Thematic, innovative and bespoke investment solutions and thinking

The Productivity Conundrum (Part III of IV)

The Productivity Conundrum (Part III of IV)
– The Power of Digitisation

Issues to be addressed in this paper

Life is not always predictable. I wrote parts I and II of this research paper in late 2019 and expected to do the rest within a few months, but the world unexpectedly got struck by COVID-19, and Trump started to play silly games ahead of last November’s elections. Suddenly, the world around ARP changed dramatically, and we had to change accordingly. Now, as we approach the end of 1Q21, although the world has not yet returned to normal, it is time to finish this research project.

As parts I and II are almost 18 months old, you may not even remember what they were about so allow me to briefly summarise them. Part I covered the big picture – how productivity growth affects GDP growth and why we need productivity to grow robustly in the years to come to generate any GDP growth. In part II, I dug a little bit deeper – how we are on the cusp of the next stage of the Digital Revolution, and why that could result in millions of jobs being lost to advanced robotics over the next few years, but also why unemployment will be a much bigger problem in some countries than in others. You can find both of those papers on ARP+ here. Now, in this part III, I will look at specific technologies associated with the Digital Revolution and how they will create investment opportunities.

A few words on productivity growth

Before I start, I should make a simple point. In parts of the investment community, the Digital Revolution is sometimes referred to as the Second Coming; however, I would argue that technology has improved pretty much nonstop since the beginning of the Industrial Revolution. Ever since Thomas Savery invented the steam engine toward the end of the 17th century, technological progress has been a key driver of productivity growth. The Digital Revolution adds nothing new in that respect.

The most powerful spike in productivity growth the Western world has ever experienced (at least since we began to compile data) occurred during a 10-year period from the mid-1950s to the mid-1960s. Eisenhower had returned from the war in Europe and had told Congress how quickly Hitler could move his army around because of something called autobahns. Congress followed through by approving the establishment of the interstate highway system.

To continue reading...

We publish investment strategies and opportunities in our research papers. This research paper is available to professional investors as part of ARP+ subscription.
More about subscription
Already a subscriber? Login

At about the same time, commercial aviation took off (no pun intended), and both Americans and Europeans began to fly over longer distances. A post World War II drive away from using coal as a direct source of motive power towards using it mostly as fuel in electric power plants also enhanced productivity. As a result of all those factors, productivity got a significant boost, and, for a 10-year period, it grew by about 2.5% annually.

The second most powerful productivity boost we have ever enjoyed occurred during the first stage of the Digital Revolution – when the IPO of Netscape in 1995 allowed everybody access to the internet – but not even the arrival of the World Wide Web could enhance productivity by much more than 2% annually. Again, productivity growth lost pace after about ten years and has been on a decline ever since.

New technologies in the offing

As I began to assess which of all the new technologies that could possibly have the biggest impact on productivity going forward, I quickly realised that to cover them all would be a mammoth task. Hence, many innovative technologies will not be covered in this paper, and fusion energy – probably the most ground-breaking technology of them all – will be covered separately in part IV. In the following, I will zoom in on:

  1. Artificial Intelligence (AI);
  2. Internet of Things (IoT);
  3. driver-less cars; and
  4. blockchain.

Should you want to read more about some of the other emerging technologies not covered in this paper, I suggest you read Bill Gates' take on 10 breakthrough technologies or World Economic Forum's take on the Top 10 emerging technologies.

Artificial Intelligence (AI)

The rollout of AI will have a massive impact on society. When robots first arrived, they displaced mostly workers engaged in assembling – virtually all displacements were on manufacturing floors. With AI now being integrated into the next generation of advanced robotics, the service industry will also be affected. So will banks and many other corporates not impacted by the first wave.

As far as AI is concerned, pretty much the only thing everybody agrees on is that productivity will benefit from the rollout of AI. Beyond that, nobody agrees on anything to do with this new technology. The most optimistic commentators argue that AI is going to create many new jobs. We just don’t have the imagination to grasp precisely what those jobs will contain, the optimists argue. See for example this report.

An argument I often come across from the optimists is that implementing a new technology has never cost job losses on a net basis and neither will AI. I am not convinced that is a valid argument, though. The overwhelming impact this particular technology is going to have on job security makes comparisons with other, new technologies less relevant, but that is a story for another day.

In my opinion, the best argument in favour of rolling out advanced robotics and AI as soon as possible is the fact that large parts of the world will suffer from adverse demographics for many years to come. If the workforce is going to shrink, we will quite simply need those new technologies, unless we are prepared to accept a lower output.

A more balanced argument I often come across is that, whilst AI may cost some job losses, most people will benefit as a result. The average job will get more interesting, and general health, possibly even life expectancy, will improve. As a result, society at large will be better off, even if some people will struggle to land a new job, the argument goes.

The more pessimistic commentators argue that, whilst skilled workers who lose their job can be successfully re-trained, unskilled workers cannot. You don’t easily retrain a lorry driver to sit in front of a computer, is the argument. Hence, for the first time in the post-Industrial Revolution era, the introduction of a new technology will cost many their livelihood, the pessimists argue.

The most pessimistic argument I have come across goes even further. AI will, at first, destroy both jobs and privacy and eventually, as AI moves closer and closer to human intelligence, it will become a threat to human existence, the argument goes.

Whether you believe the optimists or the pessimists, there can be no doubt that AI-related revenues will multiply in the years to come, and it is expected that nowhere will those revenues grow faster than in North America, which has taken a commanding lead in the race for AI supremacy (Exhibit 1).

Exhibit 1: AI revenue forecast by continent, 2016-2025 ($Bn)
Source: Financial Times

There is another reason why the trend towards moreautomation will most likely accelerate, and that has to do with the current taxsystem. Under existing tax rules, capital investments offer a tax break,whereas labour is taxed. In other words, companies have a powerful incentive toautomate as much as possible. Unless the rules are changed (and they probablywill be), the biggest loser could end up being the government, as automationcould result in a dramatic drop in tax revenues.

The Internet of Things (IoT)

IoT works through intelligent gathering of data. Sensors are embedded in various locations, and the data those sensors provide is used to enhance efficiency. Let me provide a few examples. Smart devices can now be inserted in the human body. Alternatively, they can either be injected or swallowed. Those sensors will detect problems at an early stage, and healthcare providers will be able to significantly reduce the cost of treatment. Adding to that, as a result, longevity should improve.

At the household level, by far the biggest opportunity is in automating domestic tasks such as cleaning and shopping. Imagine, for example, that you forgot to buy milk for the kids’ breakfast. Once the IoT technology is in place, your fridge will send a message to your local supermarket, which will deliver the milk on your doorstep overnight. Automating trivial tasks like that will have an immense impact on how we spend our spare time, even if the benefit cannot be measured in official GDP data. By far the biggest opportunity, though, is to be found in industry. Optimising a firm’s operational set-up, managing its inventories more efficiently, etc., will save trillions of dollars.

Public authorities will also benefit. I have already mentioned the impact IoT is likely to have on the healthcare budget, but other areas will benefit as well. Traffic flows can be optimised, water leaks can be quickly detected, and air and water quality can be better monitored. In a recent trial in Denmark, they introduced smart traffic lights that assembled traffic flow data from oncoming cars’ GPS data. On average, the time spent commuting in the area was reduced by over 20% during the trial period.

Despite (or maybe because of) the many obvious benefits to society from IoT, estimates on the economic impact vary dramatically. At the most bullish end, McKinsey Global Institute published an extraordinarily bullish paper on IoT all the way back in 2015, suggesting that the annual global GDP impact could be as much as $11 trillion by 2025. Subsequent estimates have come in lower than that, as various barriers to entry have slowed down the rollout. In a more recent paper, Bain & Company concluded that, due to barriers to IoT adoption, implementation will not be rolled out as quickly as previously predicted but that the impact will still be dramatic.

Any negatives? The cost of implementation is very high, so that is one, but there are other negatives too. IoT heightens risks to personal privacy and the confidentiality and integrity of organisational data. For starters, theft of private information (whether personal or corporate) may lead to illegal exploitation, but that is not all. Companies or individuals may lose control of confidential data, or incorrect data that has been (legally) resold may cause all kinds of problems, because incorrect data cannot easily be deleted.

For example, a bank may decline a request for credit because of duplicate names and, suddenly, an otherwise respectable client is listed as a bad credit for ever and ever. When developing IoT, it is therefore important that customers are offered transparency into the collection and use of data, and that customers can specify what they will allow the data to be used for.

Driverless cars

Driverless cars will affect society much more dramatically than generally perceived. To begin with, safety will improve, as around 95% of all traffic accidents are due to human error (source:, and human error will be non-existent, once everybody has converted.

A significant reduction in the number of injuries and fatalities will benefit public healthcare budgets but driverless cars could, at the same time, spell problems for the insurance industry. If there are virtually no accidents, insurance premia will plummet, and auto insurance is the biggest source of profits for the insurance industry today. Garages will also struggle to maintain current business levels, and millions of delivery van drivers all over the world will find themselves displaced.

On a more positive note, traffic congestion will ease. Driverless cars do not accelerate and brake unnecessarily, and such human behaviour is the cause of much congestion. Also, as the number of accidents drop, traffic will flow more freely. All this means a significant cut in commuting time which will have a positive effect on productivity.

Another piece of good news is that driverless cars emit far less carbon dioxide and other gases into the atmosphere than vehicles driven by humans do. Today, cars and trucks account for about 20% of all emissions, emitting 24 pounds (~10.9 kg) of carbon dioxide and other damaging greenhouse gasses for every gallon (~3.8 litres) of petrol used (source: Union of Concerned Scientists). As driverless cars take over, groups of cars will accelerate and brake simultaneously, leading to better fuel economy and lower emissions. EVs driven by humans will have some positive effect on emissions but not to the same extent.

The transition to driverless cars is going to be very slow, though, meaning that all the implications mentioned above will only be felt gradually. Most of the estimates I have come across suggest that only by the 2050s will most cars be fully autonomous. See for example this article.


Blockchain is an emerging technology mostly mentioned in the context of cryptocurrencies. At first, it was indeed developed as a supporting technology for bitcoin but has since taken a life of its own. In essence, it is a tool to maintain transparent ledgers that verify transactions with minimal third-party involvement. The ledgers are not held in a central location (at banks) but are spread across a network of computers, and it is transparent in the sense that every transaction is made public for all to see.

Eventually, blockchain will revolutionise all payment systems but, to begin with, expect it to be used mostly in connection with cross-border payments. About $15bn worth of fees are earned by commercial banks every year on fees to do with cross-border payments. That is a massive amount of money, and it is this fee income that banks stand to lose first as a result of blockchain.

Before blockchain will take off, two things need to happen, though. Blockchain provides (by design) full transparency, but company ABC does not necessarily want everybody else to know that it does business with company XYZ. Secondly, bitcoin must be replaced by ‘stablecoin’. Today’s digital currencies are far too volatile to ever be broadly accepted as a means of payment. Furthermore, I don’t believe central banks will ever surrender their monopoly on issuing money. You should therefore expect the cryptocurrency market to be regulated at some point, and you should expect central banks to begin launching their own digital currencies in the not so distant future. Having said that, all those digital currencies will probably embrace the underlying blockchain technology as well.

Commercial banks have been reluctant to embrace blockchain up to this point and that is, in my opinion, a massive mistake. The blockchain technology could certainly undermine a major source of income for commercial banks, but the alternative is much worse. Whether a bank embraces the technology or not, blockchain is coming and, if used intelligently, it can help banks save billions of dollars in other areas (Exhibit 2).

Exhibit 2: Savings potential from blockchain-based solutions ($Bn)
Source: McKinsey & Company

If commercial banks keep sitting on their hands, the banking industry is at risk of becoming the next fatality of disruption – just like many brick and mortar retailers seriously underestimated online retailing and paid the ultimate price for that mistake. Unless banks take much more decisive action, I am not sure there will be many commercial banks left 15-20 years from now. The blockchain technology is so encompassing that it could potentially make banks redundant.

Gartner have done a very comprehensive study on blockchain and have found that, by 2030, the value-added of the blockchain technology will most likely reach more than $3 trillion worldwide. As per the study, the technology will likely be ramped up pretty quickly, as the benefits are obvious. In no particular order, five of the most significant advantages are:

  1. Faster transactions
    Whereas many transactions take days to settle now, the blockchain technology should allow most transactions to settle in minutes.
  2. Reduced transaction costs
    Peer-to-peer transactions can now be settled with no intermediary (bank) in between.
  3. Process integrity
    Unlike now, where many aspects of a transaction can (and do) go wrong, users of the blockchain technology can trust that transactions will always be executed exactly as commanded.
  4. Durability and reliability
    Because blockchain networks are decentralised, there is no central point of failure, and users will be better able to withstand malicious attacks.
  5. Empowered users
    Cutting out the middleman (bank) by definition grants more powers to the users, as they are now in control of all information and all transactions.

One significant implication of blockchain is what we call Electrification of Everything which, as you may be aware, is an investment theme we associate with two of the megatrends we have identified – Climate Change and the Era of Disruption. Electrification shall be required if we are gong to take full advantage of the blockchain technology and electrification of all heating and transportation could almost double the demand for electricity over the next 30 years.

Electricity can be produced in different ways, and it is estimated that about two-thirds of all electricity today is produced by burning fossil fuels, mostly coal and natural gas. That said, electricity is increasingly generated from renewable sources like wind and solar and will eventually come from fusion energy, but more about that in part IV of this paper.

Investment opportunities in technology

When seeking to take advantage of the Digital Revolution investment-wise, the starting point has to be technology but, from that point onwards, it gets murky quite quickly. For starters, today’s disruptors may turn into tomorrow’s disrupted, so investing passively in technology is not necessarily a winning strategy. Secondly, tomorrow’s winners are not necessarily household names today and may only be accessible via risky venture capital funds. Thirdly, when investing in listed equities, you take on a great of deal of equity beta risk, which may not be what you have appetite for in the current environment.

Talking about equity beta risk, allow me to make a couple of points. Firstly, the simplest and cheapest way to gain exposure to AI, IoT, driverless cars and/or blockchain would be to invest in an ETF, designed to benefit from the Digital Revolution, and there are many of them; however, it is not that simple to manage your equity beta risk in a passive investment vehicle like an ETF. I would therefore urge you to consider this aspect before ploughing into an ETF of this sort.

Secondly, almost daily, somebody tells me that a new tech bubble is forming in the US – very much a repeat of the dotcom bubble some 20 years ago – but that is quite simply not true. I agree that Big Tech have done spectacularly well since the Global Financial Crisis, particularly over the last 12 months, where the world has been struck by COVID-19. As you can see in Exhibit 3 below, almost all the solid performance in the S&P 500 last year can be explained by the five FAAMGs.

Exhibit 3: Performance of FAAMGs and S&P500 since 2009
Source: Goldman Sachs Global Investment Research

However, US technology stocks are nowhere near 1999 levels, valuation-wise. According to Goldman Sachs, over 50% of the rise in Big Tech since the Global Financial Crisis can be explained by rising earnings, so the steep rise in FAAMG prices can to a reasonable degree be explained by improving earnings.

I should add that valuing companies in the technology industry on a price-to-earnings basis, as you would do in most other industries, is a flawed approach, as many technology companies with great prospect do not yet generate any earnings. Valuing those companies on a price-to-revenues basis instead is therefore standard practice and, as you can see in Exhibit 4 below, US technology companies (represented by the Nasdaq 100 index in Exhibit 4) are nowhere near the valuation levels we saw just before the dotcom boom turned into a bust in 2000.

Exhibit 4: Ratio of market value to total revenues for Nasdaq 100 index
Source: Financial Times

Furthermore, fundamental conditions for many technology companies are quite simply outstanding at present. The combination of continued digitisation of society and the fallout from COVID-19 have resulted in extraordinary benign business conditions for the FAMMGs – a fact that, in my opinion, can easily justify the P/E multiple expansion the FAAMGs have enjoyed more recently.

Other investment opportunities

The Digital Revolution has thrown off quite a large number of investment opportunities away from the technology industry. On the short-side, the most obvious investment opportunities appear to be in fossil fuels and banks, but certain car manufacturers could also lose out. BMW, for example, seems to have been sleeping in class at a time where most of its competitors have been busy rolling out their ‘Tesla-beaters’. As mentioned earlier, longer term, insurance companies could also find themselves in a great deal of trouble.

That said, I would strongly advise you against going short fossil fuel futures. Futures markets are best left to the specialists who do nothing but that all day long. Although I expect fossil fuel prices to eventually go to zero, you could be wiped out long before then if you short, say, oil futures. If you want to short anything to do with fossil fuels, I would suggest you stick to equities.

As far as banks are concerned, there are two reasons why I think they could be in trouble. Firstly, as I said earlier, the blockchain technology could greatly reduce the need for banks and, within 20 years, they may be largely obsolete. Secondly, Amazon has already announced that it wants to use its mountain of cash to get involved in lending, and Amazon could do to banks what they have already done to brick and mortar retailers. In fact, Amazon has been relatively benign to brick and mortar retailers when compared to the damage it may inflict on banks.

On the long-side, the list of attractive investment opportunities is long. One not so well-known investment strategy that stands to benefit from the Digital Revolution are carbon credits. A carbon credit is a tradable certificate which represents the right to emit one tonne of carbon dioxide into the atmosphere. Carbon credits are part of international attempts to limit the growth in greenhouse gases.

As mentioned earlier, electrification of heating and transportation will drive the need for electricity much higher in the years to come. Those power plants that have already switched to wind and/or solar, will be able to sell their carbon credits, whilst those that continue to base their power generation on fossil fuels will have to buy them. This will establish an active market for carbon credits that investors can benefit from. The market already exists and is likely to present plenty of opportunities in the years to come.

Finally, let me wrap it up by mentioning my all-time favourite investment strategy for the more conservative investor – music royalties. It is a remarkably simple investment strategy and one that, unbeknown to many, stands to benefit from the Digital Revolution. Streaming, which is one of the many consequences of digitisation, has more or less destroyed CD sales; however, royalty incomes have almost doubled, as streaming generates plenty of royalties. Even better, as living standards continue to improve in many emerging economies, streaming will likely continue to grow for many years to come.

Final few words

The opportunity set is vast, but this paper is already longer than I would ideally like, so let me stop here. In the final part IV of this paper on the productivity conundrum, I will zoom in on what I believe will become the ultimate game changer, the switch to fusion energy, which will have enormous, and positive, implications for productivity.

Niels C. Jensen

16 March 2021

About the Author

Niels Clemen Jensen founded Absolute Return Partners in 2002 and is Chief Investment Officer. He has over 30 years of investment banking and investment management experience and is author of The Absolute Return Letter.

In 2018, Harriman House published The End of Indexing, Niels' first book.