An Important Point About Increasing Returns
Today: W. Brian Arthur, my perspective on increasing and diminishing returns, industrial policy, Singapore 🇸🇬
The Agenda 👇
The life and works of W. Brian Arthur
A frequent mistake regarding increasing returns
A few additional ideas on capitalism
A primer on industrial policy
A great article about Singapore
Today, I’d like to dedicate a brief essay to economist and complexity theorist W. Brian Arthur and his pioneering research on increasing returns.
Here’s Brian’s short bio: he grew up Catholic in Northern Ireland, did a PhD thesis in operations research at the University of Michigan, worked at McKinsey (I featured his thoughts about the firm in my 11 Notes on McKinsey), specialized in modeling demographics in developing countries, then (at age 37) became the youngest endowed chair holder at Stanford University.
In the second, more consequential part of his career, he contributed to the emergence of the Santa Fe Institute, now the most renowned research organization in the world when it comes to complexity theory.
I briefly met Brian when visiting my friends at CASBS in Palo Alto back in February; he kindly offered to meet for lunch but alas we couldn’t find a mutually convenient date, and a few days later I was back on a plane to Europe. Now, with the pandemic and all, I don’t know when I’ll be back in California or if I’ll have the opportunity again 😢
Brian has become a well-known figure in the tech world. If his name sounds familiar, it’s likely you’ve discovered him through one the following channels:
Maybe you’re interested in complexity theory in terms of the history of science and you’ve read M. Mitchell Waldrop’s Complexity(which features Brian prominently—in 1992).
Maybe you’re a long-time admirer of complexity economics and you’ve read Brian’s academic writings on the matter, such as Increasing Returns and Path Dependence in the Economy (1994).
Most likely you’ve read Brian’s seminal 1996 article Increasing Returns and the New World of Business in Harvard Business Review (here’s the background story).
Or, more recently, you could have listened to Brian’s discussion with Sonal Chokshi and Marc Andreessen on the a16z podcast (or read one of Brian’s articles in McKinsey Quarterly).
As for me, I first discovered the 1996 article through Bill Gurley (2014), then bought Brian’s books as well as Waldrop’s, then stumbled upon articles and podcasts published more recently.
So, what about today’s ‘Important Point’? Many people who read the article in HBR come out of the experience having quite the binary view of the economy:
On one hand there’s the old industrial economy, that of the age of the automobile and mass production, and it’s marked by diminishing returns to scale.
On the other hand, there’s the new digital economy, that of ubiquitous computing and networks, and it’s all about increasing returns (or, as the Bible puts it, “To them that hath shall be given”).
However, that’s too simplistic, and I’ve realized there’s a different way of telling that story. It requires focusing on the various stages of development in both techno-economic paradigms:
Fordist-Age companies seem subject to diminishing returns, but that’s because those that are still around these days have reached an advanced life stage and as a result are nearing exhaustion.
Meanwhile, today’s tech companies strike us as being driven by increasing returns because for most of them, it’s still Day One. They’re still in their infancy!
Indeed, the whole point of capitalism is the pursuit of increasing returns. And the fact that the Industrial Revolution and the following technological revolutions have all been such a boon for capitalism is because each contributed better ways of generating increasing returns and arriving at an ever larger scale.
Corporate giants in the age of the automobile and mass production grew larger than those in the age of steel and heavy engineering, which themselves had grown larger than those in the age of steam and railways. And today, we already realize that the largest tech companies will be larger than the manufacturing giants that dominated the global economy in the 20th century.
A few additional ideas to illustrate that point:
When large players reach the stage where returns start to diminish rather than increase, that’s when you’d expect new entrants to be able to seize the opportunity to tackle incumbents. But that’s not what happens, for a very simple reason: returns that start to diminish are correlated with having a high level of defensibility. As an example, returns stopped increasing a long time ago in the car industry, still who would be crazy enough to launch a new car manufacturer (assuming, that is, that you’re not being lifted up by technological change in batteries)?
In his podcast conversation with Brian, Marc Andreessen quotes Peter Thiel and makes the case that “if you don’t have increasing returns, you’re on a long-term downward slide to commodity”. If you’ve read Walter Kiechel’s The Lords of Strategy (highly recommended!), this is precisely where strategy starts to become relevant. In the presence of increasing returns, it’s anything goes, because you want to detach and race ahead so as to secure lock-in. But when returns eventually start to diminish, that’s when you need to focus on strategy.
CEOs from the old paradigm seem clueless in the age of computing and networks (take John Sculley, the former CEO of PepsiCo, almost leading Apple into the ground). But that’s because they only know the late stage: their main skill as a corporate CEO is about managing a large organization cursed with diminishing returns to scale. So when you put them in charge of a smaller/younger company that still enjoys increasing returns, things go wrong quickly.
In a Stratechery update last year, Ben Thompson observed that “silicon-based chips have similar characteristics [as software]; there are massive up-front costs to develop and build a working chip, but once built additional chips can be manufactured for basically nothing”. I found this observation odd at the time, and I still do, because for me building silicon chips is a manufacturing business, and the economics pointed out by Ben are characteristics of manufacturing...in its early stages of development.
If only to reinforce the point, doesn’t it strike you that all of today’s tech giants are reaching the point where returns are finally starting to diminish? See Apple and its problems with the App Store (and manufacturing in China!); Amazon failing to manage trust on its marketplace, notably because of fraud in user-generated reviews and the bad quality of products sold; Google being unable to preserve its culture and finally becoming an ‘evil’ company; and Facebook, well, being vilified for the divisive impact of their core product.
So here’s my ‘Important Point’: don’t misread Brian’s work and think that we’re living in a completely different world now. It’s something else: because it’s still early in the Entrepreneurial Age, most dominant companies today are enjoying increasing returns similar to (even better than) those that leading car manufacturers enjoyed in the first half of the 20th century. This is why they succeed at doing capitalism, but this doesn’t mean that it will last forever—or that increasing returns are a feature that’s perennially inherent to the Entrepreneurial Age.
Also (as I’m re-reading Brian’s article and am again struck by his harsh criticism of legacy economists), remember Mary O’Sullivan’s illuminating observation:
If you want to understand capitalism, talk to capitalists, not economists.
Indeed, as revealed by Fernand Braudel, capitalists, unlike economists, have always known it’s all about increasing returns—long before the rise of computing and networks. And in the words of Brian himself:
Mechanisms of increasing returns exist alongside those of diminishing returns in all industries.
What do you think?
You might have read about Dominic Cummings, a senior advisor to Britain’s Boris Johnson, wanting to build trillion-dollar tech companies in the UK (and being willing to botch Brexit negotiations in the process). At the same time, Macron is schmoozing tech entrepreneurs in Paris, promising them billions of euros to be allocated “artificial intelligence, cybersecurity and quantum computing” 🤔
And so I thought I would again highlight what industrial policy is all about. Here’s a bullet-point primer:
Industrial policy should be designed differently if you’re a leader or a follower. A leader is the most advanced in terms of economic development and is at the frontier from a technology perspective. Being a follower means trying to catch up to that.
A country that’s the leader (like the US these days) should implement a policy that’s all about waste, trials and errors, and cutting-edge technology. It only helps to welcome immigrants (both high-skilled and low-skilled).
A follower, on the other hand, should be focused on catching up. That requires active management of the workforce (keep workers where they are, and make it rewarding), supporting emerging champions with proven increasing returns, and financial repression.
A frequent mistake is being a follower and implementing an industrial policy that’s only fit for a leader. That’s what both Cummings and Macron have in mind, with their obsession over basic research and high tech. (Meanwhile the UK let ARM slip away, sold to Nvidia.)
In both cases (leading and following), the government needs to provide a direction, usually using the terms of waging a war. If you’re the leader, it might be an actual war (as with WWII, which gave birth to Silicon Valley). If you’re a follower, it’s simpler: just wage a war against under-development—like Taiwan, South Korea, and Mainland China successfully did in the second half of the 20th century!
But this requires something painful for countries like France and the UK: admitting that they’re lagging behind from an economic development perspective. I wrote about this here: Europe Is a Developing Economy.
🇸🇬 An excellent article about Singapore succeeding as a follower made the rounds recently (I discovered it in Margins). I wrote about languages in Singapore in a previous edition of this newsletter, but this really explores all the dimensions of what a successful industrial policy is all about: The True Story of Lee Kuan Yew’s Singapore (Haonan Li & Victor Yaw, Palladium, August 2020).
If you’ve been forwarded this paid edition of European Straits, you should subscribe so as not to miss the next ones.
From Normandy, France 🇫🇷
Nicolas