Is AI good for growth?
The US grows faster than the UK and Europe. Does AI mean the gap will widen? It’s time to take The Sniff Test.
The New Elite
“Your margin is my opportunity” – Jeff Bezos
When Amazon’s boss and biggest shareholder threw down this challenge to the retail world, he captured the technology-led model of economic growth. If you rest on your laurels then startups will unseat you. Technological progress keeps the rest of us honest.
Amazon’s model was to strip the middleman from the supply chain. Products are sold to consumers via its marketplace, rather than going through distributors into stores and then to customers. Fewer links in the chain mean less cost. This is great for Amazon because lower prices and convenience allow it to dominate internet shopping. Is it good for the economy?
For centuries there was little challenge to economic power. Countries fought over empires, but back home the landowning families stayed in charge from one generation to the next. With little incentive to change, both economic and population growth were minimal.
Technology changed this. The Enlightenment and the Industrial Revolution unleashed forces that are still with us today. Byron Gilliam calls this “New things vs old things done better”.
The Industrial Revolution happened in a handful of countries. Without them, it is conceivable the rest of the world would still be in the medieval state of minimal growth. Instead, autocracies embrace new technologies to maintain their stranglehold on power. By and large, Western democracies have given up on the fantasy that they can impose change from the outside.
Autocracies rely on others for technology. When politicians tell us we should “follow China’s example” we must remember that China is playing catch-up. Access to the global trading system this century enabled hundreds of millions to be lifted out of poverty. Now President Xi demands a greater role for state-owned enterprises as a means to shore up his power. This will slow growth in China and beyond.
Technology alone does not generate growth. It is the sharing of technology, closer connections and the consequences of greater interactions that create growth. This is why when a new technology emerges, investors rush to buy the companies that create those connections. In the longer run, it is companies emerging to take advantage of the connections that become the new elite.
The Magnificent Seven
Stock market commentators talk about the Magnificent 7. These are the companies that dominate the rise in the market’s value. The seven are Alphabet (Google), Amazon, Apple, Meta Platforms (Facebook), Microsoft, Nvidia and Tesla. Tesla has fallen off the pace as enthusiasm for electric vehicles wanes. Apple has abandoned a decade-long project to build an EV in favour of investing in AI.
Nvidia’s share price has risen 200-fold over a decade. It charts the rise of the gaming industry and now the mass use of AI. It owns the building blocks of the revolution, but it is not the revolution. That is still to happen.
Nvidia is now the third largest stock in the US, behind only Microsoft and Apple. This reflects the excitement over AI and the way investing works on stock markets. There is a great game in getting ahead of where the flow of money is going next.
A few big funds will have to buy more Nvidia soon, because their rules say they must invest in the largest companies. They adjust their investments once in a while. The larger Nvidia becomes, the more the funds must buy. Other investors with different rules, buy as much Nvidia as they dare, expecting to sell it to the funds in a few weeks. This has nothing to do with the long term prospects for AI.
This is the conundrum of the stock market. In the long run it should reflect the performance of the economy. But the long run is made up of lots of short runs, in which the performance of the economy matters much less. When does the short term become the long term? No one knows for sure.
Slower Growth and China’s Rise
The US economy has grown at an average rate of 3.1% p.a. over 40 years. The chart shows that disruptions to trend caused by recessions are followed by a surging recovery. Then growth resumes its trend. The chart also shows that disruptions are getting larger.
US growth is faster than in both the UK and Europe. Yet growth this century and especially since the 2008 financial crisis, is below the long term trend. The last sustained period of above average growth was in the 1990s.
Economic growth depends on the expansion of the working age population, its productivity and the level of debt. A smaller workforce, as baby boomers retire, might be offset by more borrowing. This describes what has happened in recent years without being a causal explanation.
Workers’ productivity grows at around 2% a year. It accelerated in the mid 1990s through to the early 2000s. The share of the working population with jobs also rose at this time. This is important, as it often feared that more investment in machines puts people out of work. The opposite is probably true.
When investing leads to faster growth then more people are hired. This is why debt adds to growth when it generates investment. This is also the argument why AI may change the jobs we do, but won’t reduce their number.
The 1990s were a time of low inflation and interest rates. It was a period of globalisation and the US workforce grew due to immigration and training. It was also the dawn of the internet age, with automation, data analysis and improved communication among the technological leaps forward.
China joined the World Trade Organisation in 2001, which gave it access to the international trade system on favourable terms. Thereafter, large amounts of manufacturing shifted to China, both deliberately and due to price undercutting. Cheap Chinese goods helped keep global prices down. Nonetheless, US economic growth fell in this period. The US took note and would respond.
Why were the gains of the 1990s not sustained?
Redistribution vs Acceleration
New technology has several conflicting effects. Firstly, there is a surge in investment, as we are seeing in the semiconductor chips that Nvidia makes. Then there is displacement of workers as the technology takes over manual work. This phase in the development of AI has begun at tech companies but not yet more widely.
Lower labour costs mean higher company profits, which generates investment. In time this leads to new jobs. Displaced workers retrain and shift sector.
Technology also allows new business models. Some replace old ones while others are breakthroughs. Netflix is a way to distribute old media, such as films and TV shows. It put Blockbuster out of business and then eroded the value of network television. It moved demand for entertainment from older models to its own. This is similar to Amazon’s impact on retail.
YouTube created a new type of user generated content. This was radical, but its $30 billion of revenues are raised from advertising. There is new content but an old way of profiting from it. Hence much of the profits are transferred from other businesses. New technologies are more efficient and a spur to economic growth, but not by the total size of the new business.
Amazon’s rise is at the expense of other retailers and in particular those with physical shops. This has a knock on impact on property, which is another vast area of the economy. It’s conceivable that Amazon’s net effect is negative, at least until we figure out how to release the vast sums trapped in real estate.
Technology transforms the way we do things but its major economic consequence may be redistribution rather than acceleration of growth. Faster growth in the US may be at the expense of other parts of the world and Europe in particular.
Wildly Optimistic
What would Vodafone charge for a call without the option to use WhatsApp?
In the early stages of the internet boom, Vodafone was a stock market darling in the UK. I bought some options in the shares and lost all my money, because I did not understand short term influences on share prices. Longer-term investors speculated that Vodafone’s network would be worth a future fortune. They overestimated wildly because two things happened.
Firstly, governments decided they wanted a large slice of that network value and auctioned spectrum at crazy prices. Driven by fear of missing out, Vodafone and other phone companies mortgaged the future to access bandwidth. Once they controlled it, they expected to charge what they wanted. They might have done without the second development.
Companies emerged that used the bandwidth to do new things. The phone companies had a business hooking people up, but the bigger value is in what those people then did. iPhones are a pocket computer through which we communicate with people we never meet and give attention to the digital world. Attention is worth a fortune to advertisers. Price setting power went to the companies able to capture and contain attention, rather than the Vodafone network on which some of it happened.
Those new companies are now dominant. They are the Magnificent 7 and it’s hard to imagine a world they don’t rule. But somewhere a few people do envisage that world and are working to make it happen.
Pick a Side
John Authers argues that the US has grown faster than Europe because it is adjusting to the post-globalisation world. I’d argue it shapes that world because of its outsized economic impact. The US strategy is energy self-sufficiency and rebuilding industrial capacity. This allows it to grow faster than economies over-reliant on fractured global supply chains. One consequence of the war in Ukraine is that Europe now buys most of its gas from the US. More US growth with little change in Europe.
The strategy works only as long as other countries don’t adopt the same me-first tactics. China already has and the European Union is closed club at heart. The UK was never comfortable in this continental club and left it hoping to be part of a global trading community. Instead it must choose sides, as almost every smaller nation does.
The right side to be on is the one with the best AI. What might that be worth?
The Value of AI
Analysts at NewStreet Research estimate the market for AI chips was worth $23 billion in 2023. Lisa Su, the CEO of chip manufacturer AMD, believes the market will grow to $400 billion by 2027. That’s about 70% growth a year.
AI is much more than semiconductors. Memory, storage, network cards, data centres, power supplies and cooling are all required. Azeem Azhar thinks this may double the market to $800bn. For context, the total US economy will be around $28 trillion this year. AI chips are less than 3% of that.
The biggest overall benefit from AI will be to total company profits. Firms that don’t adopt it will be put out of business. This applies the world over, not just in the US. It’s not clear that this will change the trend rate of global growth by much.
The biggest individual beneficiaries of AI will be a handful of companies. These will dominate the stock market much more than they do the economy. Microsoft and Nvidia may not be the long term winners, but they are the best bets we have today. Investors will over estimate their fortunes until it is clearer who will take them away.
The chances are these unknown future giants will be from the US. UK and European companies wait to see who the next Jeff Bezos coming for their margins may be.