The Relationship Between Hierarchy and Wealth

Epistemic Status: Tentative

I’m fairly anti-hierarchical, as things go, but the big challenge to all anti-hierarchical ideologies is “how feasible is this in real life? We don’t see many examples around us of this working well.”

Backing up, for a second, what do we mean by a hierarchy?

I take it to mean a very simple thing: hierarchies are systems of social organization where some people tell others what to do, and the subordinates are forced to obey the superiors.  This usually goes along with special privileges or luxuries that are only available to the superiors.  For instance, patriarchy is a hierarchy in which wives and children must obey fathers, and male heads of families get special privileges.

Hierarchy is a matter of degree, of course. Power can vary in the severity of its enforcement penalties (a government can jail you or execute you, an employer can fire you, a religion can excommunicate you, the popular kids in a high school can bully or ostracize you), in its extent (a totalitarian government claims authority over more aspects of your life than a liberal one), or its scale (an emperor rules over more people than a clan chieftain.)

Power distance is a concept from the business world that attempts to measure the level of hierarchy within an organization or culture.  Power distance is measured by polling less-powerful individuals on how much they “accept and expect that power is distributed unequally”.  In low power distance cultures, there’s more of an “open door” policy, subordinates can talk freely with managers, and there are few formal symbols of status differentiating managers from subordinates.  In “high power distance” cultures, there’s more formality, and subordinates are expected to be more deferential.  According to Geert Hofstede, the inventor of the power distance index (PDI), Israel and the Nordic countries have the lowest power distance index in the world, while Arab, Southeast Asian, and Latin American countries have the highest.  (The US is in the middle.)

I share with many other people a rough intuition that hierarchy poses problems.

This may not be as obvious as it sounds.  In high power distance cultures, empirically, subordinates accept and approve of hierarchy.  So maybe hierarchy is just fine, even for the “losers” at the bottom?  But there’s a theory that subordinates claim to approve of hierarchy as a covert way of getting what power they can.   In other words, when you see peasants praising the benevolence of landowners, it’s not that they’re misled by the governing ideology, and not that they’re magically immune to suffering from poverty as we would in their place, but just that they see their situation as the best they can get, and a combination of flattery and (usually religious) guilt-tripping is their best chance for getting resources from the landowners.  So, no, I don’t think you can assume that hierarchy is wholly harmless just because it’s widely accepted in some societies. Being powerless is probably bad, physiologically and psychologically, for all social mammals.

But to what extent is hierarchy necessary?

Structurelessness and Structures

Nominally non-hierarchical organizations often suffer from failure modes that keep them from getting anything done, and actually wind up quite hierarchical in practice. I don’t endorse everything in Jo Freeman’s famous essay on the Tyranny of Structurelessness, but it’s important as an account of actual experiences in the women’s movement of the 1970s.

When organizations have no formal procedures or appointed leaders, everything goes through informal networks; this devolves into popularity contests, privileges people who have more free time to spend on gossip, as well as people who are more privileged in other ways (including economically), and completely fails to correlate decision-making power with competence.

Freeman’s preferred solution is to give up on total structurelessness and accept that there will be positions of power in feminist organizations, but to make those positions of power legible and limited, with methods derived from republican governance (which are also traditional in American voluntary organizations.)  Positions of authority should be limited in scope (there is a finite range of things an executive director is empowered to do), accountable to the rest of the organization (through means like voting and annual reports), and impeachable in cases of serious ethical violation or incompetence. This is basically the governance structure that nonprofits and corporations use, and (in my view) it helps make them, say, less likely to abuse their members than cults and less likely to break up over personal drama than rock bands.

Freeman, being more egalitarian than the republican tradition, also goes further with her recommendations and says that responsibilities should be rotated (so no one person has “ownership” over a job forever), that authority should be distributed widely rather than concentrated, that information should be diffused widely, and that everyone in the organization should have equal access to organizational resources.  Now, this is a good deal less hierarchical than the structure of republican governments, nonprofits, and corporations; it is still pretty utopian from the point of view of someone used to those forms of governance, and I find myself wondering if it can work at scale; but it’s still a concession to hierarchy relative to the “natural” structurelessness that feminist organizations originally envisioned.

Freeman says there is one context in which a structureless organization can work; a very small team (no more than five) of people who come from very similar backgrounds (so they can communicate easily), spend so much time together that they practically live together (so they communicate constantly), and are all capable of doing all “jobs” on the project (no need for formal division of labor.)  In other words, she’s describing an early-stage startup!

I suspect Jo Freeman’s model explains a lot about the common phenomenon of startups having “growing pains” when they get too large to work informally.  I also suspect that this is a part of how startups stop being “mission-driven” and ambitious — if they don’t add structure until they’re forced to by an outside emergency, they have to hurry, and they adopt a standard corporate structure and power dynamics (including the toxic ones, which are automatically imported when they hire a bunch of people from a toxic business culture all at once) instead of having time to evolve something that might achieve the founders’ goals better.

But Can It Scale? Historical Stateless Societies

So, the five-person team of friends is a non-hierarchical organization that can work.  But that’s not very satisfying for anti-authoritarian advocates, because it’s so small.  And, accordingly, an organization that small is usually poor — there’s only so many resources that five people can produce.

(Technology can amplify how much value a single person can produce. This is probably why we see more informal cultures among people who work with high-leverage technology.  Software engineers famously wear t-shirts, not suits; Air Force pilots have a reputation as “hotshots” with lax military discipline compared to other servicemembers. Empowered with software or an airplane, a single individual can be unusually valuable, so  less deference is expected of the operators of high technology.)

When we look at historical anarchies or near-anarchies, we usually also see that they’re small, poor, or both.  We also see that within cultures, there is often surprisingly more freedom for women among the poor than among the rich.

Medieval Iceland from the tenth to thirteenth centuries was a stateless society, with private courts of law, and competing legislative assemblies (Icelanders could choose which assembly and legal code to belong to), but no executive branch or police.  (In this, it was an unusually pure form of anarchy but not unique — other medieval European polities had much more private enforcement of law than we do today, and police are a 19th-century invention.)

The medieval Icelandic commonwealth lasted long enough — longer than the United States — that it was clear this was a functioning system, not a brief failed experiment.  And it appears that it was less violent, not more, compared to other medieval societies.  Even when the commonwealth was beginning to break down in the thirteenth century, battles had low casualty rates, because every man still had to be paid for!  The death toll during the civil war that ended the commonwealth’s independence was only as high per capita as the current murder rate of the US.  While Christianization in neighboring Norway was a violent struggle, the decision of whether to convert to Christianity in Iceland was decided peacefully through arbitration.  In this case, it seems clear that anarchy brought peace, not war.

However, medieval Iceland was small — only 50,000 people, confined to a harsh Arctic environment, and ethnically homogeneous.

Other historical and traditional stateless societies are and were also relatively poor and low in population density. The Igbo of Nigeria traditionally governed by council and consensus, with no kings or chiefs, but rather a sort of village democracy.   This actually appears to be fairly common in small polities.  The Iroquois Confederacy governed by council and had no executive. (Note that the Iroquois are a hoe culture.)  The Nuer of Sudan, a pastoral society currently with a population of a few million, have traditionally had a stateless society with a system of feud law — they had judges, but no executives. There are many more examples — perhaps most familiar to Westerners, the society depicted in the biblical book of Judges appears to have had no king and no permanent war-leader, but only judges who would decide cases which would be privately enforced. In fact, stateless societies with some form of feud law seem to be a pretty standard and recurrent type of political organization, but mostly in “primitive” communities — horticultural or pastoral, low in population density.  This sounds like bad news for modern-day anarchists who don’t want to live in primitive conditions. None of these historical stateless societies, even the comparatively sophisticated Iceland, are urban cultures!

It’s possible that the Harappan civilization in Bronze Age India had no state, while it had cities that housed tens of thousands of people, were planned on grids, and had indoor plumbing.  The Harappans left no massive tombs, no palaces or temples, houses of highly uniform size (indicating little wealth inequality) no armor and few weapons (despite advanced metalworking), no sign of battle damage on the cities or violent death in human remains, and very minimal city walls.  The Harappan cities were commercial centers, and the Harappans engaged in trade along the coast of India and as far as Afghanistan and the Persian Gulf.  Unlike other similar river-valley civilizations (such as Mesopotamia), the Harappans had so much arable land, and farmsteads so initially spread out, that populations steadily grew and facilitated long-distance trade without having to resort to raiding, so they never developed a warrior class.  If so, this is a counterexample to the traditional story that all civilizations developed states (usually monarchies) as a necessary precondition to developing cities and grain agriculture.

Bali is another counterexample.  Rice farming in Bali requires complex coordination of irrigation. This was traditionally not organized by kings, but by subaks, religious and social organizations that supervise the growing of rice, supervised by a decentralized system of water temples, and led by priests who kept a ritual calendar for timing irrigation.  While precolonial Bali was not an anarchy but a patchwork of small principalities, large public works like irrigation were not under state control.

So we have reason to believe that Bronze Age levels of technological development (cities, metalworking, intensive agriculture, literacy, long-distance trade, and high populations) can be developed without states, at scales involving millions of people, for centuries.  We also have much more abundant evidence, historical and contemporary, of informal governance-by-council and feud law existing stably at lower technology levels (for pastoralists and horticulturalists).  And, in special political circumstances (the Icelanders left Norway to settle a barren island, to escape the power of the Norwegian king, Harald Fairhair) an anarchy can arise out of a state society.

But we don’t have successful examples of anarchies at industrial tech levels. We know industrial-technology public works can be built by voluntary organizations (e.g. the railroads in the US) but we have no examples of them successfully resisting state takeover for more than a few decades.

Is there something about modern levels of high technology and material abundance that is incompatible with stateless societies? Or is it just that modern nation-states happened to already be there when the Industrial Revolution came around?

Women’s Status and Material Abundance

A very weird thing is that women’s level of freedom and equality seems almost to anticorrelate with the wealth and technological advancement.

Horticultural (or “hoe culture“) societies are non-patriarchal and tend to allow women more freedom and better treatment in various ways than pre-industrial agricultural societies. For instance, severe mistreatment of women and girls like female infanticide, foot-binding, honor killings, or sati, and chastity-oriented restrictions on female freedom like veiling and seclusion, are common in agricultural societies and unknown in horticultural ones. But horticultural societies are poor in material culture and can’t sustain high population densities in most cases.

You also see unusual freedom for women in premodern pastoral cultures, like the Mongols. Women in the Mongol Empire owned and managed ordos, mobile cities of tents and wagons which also comprised livestock and served as trading hubs.  While the men focused on hunting and war, the women managed the economic sphere. Mongol women fought in battle, herded livestock, and occasionally ruled as queens.  They did not wear veils or bind their feet.

We see numerous accounts of ancient and medieval women warriors and military commanders among Germanic and Celtic tribes and steppe peoples of Central Asia.  There are also accounts of medieval European noblewomen who personally led armies. The pattern isn’t obvious, but there seem to be more accounts of women military leaders in pastoral societies or tribal ones than in large, settled empires.

Pastoralism, to a lesser extent than horticulture but still more than plow agriculture, gives women an active role in food production. Most pastoral societies today have a traditional division of labor in which men are responsible for meat animals and women are responsible for milk animals (as well as textiles).  Where women provide food, they tend to have more bargaining power.  Some pastoral societies, like the Tuareg, are even matrilineal; Tuareg women traditionally have more freedom, including sexual freedom, than they do in other Muslim cultures, and women do not wear the veil while men do.

Like horticulture, pastoralism is less efficient per acre at food production than agriculture, and thus does not allow high population densities. Pastoralists are poorer than their settled farming neighbors. This is another example of women being freer when they are also poorer.

Another weird and “paradoxical” but very well-replicated finding is that women are more different from men  in psychological and behavioral traits (like Big 5 personality traits, risk-taking,  altruism, participation in STEM careers) in richer countries than in poorer ones.  This isn’t quite the same as women being less “free” or having fewer rights, but it seems to fly in the face of the conventional notion that as societies grow richer, women become more equal to men.

Finally, within societies, it’s sometimes the case that poor women are treated better than rich ones.  Sarah Blaffer Hrdy writes about observing that female infanticide was much more common among wealthy Indian Rajput families than poor ones. And we know of many examples across societies of aristocratic or upper-class women being more restricted to the domestic sphere, married off younger, less likely to work, more likely to experience restrictive practices like seclusion or footbinding, than their poorer counterparts.

Hrdy explains why: in patrilinear societies, men inherit wealth and women don’t. If you’re a rich family, a son is a “safe” outcome — he’ll inherit your wealth, and your grandchildren through him will be provided for, no matter whom he marries. A daughter, on the other hand, is a risk. You’ll have to pay a dowry when she marries, and if she marries “down” her children will be poorer than you are — and at the very top of the social pyramid, there’s nowhere to marry but down.  This means that you have an incentive to avoid having daughters, and if you do have daughters, you’ll be very anxious to avoid them making a bad match, which means lots of chastity-enforcement practices. You’ll also invest more in your sons than daughters in general, because your grandchildren through your sons will have a better chance in life than your grandchildren through your daughters.

The situation reverses if you’re a poor family. Your sons are pretty much screwed; they can’t marry into money (since women don’t inherit.) Your daughters, on the other hand, have a chance to marry up. So your grandchildren through your daughters have better chances than your grandchildren through your sons, and you should invest more resources in your daughters than your sons. Moreover, you might not be able to afford restrictive practices that cripple your daughters’ ability to work for a living. To some extent, sexism is a luxury good.

A similar analysis might explain why richer countries have larger gender differences in personality, interests, and career choices.  A degree in art history might function as a gentler equivalent of purdah — a practice that makes a woman a more appealing spouse but reduces her earning potential. You expect to find such practices more among the rich than the poor.  (Tyler Cowen’s take is less jaundiced, and more general, but similar — personal choices and “personality” itself are more varied when people are richer, because one of the things people “buy” with wealth is the ability to make fulfilling but not strictly pragmatic self-expressive choices.)

Finally, all these “paradoxical” trends are countered by the big nonparadoxical trend — by most reasonable standards, women are less oppressed in rich liberal countries than in poor illiberal ones.  The very best countries for women’s rights are also the ones with the lowest power distance: Nordic and Germanic countries.

Is Hierarchy the Engine of Growth or a Luxury Good?

If you observe that the “freest” (least hierarchical, lowest power distance, least authoritarian, etc) functioning organizations and societies tend to be small, poor, or primitive, you could come to two different conclusions:

  1. Freedom causes poverty (in other words, non-hierarchical organization is worse than hierarchy at scaling to large organizations or rich, high-population societies)
  2. Hierarchy is expensive (in other words, only the largest organizations or richest societies can afford the greatest degree of authoritarianism.)

The first possibility is bad news for freedom. It means you should worry you can’t scale up to wealth for large populations without implementing hierarchies.  The usual mechanism proposed for this is the hypothesis that hierarchies are needed to coordinate large numbers of people in large projects.  Without governments, how would you build public works? Or guard the seas for global travel and shipping? Without corporate hierarchies, how would you get mass-produced products to billions of people?  Sure, goes the story, idealists have proposed alternatives to hierarchy, and we know of intriguing counter examples like Harappa and Bali, but these tend to be speculative or small-scale and the success stories are sporadic.

The second possibility is (tentative)  good news for freedom.  It says that hierarchy is inefficient.  For instance, secluding women in harems wastes their productive potential. Top-down state control of the economy causes knowledge problems that limit economic productivity. The same problem applies to top-down control of decisionmaking in large firms.  Dominance hierarchies inhibit accurate transmission of information, which worsens knowledge problems and principal-agent problems (“communication is only possible between equals.”)  And elaborate displays of power and deference are costly, as nonproductive displays always are.  Only accumulations of large amounts of resources enable such wasteful activity, which benefits the top of the hierarchy in the short run but prevents the “pie” of total resources from growing.

This means that if you could just figure out a way to keep inefficient hierarchies from forming, you could grow systems to be larger and richer than ever.  Yes, historically, Western economies grew richer as states grew stronger — but perhaps a stateless society could be richer still.  Perhaps without the stagnating effects of rent-seeking, we could be hugely better off.

After all, this is kind of what liberalism did. It’s the big counter-trend to “wealth and despotism go together” — Western liberal-democratic countries are much richer and much less authoritarian (and less oppressive to women) than any pre-modern society, or than developing countries. One of the observations in Wealth of Nations is that countries with strong middle classes had more subsequent economic growth than countries with more wealth inequality — Smith uses England as an example of a fast-growing, equal society and China as an example of a stagnant, unequal one.

But this is only partial good news for freedom, after all. If hierarchies tend to emerge as soon as size, scale, and wealth arise, then that means we don’t have a solution to the problem of preventing them from emerging. On a model where any sufficiently large accumulation of resources begins to look attractive to “robber barons” who want to appropriate it and forcibly keep others out, we might hypothesize that a natural evolution of all human institutions is from an initial period of growth and value production towards inevitable value capture, stagnation, and decline.  We see a lack of freedom in the world around us, not because freedom can’t work well, but because it’s hard to preserve against the incursions of wannabe despots, who eventually ruin the system for everyone including themselves.

That model points the way to new questions, surrounding the kinds of governance that Jo Freeman talks about. By default an organization will succumb to inefficient hierarchy, and structureless organizations will succumb faster and to more toxic hierarchies. When designing governance structures, the question you want to ask is not just “is this a system I’d want to live under today?” but “how effective will this system be in the future at resisting the guys who will come along and try to take over and milk it for short-term personal gain until it collapses?”  And now we’re starting to sound like the rationale and reasoning behind the U.S. Constitution, though I certainly don’t think that’s the last word on the subject.


Book Recommendations: An Everyone Culture and Moral Mazes

Epistemic Status: Casual

I highly recommend An Everyone Culture, by Robert Kegan, and Moral Mazes, by Robert Jackall, as companion books on business culture. Moral Mazes is an anthropological study of the culture and implicit ethics of a few large corporations, and is an eye-opening illustration of the problems that arise in those corporations. An Everyone Culture is an introduction to the idea of a “deliberately developmental organization”, an attempt to fix those problems, plus some case studies of companies that implemented “deliberately developmental” practices.

The basic problem that both books observe in corporate life is that everybody in a modern office is trying to conceal their failures and present a misleadingly positive impression of themselves to their employers and coworkers.

This leads to lost productivity.

For instance:

  • The longer one tries to cover up a mistake, the costlier it will be to fix it.
  • The less accurately credit is allocated for success or failure, the harder it will be to incentivize good work.
  • The more employees misinform their bosses, the worse-informed the bosses’ decisions will be.
  • The more people are concerned with maintaining appearances, the less cognitive capacity they will have for productivity and creativity.
  • The more unacceptable it is to acknowledge “personal” concerns (emotions, physical health, intrinsic motivation or lack thereof), the harder it is to fix productivity problems that arise from “personal” problems.

Moral Mazes basically takes the view that the Protestant work ethic really died in the mid-to-late nineteenth century, when an American economy defined by small business owners and freelance professionals was replaced by an economy defined by larger firms and the rise of the managerial profession. The Protestant work ethic declared that hard work, discipline, and honesty would bring success. The “managerial work ethic” holds that a good employee has quite different “virtues” — things like

  • ability to play politics
  • loyalty & willing to subordinate oneself to one’s manager
  • “flexibility” (the opposite of stubbornness — not holding strong individual opinions)

To give an outside example, the author of “The Western Elite from a Chinese Perspective” was coming from a “Protestant work ethic” culture of hard work (though not, of course, actually Protestant) and encountering the “managerial work ethic” culture of American office politics.

Moral Mazes relies on the author’s observations and interviews with managers. I’m sure it’s not a fully objective portrayal — perhaps the author selected the most damning quotes, and perhaps the most disgruntled and cynical managers were the most willing to talk.  But the picture the book gives is of a culture where:

  • rank is everything — contradicting your boss, especially in public, is career suicide, and deference to superiors is expected
  • beyond a certain minimum floor of competence, objective job performance doesn’t determine career success, political skill does
  • “credit flows upwards, details flow downwards” — higher-rank managers take credit for work done by their subordinates, and the higher-rank you are, the fewer object-level details you concern yourself with
  • mistakes and bad decisions are reliably concealed; then, when the inevitable catastrophe happens, whoever’s politically vulnerable takes the fall
  • managers are tested for their “flexibility” — someone with strong opinions about the best engineering decisions or with rigid ethical principles will not rise far in their career

If you watch The Marvelous Mrs. Maisel, Joel Maisel’s job at the plastics company is a classic example of the managerial work ethic; he’s basically a professional sycophant.  He’s burned out and unmotivated, and he leaves to “find himself” as a comedian, but quickly realizes he has no talent at comedy either.  Instead, working in his father’s garment business, he comes to life again.  He learns the nitty-gritty of the factory floor, the accounting, the machines, the seamstresses and their personal needs and strengths and weaknesses.  It’s a beautiful illustration of the difference between fake work and real work.

An Everyone Culture‘s prescription for the problems of deception, sycophancy, and stagnation in conventional companies is complex, but I’d summarize it as follows: creating a culture where everyone talks about mistakes and improvements, and where the personal/professional boundaries are broken down.

This sounds vaguely cultish and shocking, and indeed, the companies profiled (like Bridgewater) are often described as cults.  Kegan acknowledges that their practices are outside most of our comfort zones, but believes that nothing inside the range of what we think of as a normal workplace will solve workplace dysfunctions.

What distinguishes the companies profiled in the book is a lot of talk, about issues that would ordinarily be considered too “personal” for work.  When a failure occurs, a DDO looks for the root cause, as you would in a kaizen system, but it won’t stop there — people will also ask what personal or psychological issue caused the mistake.  Does this person have a tendency towards overconfidence that they need to work on?  Were they afraid of looking bad?  Do they need to learn to consider others’ feelings more?

It’s vulnerable to be laid bare in this way, but, at least in the ideal of a DDO, everyone does it, from the interns to the CEO, to the point that people internalize that having flaws and a personal life is nothing to hide. Some people would find this horrifically intrusive, but others find it a relief.

I’ve never worked in a DDO, but I think I might like it; with enough mandated transparency, I’d be forced to override the temptation to hide flaws and make myself look better, and could focus better on actually doing good work.

The cost, of course, is way more communication about seemingly non-work-related things. You’d be processing personal stuff with coworkers all the time. The hope is that this is actually cheaper than the costs of the bad decisions made when you don’t have enough honest communication, but it’s an empirical matter whether that works out in practice, and the authors don’t have data so far.