What if the Single Most Important Idea in Economics No Longer Works? by@jamestplunkett

What if the Single Most Important Idea in Economics No Longer Works?

In 1714, the Anglo-Dutch social philosopher Bernard Mandeville published a poem called the Fable of the Bees. The poem told the story of a bee colony that underwent a conversion: All the bees became suddenly selfless, putting their vices and self-interest to one side to pursue the wider good of the hive. The result was a disaster. Without the drive of private ambition, the hive broke down. Nectar collection plummeted. Honey production stalled.[1] It turned out that a selfless society was, paradoxically, worse for social welfare than one in which free individuals chased their own visions of a good life. Mandeville’s idea might have faded into obscurity but for a young economist called Adam Smith. He incorporated the essence of the poem’s argument into his book, The Wealth of Nations.[2] In Smith’s formulation, it was bakers, not bees, that provided the organising image: It is not from the benevolence of the butcher, the brewer, or the baker, that we expect our dinner, but from their regard to their own interest. Thanks in large part to Smith, this idea — that markets align the interests of producers of consumers — went on to become one of the most important in all of economics, framing how we conceive of the relationship between the market and the state. Yet in the last few years, as the full scale of the digital revolution has made itself felt, this old mental model has begun to look shaky. Does a market economy still foster prosperity when the bees are all on Facebook? And, if not, where would that leave us? Finding our Feet With all this talk of the metaverse, I thought it might be interesting to explore these questions in a little more depth. What was it, precisely, that the bees and the baker were originally intended to signify? What is it about a digital economy that makes this way of thinking less valid or useful? And what could this mean for a conception of the state that rests so heavily on this intellectual foundation? We’ll be talking a lot about bakers, so let’s start by being clear about the significance of this image.
James Plunkett HackerNoon profile picture

James Plunkett

Author of End State: https://amzn.to/3ybEsGd I write about how we govern a digital economy

linkedin social icontwitter social icon

Photo by Tina Rataj-Berard on Unsplash

The Fable of the Bees, if the Bees Were all on Facebook


In 1714, the Anglo-Dutch social philosopher Bernard Mandeville published a poem called the Fable of the Bees.

The poem told the story of a bee colony that underwent a conversion: All the bees became suddenly selfless, putting their vices and self-interest to one side to pursue the wider good of the hive.

The result was a disaster.

Without the drive of private ambition, the hive broke down. Nectar collection plummeted. Honey production stalled.[1] It turned out that a selfless society was, paradoxically, worse for social welfare than one in which free individuals chased their own visions of a good life.

Mandeville’s idea might have faded into obscurity but for a young economist called Adam Smith. He incorporated the essence of the poem’s argument into his book, The Wealth of Nations.[2]

In Smith’s formulation, it was bakers, not bees, that provided the organising image:

It is not from the benevolence of the butcher, the brewer, or the baker, that we expect our dinner, but from their regard to their own interest.

Thanks in large part to Smith, this idea — that markets align the interests of producers of consumers — went on to become one of the most important in all of economics, framing how we conceive of the relationship between the market and the state.

Yet in the last few years, as the full scale of the digital revolution has made itself felt, this old mental model has begun to look shaky. Does a market economy still foster prosperity when the bees are all on Facebook? And, if not, where would that leave us?

Finding our Feet

With all this talk of the metaverse, I thought it might be interesting to explore these questions in a little more depth. What was it, precisely, that the bees and the baker were originally intended to signify?

What is it about a digital economy that makes this way of thinking less valid or useful? And what could this mean for a conception of the state that rests so heavily on this intellectual foundation?

We’ll be talking a lot about bakers, so let’s start by being clear about the significance of this image.

Smith is, after all, one of history’s most misused economists, straw-manned by the left and worshipped by the right for saying things he didn’t say. So what was his point?

The phenomenon Smith had his eye on is the mysterious order that seems to emerge from free exchange in a market. No-one tells bakers what to do, or helps them coordinate with their suppliers, or forecasts how many loaves of bread they should make. And yet bakers get up damned early, and work damned hard, and make a decent fist of judging the baguettes and the braids and the rounds.

Smith’s explanation for these outcomes is that the market acts as a coordinating mechanism. As prices are set and as consumers make choices, this reveals underlying information about what people want, and it incentivises other people to meet those wants.

The market thereby aggregates self-interested preferences into an output that looks, on the face of it, selfless. It’s a form of emergent intelligence that led Smith to his most famous metaphor: the invisible hand.

What Smith wasn’t claiming, contrary to what many on the left and right like to think, is that markets are perfect, or that they can self-regulate, or that there’s no need for the state. As Jesse Norman shows in his biography of Smith, the idea of the invisible hand is both narrower and more contingent than it’s often claimed to be.

Smith saw, for example, that markets can create harmful byproducts. Our baker might, in their blinkered focus on bread, throw their litter into the street. Smith also acknowledged the risk of collusion, warning that, whenever company bosses meet, the conversation is likely to end “in a conspiracy against the public”.

Smith even saw that people are prone to irrational behaviour, a strand of thinking that later matured into behavioural economics. This all means there’s a role for the state: banning or taxing ‘bads’ like pollution, breaking up monopolies and cartels, and protecting people from exploitation.

Behind all of this, however, lies that central conviction: that a baker’s ambition is aligned to a customer’s appetite. The baker’s route to a good life is to make the customer’s life good. And it’s because of this lucky alignment that markets — if we can contain all those byproducts — lead to a prosperous social order and drive progress over time.

An Uneasy age

In the 245 years since the Wealth of Nations was published, Smith’s idea spread out under our feet like a giant tree pushing out roots.

Alongside the work of economists like Ricardo, it began by defeating the tenets of mercantilism, helping to usher in a period of economic liberalism. It went on to defeat central planning, making a mockery of the idea that any government could hope to synthesise the coordinating power of the market.

In the 20th century, as politics in the west coalesced around the social democratic mixed economy, Smith’s baker became the lens through which the state-market relationship was viewed.

Left and right continued to argue over the contours of that relationship, and economists refined concepts like ‘externality’ and ‘oligopoly’ to deal with all those so-called ‘market failures’, but still, deep down there was the premise of alignment, and so the state’s role was defined in the negative — to intervene by exception.

It would be arbitrary to say that ‘everything changed in year X’, but it does feel fair to say that, since around 2010, there’s been snowballing interest in the extent to which digital markets abide by old rules.

I first wrote about these developments in a 2018 essay about big data and dark nudges, while others have explored these ideas in much greater depth. Books like The Age of Surveillance Capitalism by Shoshana Zuboff have described the character of a digital economy and writers like Anna Weiner (Uncanny Valley) have captured the queasy vibe of what is sometimes called ‘late capitalism’.

host of new programmes and institutes have been founded at leading universities to explore how we can update economics for a digital world, while my own book, End State, explores how we govern this new phase of capitalism, sketching the contours of a new kind of state.

What a lot of these debates come down to is a growing sense that something deep has gone wrong with markets as we thought we knew them. It’s a feeling that capitalism is no longer just misfiring in the familiar ways — producing harmful byproducts, for example — but deeper down, in the belly of the thing. It feels increasingly clear that even if we dealt with all those off-shoots and externalities, this wouldn’t be enough. That something is misaligned.

In keeping with our uneasy age, the signs of this misalignment can be hard to place. At a micro level, they often show up in our lives as no more than frustrations. We check our bank account and find a subscription we don’t remember signing up to, or a bill comes out higher than we thought we’d agreed. In general, we feel a nagging suspicion that companies are trying harder to trick us than to impress us — that, as consumers, we have to be ever more on guard.

At a macro level, the signs are no less intangible, and they sum to the mood of our times. There’s the way we doom scroll as apps draw us irresistibly in; or the way we share our data, knowing we have no real choice; or the frazzle of burnout as our weeks fly past in a blur of emails and Zooms, even while productivity disappoints.

And, perhaps most tellingly, there’s the sense of unease when we see capitalism’s latest gift, Zuckerberg’s metaverse. It all feels less like lucky alignment and more like grim inevitability — a dystopian future that we don’t really want and yet that, with each scroll of the finger, we’re beckoning on.

It’s this image — that we’re somehow choosing to beckon dystopia — that most makes me want to check back to that Smithian small print. Weren’t markets supposed to satisfy our needs? Whatever happened to the baker? They’re questions that go to the heart of how digital capitalism functions, so it’s useful to start by following the money.

What’s Changed?

It was barely two years ago that Mark Zuckerberg sat before Congress wearing his best ‘I didn’t eat the cookies’ face. He promised that Facebook would become a more responsible corporate citizen, dropping the move fast and break things attitude of their early years. The humility didn’t last long. In late 2021, the company rebranded itself Meta as Zuckerberg set out his plans to replace, you know, all of lived reality.

It’s funny to think now, but back in the early years of Big Tech’s rise, when we were still young and naive, we took these companies at face value. Facebook was a social network and Google was a search engine. They did what they said on the tin.

What we know now, of course, is that the flagship products run by these companies are really just a means to an end — a step toward the thing they’re actually selling.

It’s an idea that was first captured by a neat turn of phrase:

If something is free, you’re not the customer — you’re the product.[3]

Later, as we came to understand digital platforms better, this idea was refined. As the technology philosopher Jaron Lanier has said, the product isn’t so much ‘you’ as: ‘the ability to change the behaviour of millions of people like you’.

The search engine or social media platform is little more than bait, there to hold our attention long enough to capture our likeness in data. And so, when we search or scroll, this isn’t the endpoint of a production process; it’s an intermediate manufacturing step. It’s like whisky being distilled or bread being proved, except in this case it’s us being processed.

To see how this all relates to the baker, we need to rewind to Smith’s time and stand his version of capitalism alongside this emerging form of capitalism in the early days of the metaverse.

Imagine you’re in London in 1776; a horse trots by on the cobbles, a person trundles past with a wagon, and a baker opens up shop, the bell dinging as the smell of fresh bread mingles with the sour city air.

If you’d stood and watched that baker work, I don’t think it would be too hard, whatever your politics, to empathise with Smith. The early starts, the tireless effort to perfect the crust, the customers queuing outside because they’d eaten the bread and liked it, or because they’d heard it was good from a friend.

And if you’d stepped back to see the whole chain of events that led to that bread — the miller dropping off flour, the glazer making the shop window, the candlemaker providing light — you could, I suspect, appreciate why some people see in markets a sense of magic. The whole thing seems so uncoordinated and yet so coordinated. You can sense the invisible hand.

Since Smith’s time, there have been, shall we say, some hitches to that rosy view, many related to the concept of externalities. Most disastrously, of course, in the scramble for profit, humanity poured vast quantities of carbon into the air. And there have been a few minor arguments about the other side of production too: what happens to workers when companies compete to make cheap products, and what should be done about that.

When it comes to our focus here, however — the central question of alignment — you could argue that the first really material development was the emergence of mass market advertising.

In late 19th century America, capitalism entered a new and more powerful phase as the innovations of interchangeable parts and repeatable assembly processes ushered in the age of mass production. It was suddenly possible — and explosively profitable — to sell identikit products at massive scale.

Around the same time, modern mass media took off, making it affordable to reach millions of people with identical messages via print, radio, and, later, TV.

Together, these developments transformed the economics of production and they also did something intriguing to that deep alignment of interests. It now made sense not just to make good bread but to spend lots of money and effort telling people how good your bread was — and suggesting that they might also like some jam. Soon bread became Wonder Bread or Hovis, and, while some of the baking was of dubious quality, advertisers assured us that two out of three people thought it was the crunchiest bread around.

It’s still something of a mystery why advertising didn’t inspire more intellectual angst, particularly on the right. The rise of consumer brands weakened free markets at the knees, systematically entrenching big producers who could afford to advertise. Meanwhile, ubiquitous advertising undermined a core economic assumption: the idea that consumer preferences are fixed and underlying, so that free exchange in a market does nothing more than surface a person’s preferences, ready to be scratched like a longstanding itch.

In the new world of advertising, markets didn’t so much surface our preferences as create them, and there’s a case for saying this is why the 20th century ended not with Keynes’ sunny forecast for his grandchildren — a world in which we gradually met our material needs through rising productivity, and eased back to a 15 hour work week — but with victory for Galbraith’s gloomier prophecy — a world in which we manufactured new wants as fast as we could meet old ones, leaving us all running on the hamster wheel.

Still, let’s park that for now. Because what’s been happening since the 2010s seems to be even more significant, and more destabilising of those deep Smithian ideals.

With the rise of the digital economy — as described so well in Azeem Azhar’s book Exponential — four things are changing that alter the logic of markets more profoundly than mass production.

First, markets are reaching a scale so vast it makes the Model T look like a niche product. We hear stats about Facebook having 2.9 billion active users or Google having seven products with more than a billion users and the numbers are so big we can’t comprehend them. It’s a shift of scale so significant that, like the transition from quantum mechanics to general relativity in physics, it changes the rules of the game.[4]

Second, an economy of products has been replaced by an economy of services, as physical stuff — things you can stub your toe on, to use Azhar’s lovely phrase — has declined as a share of GDP. In 1948, services made up 46% of the UK economy. By the mid-2010s this had risen to 79%. This has also changed the face of work, as the percentage of workers employed in services rose from 33% in 1841 to 80% today.

Third, services themselves are changing, as in-person services like haircuts decline relative to digital services like music streaming; things we can click, swipe, or scroll. Digital services mean companies can now scale at near-zero cost and they also favour new payment models like subscriptions, rearchitecting the way we make choices.

Fourth, the world has been networked, so that fewer of our decisions are made in isolation and more economic phenomena have the qualities of a virus. And, since these viruses are digital, their vectors of transmission and strength of contagion are defined by algorithms, with these algorithms in turn being designed by profit-seeking technology platforms.

Each of these developments would, on its own, have been a big deal but together they justify that tech sector cliche: this changes everything. In effect, they swap one type of economy for another. We used to buy bread, now we sell software services at scale. And it’s this shift that makes the baker less apt as a model.

What do we Need From a Theory?

In his 1966 essay, The Methodology of Positive Economics, Milton Friedman explored what we should and shouldn’t expect from an economic theory. For Friedman:

…the relevant question to ask about the “assumptions” of a theory is not whether they are descriptively “realistic,” for they never are, but whether they are sufficiently good approximations for the purpose in hand.

An economic theory doesn’t need to be accurate, it just needs to be close enough to be useful. We need to be confident that the details we lose in the process of abstraction aren’t pertinent to the matter at hand.

So how about the baker and the bees? Of course they’re not perfectly realistic, but are they close enough? If we treat them as decent low-fidelity guides to the task at hand, will things go well?

Here are some observations about life in 2021 that make me feel like the answer to these questions is no.

  • Some of the world’s most valuable companies and the richest people in history make their money not by selling great products to consumers but by selling to other companies a meta-product: the ability to change people’s behaviour. This isn’t a side hustle, it’s their business, and so the ability to alter consumer behaviour is what they optimise for.[5]

  • The incentive to get better at these techniques of behaviour modification is evidently really strong. The companies themselves are relentless in their pursuit of this goal, throwing the full weight of their institutions — the most sophisticated in human history — behind it. Meanwhile the system itself (i.e. capitalism) is doing likewise, so that the world’s most responsive investor dollars and finest minds are being sucked into this area of innovation like prospectors to a fresh seam of gold.

  • It’s working. So well, in fact, that in little more than a decade it’s changed the character of our lives and societies. Already (and remember we’re still in the opening credits of this thing) these companies have grabbed so much of our attention that the average person in America spends nearly a quarter of their waking lives on social media. And the same companies now have so much of our data that they literally know things about us that we don’t yet know ourselves.

  • Governments seem baffled, like an adult who’s forgotten how to walk. Even ideologically liberal governments are having to intervene in markets in ways they’d rather not, while previously established terms of engagement no longer seem to work. Debates we thought we’d closed — should speech be regulated? — are being reopened and worms are spilling everywhere.

  • The whole dynamic has a totalising air to it, some of which has already manifested but most of which is still to come. Already we can’t really escape from the dominant tech platforms while still playing a full role in society, so we have no choice but to engage daily with services that we know are designed to be addictive. And as to the future, well, it’s all a bit ominous. It seems increasingly likely that the moonshot will succeed, and that, through a blend of augmented and virtual reality, most of our interactions with the physical world will ultimately be mediated by a privately-configured digital layer. a.k.a. The metaverse is coming. (If you want to peer through the looking glass, read this and don’t stop until you get to compulsion loops and the use cases for $BANANA).

  • Now I don’t know about you, but when I sit with these facts for a minute I can’t help but feel that (a) things aren’t going very well and (b) bakers and bees are no longer all that reassuring a guide.

New Metaphors, new Models

A couple of years ago, my partner and I moved onto a narrowboat. Or I should say the shell of a narrowboat because when we moved onboard the boat had no power or water and the plan was to live there while we renovated the space into a warm and comfortable home.

This was, needless to say, a mistake. And one day, as part of the process by which this dawned on us, our generator broke down. At first, when you started the generator up, its motor seemed to run OK. But as the revs climbed higher the motor quickly jammed.

This being the third generator we’d bought — the first two were stolen — we had little choice but to fix it, so one Saturday afternoon we sat on the towpath and stripped the motor back to its parts. Once we had the machine in pieces, we saw the problem: the piston seemed ever so slightly off-center. Only barely, but enough to be decisive at high speed.

I think about this motor quite a lot, and not just because the whole experience was so memorable, but also because the motor feels like such an apt metaphor for how markets are fairing in the transition to digital capitalism.

If we think back to Smith’s time, maybe the motor of capitalism was still misaligned, but markets moved slowly enough — or were low powered enough — that, save for the exhaust fumes coming out the side, the whole thing just about functioned OK.[7]

Maybe the baker wasn’t quite incentivised to make crunchy bread, but to do something similar — something like: ‘design a bread-buying experience to maximise lifetime profit’. But maybe, in the technological conditions of the time — small markets in which customers made daily independent decisions and bought physical stuff — that nuance didn’t matter. The baker’s best bet was simply to make great bread. And so we could act, individually and collectively, as if this was the case.

But what about now? What if the motor has sped up, as capitalism has become more sophisticated, so that other ways to optimise profit are possible, and not just possible but way more profitable than making great bread? Maybe at this point the misalignment is decisive, meaning that it’s no longer a fact we can afford to downplay or ignore. And maybe the fact that we are currently ignoring this misalignment — by governing markets with a policy settlement that is premised on alignment — explains why we can all smell smoke.

Addressing a Counterargument

I’m keen to get onto the policy implications of all this but first I want to address a counterargument. (If you’re fully on board with what I’ve suggested so far, feel free to skip to the next section.)

I suspect some people will buy the general thrust of my argument but will say I’ve massively overstated the case. Sure, some parts of the economy now work in new ways as a result of digital technology. But doesn’t the vast majority of economic activity still look more like the baker? Or at least close enough to the baker that the Smithian mental model can still serve as a premise for the relationship between the market and the state.

My instinct here is to buy the first part of this counterargument but not the second. i.e. I agree that, as of today, most of the economy doesn’t function with the logic of a digital market. But I don’t think it follows that we can stick with our old mental models.

To unpack why I think this, let’s try to give as balanced an account as we can of what markets look like today.

If you step back to think about today’s economy, it does seem fair to say that most markets don’t yet look like digital markets — an idea that I’m going to abbreviate as SSaS, or software services at scale.

For starters, we still buy lots of stuff — from bricks to books to, well, bread — and, byproducts aside, these markets still seem to exhibit a strong degree of alignment; i.e. companies work really hard to make us nice things.

Then there are all those companies that sit far away from the frontier of technological change. The many local businesses, for example, in which people work hard every day to make us lovely handmade products, or to deliver charming personal services.

So I buy the point that, in many markets today, a strong form of that serendipitous alignment still seems to operate. But what I don’t buy is that we can therefore relax.

The main reason I say this is that, even if the majority of economic activity today doesn’t have the qualities of SSaS, this is nonetheless clearly the direction of travel — and there are hugely powerful economic incentives to move us ever quicker in this direction.

You can see this if you think about how hard companies are trying to turn every transaction we have with them into a software-selling experience.

Think about when you buy something on Amazon for example. How often are you now nudged to ‘subscribe’ to a regular purchase of the toilet rolls or carrots? That’s Amazon trying to ‘service-ise’ a product, giving the interaction some of the qualities of a digital service (e.g. a recurring payment).

Or think about how products like mobile phones are often given away for ‘free’ with (i.e. paid for through) an accompanying service contract, again turning products into services. Or consider the boom in leasehold cars, which in essence does the same thing.

Or look at how companies from John Deer to Sony are questioning what it even means to ‘own’ something in a digital world. Both companies have been in legal disputes with their customers after trying to stop them from repairing or modifying the tractors/PS4s they had bought and thought they owned. It turns out the companies own the software that makes the hardware function, which makes it hard to know if there’s even a ‘thing’ there to be owned.

Or, even more fundamentally, think about changes to what we might call the infrastructure layer of the consumer economy. The no-checkout shops, which work by modelling an in-person shopping experience in data, essentially digitising a physical experience. Or the rise of buy-now-pay-later payment options, which turn a one-off payment into an enduring financial relationship. Or, at the highest level of abstraction, the metaverse-shot: the attempt to create that digital mediating layer, so that life itself becomes software.

When you step back to think of this diverse set of examples, you might say that this process of conversion —of trying to turn bread-buying into software selling — is one of the main economic forces at work in the world today.

So it’s clear where we’re headed, even if we’re not there yet. And I want to go further than this by borrowing an argument from Peter Drucker. This is to say that, precisely because the direction of travel is so clear and inevitable, SSaS is not only the dominant economic and social logic of the future but also the dominant logic of the present.

Sure, SSaS isn’t yet in the mathematical majority, but when it comes to determining an era’s decisive economic and social logic, what matters isn’t maths, it’s momentum. It’s already clear that the decisive dynamic of our times isn’t bread-buying but software selling, and so our governing institutions need to function on that basis.

And let’s end this section by reminding ourselves of where we started: none of these behaviours are part of a conspiracy, or some kind of planned initiative. It’s just the social order that is emerging spontaneously from the self-interested behaviour of free individuals in a digital economy. It’s the 21st century invisible hand.

The Task and why it’s Hard

So let’s talk finally about policy implications. And let’s start by pinning down the task at hand. What specifically is it that we need to do from a public policy perspective? And how does this differ from what we used to need to do?

In the 20th century, we needed to build a regulatory settlement for industrial conditions based on the close-enough premise that producer interests aligned to consumer interests. The state could therefore focus on byproducts/market failures (summed up by the three c’s of carbon, cartels, and cons).

In the 21st century, we need to do something that looks, on the face of it, similar but that is in fact profoundly different: we need to build a policy settlement under the technological conditions of digital capitalism based on the premise that producer interests are meaningfully misaligned from consumer interests (while also doubling down on the ongoing task of tackling classical externalities).

This new task isn’t just different to the old one, it also seems to be meaningfully more difficult for a couple of reasons. Both of which are, I think, capitalisable Really Big Deals.

First, think about where this new problem is situated with respect to the logic of capitalism. What we’re talking here about aren’t byproducts, like an exhaust that vents off the side. What we’re talking about is a misalignment of the market itself.

Contrast this with pollution as a classic problem of the old type. When you’re trying to stop a company from polluting, there’s one big thing in your favour: the company doesn’t literally want to pollute. No CEO wakes up rubbing her hands at the thought of filling a river with sludge; it’s a byproduct of the chase for profit. So while companies of course resist attempts to stop them polluting — because it’s cheap and therefore profitable — when we muster the courage to stop them (by banning CFCs or leaded petrol, for example) the market responds relatively well.[7] The hard part, as we’re seeing with climate change, is mustering the courage to act in the first place.

Contrast this with a misalignment problem, such as when a firm exploits our behavioural biases with a contract that quietly auto-renews. In this case, we need to stop the company from doing precisely the thing it most wants to do. The problem isn’t a byproduct, the problem is the product itself.

We only have to look at the UK insurance industry to see how tricky this all is. When the UK regulator, the FCA, tried to stop firms from luring people to unintentionally auto-renew their contacts, the industry did everything it could to wriggle past the new rules, pulling the FCA into an absurd cat and mouse game in which the regulator had to set ever more detailed rules, down to specifying the design of auto-renewal letters.[8]

It’s easy to see why the same thing will happen whenever we try to regulate misalignment. Imagine, for example, trying to force Facebook to make their algorithm less attention-grabbing, or forcing Google to capture less data. The companies would likewise do everything possible to evade the new rules, and I’d wager they’d win.

So that’s issue one: the locus of the problem is really close to the burning heart of capitalism — much closer than externalities are — and that’s a hard, even dangerous, place for the state to be.

The second issue relates to the nature of the information revolution, the defining feature of which is, of course, that its constitutive material is intangible, not physical.

What this means is that the social fallout from the digital revolution is also mostly intangible — it plays out in information and ideas.

Picture the fallout from the Industrial Revolution as a contrast. It was a steaming pile of sewage, sickness, and soot. Now picture the problems of our digital age. It’s a cloud of mental illness, burnout, conspiracy theories, and distrust.

These are harder materials for the state to work with and, as a result, public policy is being pulled into territory where it’s (rightly) reluctant to go.

Think again of the contrast with an old problem like pollution (not old, of course, in the sense that it’s solved, but old in the sense that it emerged in the industrial age). A problem like pollution is, it’s clear, difficult to solve, and even more so when the pollution floats across borders. But at least it’s tractable. If you want to ban a pollutant, you can specify the chemical in law. Or consider another set of problems that emerged in the twentieth century: the damage to human health caused by smoking, drinking, or obesity. Again, when it comes to cigarettes, alcohol, and sugar, you can tax them or limit their sale.

But what do you do when society is being harmed not by physical toxins, but by toxic ideas? When companies aren’t tempting us with sugar but with a news story that satisfies our prior beliefs? Or when an algorithm radicalises people by keeping them gripped with ever more extreme content, as some studies suggest YouTube’s algorithm might do. Or what if we’re being tricked with weasel words that exploit our inertia? Or what if — because remember how viral content thrives in a networked world — we’re being fed a conspiracy theory so juicy we can’t wait to tell our friends?

It’s not a coincidence that these are precisely the issues that have governments across the west feeling all clammy, while the Chinese government looks on, glowing with self-confidence.

And it’s from this global perspective that the whole thing starts to feel a bit existential. Because of course other social settlements are available.

Remember that quaint old idea that free markets and free societies go together nicely? That capitalism and democracy each encourage the spread of the other, because their logics are so similar?

Well what if that was true for industrial capitalism, but isn’t true for digital capitalism? What if, in fact, digital markets are on some level incompatible with democratic institutions, or at least in tension with these institutions as they’re currently designed? And what if it’s another model — something closer to the digital authoritarianism of China — that clicks together neatly with digital capitalism? What if this mess we’re in doesn’t just signal the end of a phase of capitalism, but of democracy too?

Three Policy Solutions

Ok, so I can feel a sense of vertigo coming on, so let’s grab hold of some policy solutions. Because at the end of the day I’m an optimist on all of this. My basic view is that this is saveable, and in fact more than saveable; I think there’s reason to believe that we can make a digital economy not just richer but fairer and greener than anything we’ve seen before.

In fact, this is precisely the twist concealed in the title of my book, End State. I’m not arguing that we’ve reached the end. I’m arguing that we’ve reached an ending. We’re at an inflection point — a point at which the economy has transitioned into a new technological era and we now need the state to do the same.

Come on then, how? This is complicated stuff but here are three promising areas for policy development — essentially three pillars of a new regulatory settlement that’s premised not on Smith’s serendipitous alignment but on misaligned markets in the digital age.

First, it’s pretty clear that we need to liberate ourselves from the idea of market failure and replace it with a concept more like market design. This means working not from the premise that ‘markets work well until they don’t’ but from the premise that ‘all markets have initial design conditions, and these conditions are decisive in determining the social order that emerges from the market’.

What do I mean by design conditions? An example is the way choices are framed: does a payment recur automatically, or is it a one-off, and what is the default outcome if a consumer doesn’t proactively change their mind? The key point is that there’s no such thing as neutral design conditions, and nor can the state be agnostic.

Some design conditions mean the misalignment in markets doesn’t really tell, so the market will create innovative products and happy consumers, just like our industrious baker. Other design conditions make the misalignment decisive, so that companies compete over how best to exploit our biases, while product quality/price stagnates. The state must therefore care about and shape a market’s initial design conditions. And, because this whole process is dynamic and unpredictable, this isn’t a ‘once and done’ job, it’s an ongoing task of refinement.

This is all very vague, so here are some concrete examples, some of which have already been tried by forward-thinking regulators.

  • We could stop companies using auto-renewal to exploit people’s inertia, either by defaulting people onto a backstop tariff when their first deal ends, or by banning so-called price-walking (the FCA’s eventual solution in insurance), or by capping the price difference between loyal customers and new customers.

  • We could ban subscription traps by requiring that it be as easy to leave a rolling deal as it was to sign up.

  • We could stop companies from exploiting situations of constrained choice, for example by requiring them to match the price they charge people in an open/active domain (like when we book a rental car online) with the price they charge people in a closed/stitch-up domain (like when we have to book a rental car last minute at the airport).

  • We could reset the economics of platforms by limiting the extent to which companies can build and take advantage of closed ecosystems, and by changing the way we govern the sharing and ownership of data (more on this below).

In some cases, steps like these have already been taken. But they’re still occasional forays, most of which are still justified under the cover of old arguments, rather than being a new regulatory paradigm.

The second pillar, which complements the first, is to shift away from a compliance / checkbox model of prescriptive regulatory rules to a model based on outcomes.

At root, this means being more curious as to whether or not a particular market is delivering good outcomes rather than just assuming things are OK unless something goes horribly wrong. This is a big deal because so many of the mechanisms of economic regulation are still tied to that baker-based philosophy of ‘don’t intervene unless you have to’. Regulators have to prove beyond doubt that something has ‘gone wrong’ before they’re allowed to ‘intervene’.

In an outcome-based approach, regulators could set an outcome like ‘treat customers fairly’, and they could monitor this with a basket of metrics including the variance in prices paid by people from different demographic groups. More creatively, regulators could monitor indicators of when a market is functioning not on the basis of active choices but via the exploitation of behavioural biases, for example by measuring consumer sentiments like surprise or regret.

This shift to outcomes-based regulation isn’t easy or risk-free, and there are thorny challenges around enforcement. Even so, some regulators have already moved decisively in this direction, with the UK finance regulator, the FCA, leading the way, in part because financial services are at the forefront of technological change and it’s so clear that prescriptive rules can’t handle a world of algorithms. Andrew Bailey, former CEO of the FCA and now Governor of the Bank of England, has spoken in favour of a shift to outcomes-based regulation and the FCA is already consulting on a new, stronger ‘consumer duty’ that sits in precisely this space. Still, when you step back to look across regulators, this work still feels like the exception, not the rule.

Third, we need to change how we classify and regulate digital platforms. We should stop treating the world’s biggest digital platforms as companies and instead class them as a new kind of entity with new responsibilities.[9]

What digital platforms are incentivised to do, and what they’re doing with great energy and ingenuity, is to establish monopoly positions in the markets in which they operate, and to extend horizontally outward from there.

For platforms that use a free-to-consumer business model, the root purpose of this work is — not through ill-will, but simply because it’s profitable — to accumulate ever more control over our mental environment so that this control can be sold to other companies with the intention of adjusting our behaviour (e.g. getting us to buy things). To treat these institutions within the 19th century regime of company law is patently absurd.

This is, I hope, the least contentious of the three pillars. In fact it feels like one of those areas of public policy where, in 10 years’ time, we’ll look back and wince at how naive we were. It’s like that point in a horror film, just before the protagonist wises up, when you rise from your seat in exasperation to scream: HE’S BEHIND YOU!!

So if digital platforms aren’t companies, what are they? I’ve written about this elsewhere but in essence they’re a new kind of institution that the world hasn’t seen before. For want of a better name, let’s embrace Zuckerberg’s rebrand and call them metas.

Being a meta can’t just be a label. It needs to be a concept — like ‘company’ or ‘charity’ — with legal significance, brought to life with a suite of attendant rules and institutions. Metas should be required to do/not do some things, and they should sit in an ecosystem of institutions that give them form.

What responsibilities should a meta have? Here are three examples:

  • Be interoperable so that people can move freely between your ecosystem and others. The point here is to stop metas from becoming private walled kingdoms, in which access is controlled capriciously in order to maximise profit.

  • Make data open (in aggregate form) for entrepreneurs and researchers to use. The point here is that the data we generate when using a meta shouldn’t ‘belong’ to the meta any more than the air we breathe out when we’re standing in McDonalds belongs to McDonalds. It’s our data, collectively, and it’s one of our most extraordinary sources of value, so we need access to its abundant potential.

  • Follow the digital rule of law so that metas don’t become more powerful than governments, and so that decisions of social or political significance are made by institutions with legitimacy in the relevant domain. This is how we avoid a world in which private companies determine things like freedom of speech.

Stepping back from the specifics, what I find reassuring about all of this is that it seems at least possible to build a regulatory settlement that doesn’t sit on the premise of the baker. I’m not saying that this will be easy, but there’s hope. We don’t have to cling fearfully to old orthodoxies as they disintegrate in our arms. We can ease ourselves onto something newer and more durable.

The end

The length of this piece is getting frankly embarrassing, so let’s wrap things up there.

What I’ve tried to surface is how digital capitalism affects one of the most important premises in economics — the idea that markets bring about a fortuitous, if heavily caveated, alignment of interests between producers and consumers. Alongside this I’ve tried to question a corresponding conception of the state, with its limited and prescriptive focus on byproducts and exceptional cases of market failure.

What I hope I’ve done, at least, is to sow some doubt. Are we sure the mental model of bees and bakers still works in the age of digital capitalism? And, if we’re not, shouldn’t we invest some proper time together answering the following question:

What kind of social order emerges from the behaviour of self-interested individuals in the technological conditions of a digital economy? a.k.a. What if the bees are on Facebook?

Judging from how things are going so far, it feels like the answer to this question is probably not a happy hive. In which case those old premises will need to be updated or, more likely, replaced.

This new way of thinking about the state’s relationship with markets is, it seems to me, just one part of a wider project: to upgrade our 20th century policy settlement for a 21st century digital economy.

Did I mention I’ve written a book about that? And that it’s available now, for a suspiciously tempting price, on everyone’s favourite meta?

End State: 9 ways society is broken — and how we can fix it is out now and available here.



  1. I’m paraphrasing here. In the original poem, Mandeville gave himself more latitude with the metaphor. As the hive breaks down, deflation kicks in, trade and manufacturing fall into a depression, the legal and medical systems fail, and pubs close as the bees stop buying beer.
  2. The link between Smith and Mandeville is more nuanced than I’ve implied here. Mandeville’s poem wasn’t making an economic point, it was a swipe at moralisers. His point was that you can’t rid society of vice without also losing the benefits it brings. (The original poem is here, along with a preface added later.) The line to Smith runs through Mandeville’s later work, in which he built out his earlier idea — of self-interested people collaborating to produce things of value — into an early incarnation of the division of labour. Smith never accepted the ethical aspects of Mandeville’s work and argued against them in The Theory of Moral Sentiments.
  3. This phrase came originally from a 1973 Richard Serra documentary,  and was later re-popularised by .
  4. Ok, so this analogy isn’t quite right. The move from QM to GR doesn’t ‘change the rules’ as such. It’s more that we haven’t yet found a unifying model that bridges the two.
  5. Some people will say this is just online advertising but that seems to me profoundly naive. It’s a bit like turning up at the Deathstar with a suitcase thinking it’s a really big Travelodge. Sure, it’s got loads of rooms and a pretty nice view, but as you ring the bell and stand there waiting, surely you’d wonder if you’d made a category error.
  6. Here’s a point I can’t stress enough: when I say ‘markets functioned OK’ my meaning is really narrow. I’m not saying that there weren’t harmful externalities, or that workers weren’t exploited, or that cartels didn’t stitch things up. I’m just putting these familiar failings of markets to one side to focus on the topic at hand: the way markets align the interests of bakers with the interests of people queuing for bread.
  7. I know people will shout at me for this, so just to emphasise: I’m not saying it’s easy to tackle an externality like pollution. Passing environmental legislation is really hard and companies do of course try to game these laws. Still, my instinct is that these laws are relatively easier to enforce than laws in the space of misalignment/behaviour manipulation, if only for the simple reason that you’re dealing with unintended byproducts, not intended/integral aspects of the ‘buy this thing’ experience.
  8. The FCA later promised to ban so-called ‘price-walking’ altogether, with the new rules coming into force in 2022, although as I write it’s unclear how industry will respond.
  9. The UK competition authority, the CMA, has come admirably close to this approach already. Building on the Furman Review, they’ve proposed designating the biggest tech companies with ‘Strategic Market Status’ and then treating these institutions differently to other companies in the regulatory regime.

Previously published here.

react to story with heart
react to story with light
react to story with boat
react to story with money
. . . comments & more!