I have a new paper out in the Evolutionary and Institutional Economics Review. It’s called ‘Economic Development and the Death of the Free Market’. The paper explores a question that has long fascinated me: how do human societies change as they consume more energy? As you probably guessed, the paper’s title gives away the conclusion. Consuming more energy seems to involve less free market and more hierarchy.
(Side note: yes, the paper is paywalled. Yes, this annoys me. No, I can’t afford to pay open-access publishing fees. Yes, I have a preprint up with content that is identical to the published version. Read it here.)
In markets we trust
Let’s start with what mainstream economists have to say about social change vis-à-vis consuming more energy. The short answer is that they say very little, since they don’t care much about energy. The long answer, however, is that if we look at the wider body of neoclassical economics, we see that the prime mover of pretty much everything is the ‘free market’.
Want to solve poverty? Use the market. Want less pollution? Use the market. Want economic growth? Use the market. And so on …
Outside the economics discipline, this relentless appeal to the free market is somewhat of a joke. But within the core of economics theory, it’s still taken seriously. The reason is that economists claim to have ‘proved’ something astounding: in a perfectly competitive market, the distribution of resources is ‘Pareto optimal’, meaning no person can be made better off without making another person worse off.
The ‘proof’ is called the first fundamental theorem of welfare economics, and it depends on a host of assumptions, all of which are violated in the real world. But I’m not concerned about that here. Instead, what concerns me is the bigger picture. Basically, what free-market theory tells us is that the route to all social ‘good’ is through individual ‘selfishness’. Stoke individual self interest, the thinking goes, and the invisible hand of the market will maximize social welfare.
By now, there’s a century’s worth of debate about these ideas. And to be blunt, I find much of it not worth reading. The reason is that almost all of the debate has taken place on theoretical grounds, without looking at actual human behavior. Moreover, the debate has disregarded fascinating developments in the rest of science.
The architecture of complexity
Many scientists have realized that complex structure seems to always arise in the same way. It is built using hierarchy. Noting this fact, the polymath Herbert Simon called hierarchy the ‘architecture of complexity’.
If you look at complex structures, you’ll see that they are typically composed by merging simpler components. This nesting behavior is easy to explain. It’s a result of cosmic evolution. As far as we know, the universe started off as a boring place — a homogeneous soup of sub-atomic particles. Everything that followed — galaxies, stars, planets, life, humans — had to be built from this simple starting point. Because the universe began as a simple place, there was no alternative but to build complexity by merging simpler components. Thus, complexity is always hierarchical.
In addition to this ‘nesting’ sense, complexity can be hierarchical in the sense of having a chain of command. True, galaxies have no command structure, nor do planets or any other non-living structure. But among life, command structures are everywhere.
Your body is a good example. You are composed of trillions of cells that work together to survive. And you are not simply an amalgamation of autonomous cells. No, your body has a distinct hierarchy. The central nervous system sends signals that control the rest of the cells in your body, allowing complex activities like running.
If you look across life on Earth, you see the same pattern everywhere. Large, complex organisms are not lumps of otherwise autonomous cells. They are hierarchically organized. The result is enhanced freedom for the organism, but a distinct lack of freedom for the constituent cells.
Selecting against selfishness
Why do complex organisms turn to hierarchy (and the associated command structure) to organize? The theory of ‘multilevel selection’ gives a possible answer.
If complex organisms are built by merging previously autonomous components, this comes with an inherent problem. The merger creates a tension between units of selection. It’s usually best for individuals with a group to act selfishly, because doing so will increase their relative fitness (i.e. reproduction). But such selfish behavior tends to corrode the group’s cohesion. So we have a dilemma. The behavior that is best for the group is not necessarily what is best for relative fitness within the group. This tension between units of selection means that if complex structure is to evolve, it must suppress selection at lower levels. The hierarchical chain of command seems to be a common solution to this problem.
I find multilevel selection theory fascinating for many reasons, one of which is that it contradicts (in almost every way) what neoclassical economics tells us about society. Ever since Adam Smith, economists have praised the social merits of individual selfishness, claiming that stoking self interest will (paradoxically) maximize social welfare. But if multilevel selection theory is correct, this free-market thinking is bollocks.
Looking at the rest of life on Earth, there’s no evidence for the kind of ‘invisible hand’ theorized by economists. When life becomes complex, it suppresses the autonomy of sub-components. That said, it’s possible that humans are the exception to the rule. With this possibility in mind, let’s look at the evidence.
Evidence for the growth of hierarchy: firms and governments get bigger
What I’m going to do now is bombard you with evidence suggesting that as human societies consume more energy, they become more hierarchical. After I show you the evidence, I’ll tell you what I think it means.
Let’s start with ‘firms’, the quintessential unit of market competition. Neoclassical economists theorize that the best kind of firm is a small one. That’s because the smaller firms become, the closer we get to the magical ideal of the ‘perfectly competitive market’.
Unfortunately (for neoclassical economics), it seems that real-world societies ignore this ideal. Everywhere we look, economic development comes with bigger firms. Figure 1 shows the trend across countries. Societies that consume more energy tend to have larger firms. And as the inset chart shows, the entire firm-size distribution changes with energy use, shifting towards bigger firms.

In the main panel, each point shows how average firm size relates to energy use per capita within countries. The inset chart shows the size distribution of firms, differentiated by energy-use quartile (color).
It’s not just firms that get bigger. As energy use increases, governments get bigger too. Figure 2 plots the international trend in the government share of employment as it relates to energy use per capita. More energy means more government.

As you’d expect, politics play a big role in shaping the size of government. As Figure 3 shows, communist countries tend to have larger governments than non-communist countries. What’s unexpected, though, is that both types of countries exhibit the same trend (shown in the inset plot): governments get bigger as energy use increases.

I’ve replotted the data from Figure 2, but this time differentiated between countries that have (or once had) a communist regime, and countries that have never had a communist government. The inset panel shows the smoothed trend across each type of country.
Evidence for the growth of hierarchy: more managers
What I’m ultimately interested in is not the size of firms and governments per se, but what this size represents. I think that the growth of these institutions represents the growth of hierarchy.
If I had my way, I’d measure the growth of hierarchy directly. But the truth is that we don’t have the data that is required. What we do have, though, is data that is plausible related to hierarchy. To proxy the growth of hierarchy, I propose we count managers. Since their job is to coordinate the activity of others, managers live at the tops of hierarchy. So as societies accumulate managers, it’s reasonable to suppose that they are also accumulating hierarchy.
On that front, the evidence is pretty clear that as societies consume more energy, they accumulate more managers. Here’s the international trend.

Each line represents the path through time of an individual country. The black line shows the smooth trend across all observations.
Like government size, it seems plausible that the number of managers is influenced by politics. But it turns out this isn’t the case, as shown in Figure 5. Here I’ve taken the energy-manager data and differentiated between communist and non-communist countries. Surprisingly, there’s no discernible difference in the trend. So whatever is driving societies to accumulate managers, its not politics.

I’ve replotted the data from Figure 4, but this time differentiated between communist and non-communist countries. As shown in the inset panel, when it comes to the accumulation of managers with energy use, there’s no discernible difference between the two types of countries.
Inferring the growth of hierarchy
Let’s recap what we know. As energy use per capita increases:
- Firms and governments get bigger;
- Societies accumulate managers.
Both of these trends, I propose, strongly hint that increasing energy consumption is associated with the growth of hierarchy.
To make the case even stronger, we can use a model to work backwards. Suppose that firms and governments are all hierarchically organized. And suppose that these hierarchical institutions grow larger with energy use, as suggested by the empirical data. If so, what does this say about the accumulation of managers?
Well, it turns out that it says quite a lot. I won’t go into the details of the model here, but instead show you the results. Basically, if firms and governments get larger with energy use like the empirical data suggests, and these institutions are hierarchically organized, then we get the pattern shown in Figure 6.
The black points show the empirical trend between energy use and the relative number of managers. The rainbow shows the trend predicted by my ‘energy-hierarchy’ model. Color indicates the ‘span of control’ — the number of subordinates controlled by each superior in the model. (The span of control is a free parameter, so I let it vary.) What’s most striking is the trend in the inset panel. Here the blue line shows the international trend among countries. The red line shows the prediction made by the best-fit energy hierarchy model. As you can see, there’s not much error. In other words, the accumulation of managers is a straightforward consequence of the growth of hierarchy.

The black points replot the empirical data from Figure 4. The rainbow shows the results of my ‘energy-hierarchy’ model. The inset model shows the smoothed trend across countries, and the corresponding best fit model. The ‘span of control’ (a free parameter) determines the number of subordinates controlled by each superior in the model.
With the energy-hierarchy model in hand, we can then work backwards and infer how hierarchy grows with economic development. In technical terms, a ‘hierarchy’ is a particular form of network that has a treelike structure. The more treelike the network, the more hierarchical it is. How we actually quantify this treelike structure gets a bit technical. One measure is called ‘global reaching centrality’, which basically measures the degree of centralization in a network.
Another measure (that I find more intuitive) quantifies the inequality of ‘hierarchical power’. We assume that within a hierarchy, power is proportional to the total number of subordinates you control. We measure this ‘hierarchical power’ for every individual, and then calculate its inequality using the Gini index. I call the result the ‘concentration of hierarchical power’ — a metric that tells us about the accumulation of hierarchy.
The catch is that we don’t actually have the data needed to apply these metrics to real-world societies. (Doing so would require knowing the network structure of the whole society.) However, what we can do is match the energy-hierarchy model with a real-world society, and then use the model to infer the degree of hierarchy in the associated society.
Figure 7 shows the results of this estimate. Since it’s a model-based inference, you have to treat it with caution. Still, the results are pretty stunning. I infer that as energy use increases, there’s a massive increase in the hierarchical structure of society.

This figure plots the results of a two-step calculation. First, I match the energy-hierarchy model with a real-world country. Then I use the data from the model to infer the degree of hierarchy in the given society. The main panel shows this degree of hierarchy, measured using the ‘concentration of hierarchical power’. The inset panel measures the degree of hierarchy using ‘global reaching centrality’. Both metrics suggest that as societies consume more energy, they accumulate hierarchy.
Where’s the free market?
The empirical data suggests that to ramp up the consumption of energy, humans turn to hierarchy. That leaves free-market theory in a tough spot. If the perfectly competitive market is so ‘optimal’, why don’t societies use it to industrialize? Why do they instead head in the opposite direction towards less market competition?
One possibility is that neoclassical economists are correct, but human societies have failed to get the message. Thus, we’re stuck with social configurations that are ‘sub optimal’. While this could be true, it implies that humans are rather stupid. A better social arrangement sits before our eyes (and is preached constantly by economists) yet we refuse to see it.
A more plausible reason that humans do not listen to neoclassical economics is that it is wrong. In other words, the market is not the solution to all of humanity’s problems. Some problems, it seems, are solved by accumulating hierarchy.
Still, we are left with the stunning persistence of neoclassical theory. Critics have been railing against it for a century, yet free market ideas have nonetheless spread. In fact, they have spread at the very time that hierarchy exploded. Figure 8 shows this trend in the United States. Paradoxically, in American English the word frequency of free-market jargon grew in tandem with the growth of government and the growth of managers.

I’ve plotted here the word frequency (in American English) of 4 free-market terms. Par.adoxically, this frequency increased at the same time that the US accumulated more government and more managers.
So what’s going on here? One possibility is that free-market ideologues are simply responding to the growth of hierarchy. In other words, they’re shouting louder as the free market dies, but to no effect.
Another possibility is that free market ideas actually do the opposite of what they claim. While free-market theory champions the autonomy of individuals, perhaps its effect is actually to subvert this autonomy by enabling the growth of hierarchy.
This subversion sounds odd … until you look more closely at the neoclassical notion of free markets. You see, there’s an odd asymmetry in neoclassical theory. Consumers are always treated as utility maximizing individuals. But producers are treated as profit-maximizing firms. So when neoclassical economists champion the free market, they are actually championing the autonomy of firms. And that autonomy translates into power for firm owners. In other words, when you champion the ‘free market’, you are backing owners’ power to accumulate hierarchy.
This idea might sound far-fetched. After all, championing ‘freedom’ doesn’t sound like a great way to facilitate subordination in a hierarchy. But there’s actually data to back up this hypothesis.
Take, as an example, ‘individualistic psychology’. You’d think that societies that are more ‘individualistic’ would be breeding grounds for the free market, not bastions of hierarchy. Yet it turns out that the reverse is true. As Figure 9A shows, countries with a more individualistic psycology tend to have more managers (and hence, more hierarchy). Likewise, societies that are ‘looser’ — meaning they are more tolerant of deviant behavior — also tend to have more managers (Figure 9B).

Societies with more managers tend to have a more ‘individualistic’ psychology, as measured by Geert Hofstede ‘individualism index’ (panel A). They also tend to be ‘looser’, as measured by Michele Gelfand’s index of cultural ‘tightness’ (panel B).
Admittedly, these correlations are not tight, so they are far from conclusive evidence. But the point is that they support the idea that, contrary to what we might think, free-market thinking might actually stoke the growth of hierarchy.
A ‘massive fiction’
Heterodox economists have long criticized free-market theory as being wrong. Less attention, however, has been given to how this ‘theory’ might affect human behavior.
On that front, let’s return to evolutionary biology. According to the theory of multilevel selection, successful groups must suppress the self interest of individuals and stoke pro-social behavior. The obvious way to do this, you’d think, would be to praise the merits of altruism and self-sacrifice. But that, it turns out, is a terrible way to create an ideology.
Interestingly, we know this from the word ‘altruism’ itself. It was coined in the 19th by Auguste Comte who proposed it as the basis of a new ‘Religion of Humanity’ to be based explicitly on self sacrifice. Comte’s movement was a dismal failure.
Evolutionary biologist David Sloan Wilson thinks the problem was simple: Comte preached the truth. And the truth is that prosocial behavior does involve sacrifice on the part of individuals. That’s because it often pays to lie and cheat. Therefore, when you act pro-socially by not lying and cheating, you’re implicitly making a sacrifice.
Comte’s mistake was to emphasize this sacrifice. Successful ideologies, Wilson observes, do the opposite — they proclaim that what is good for the group is also good for individuals (and vice versa). Thus, many religions declare that pro-social behavior is good for individuals too. In other words these ideologies promote pro-social behavior by being ‘massively fictional’ (Wilson’s words).
If you look closely at free-market theory, you see that it does the same thing. It claims that what’s in the interest of the individual is also in the interest of the group. The twist is that free-market theory declares that rampant selfishness is good for the group. That is a ‘massive fiction’. But it may nonetheless serve a purpose. Indeed, the evidence suggests that such thinking may actually aid the growth of hierarchy.
Whether this idea is true, of course, is an open question. But I think it’s a fascinating topic for future research. Assume that humans follow the wider trends found in evolution. Treat economics as an ideology that helps accomplish these feats. And treat economists as high priests who promulgate ‘massive fictions’.
(Warning: doing this research will make you unemployable in a mainstream economics department.)
Support this blog
Economics from the Top Down is where I share my ideas for how to create a better economics. If you liked this post, consider becoming a patron. You’ll help me continue my research, and continue to share it with readers like you.
Stay updated
Sign up to get email updates from this blog.
This work is licensed under a Creative Commons Attribution 4.0 License. You can use/share it anyway you want, provided you attribute it to me (Blair Fix) and link to Economics from the Top Down.
Interesting post!
What is intruiging in many of the figures is the apparent leveling off and even downturn of complexity with increasing energy use. I suspect that it is an artifact of using only human levels of hierarchy and not including the rapid increase in use of machine mechanisms for control of complexity. Moore’s Law and all that.
Ecomodernists want to extend those curves much further to the right. Perhaps 5,000 GJ per capita might be enough to reduce the human participation in control of the economy to zero, resulting in “fully automated luxury communism”, where robots do all the work of managing a very complex economy and humans are on permanent vacation.
Such huge energy expenditures could be problematic for the environment, though, even if we could produce that much energy, plus I wonder where the other resources needed to provide for people’s needs and maintain the machines will come from.
But I expect that the global economy will fail from overcomplexity and resource depletion long before we get to our utopia of 100% leisure. I can’t prove it, but I expect that there is a “Peter Principle” aspect to economic and social complexity, wherein complexity rises to just past the point where it can be successfuly managed, at which point everything falls apart.
Hi Joe,
I can’t speak to ‘complexity’, but the levelling off of the number of managers is easy to explain. In the limit that the entire society is a single hierarchy, the number of managers approaches 1/s^2, where s is the span of control. I think I put the derivation somewhere in the paper’s appendix.
I just read Scale by Geoffrey West in which he shows that when organisms grow in size they minimize their energy use by a hierarchy of capillaries which transmit blood to every part of the body. By dividing in two they reach areas logarithmically so that a larger animal requires many less veins to transmit energy than a smaller one whose metabolism is proportionally much higher. The same hold for companies who develop hierarchies analogous to the circulatory system. Consequently there is a maximum size for organism because of the need to reach each cell. Although organisms and companies must stop growing and die. Interesting this is not true for cities
[…] century, big corporations increasingly dominated society, all the while championing the ethos of ‘free markets’. But we should not forget that an ethos can change, sometimes in a shockingly short […]