volume of goods and services produced in an economy. So as technology prices decline, the effect on “output” and in turn on productivity growth will be negative. And since technology has increasingly pervaded so many aspects of our lives, through the office, our home, the hospital and the factory, it is likely that the total exposure to technology for the economy as a whole has steadily risen. Data from Moody’s illustrates the extent of technology growth, although not the full extent of the exposure since it only captures the IT sector (hardware, software, services and the internet) and does not extend to technology used in telecoms, media, finance and the consumer or retail sectors. A study from economist Ian Hathaway shows that the IT sector accounted for 5.2% of US GDP in 2015, a massive increase of approximately 700% since 1980.11


We may simply be too early in the adoption of internet, mobile networks, robotics and automation systems to have yet seen the real impact of these innovative technologies on productivity growth in the economy. Price deflation so common in technology products and solutions, combined with the growing importance of technology in the economy may also have become factor in hiding some of the productivity gains over the last two or three decades. In fact, we believe that both arguments may be helpful to explain the “productivity paradox” and may well co-exist. They are certainly not-mutually exclusive.

As technology advances and machines become both smarter and cheaper, robotics and automation systems will likely be deployed more broadly not just in factories, but also in our homes, offices, hospitals, and infrastructure and transport systems. We believe these advances in technology herald the dawn of a new era and are likely to trigger a great leap forward in productivity growth in the global economy.

Credit Suisse Asset Management has designed a strategy to provide clients with “pure-play” exposure to the compelling long-term secular growth theme of robotics & automation.

For further information please click here