Empire Theory, Part II: Power Dynamics

by | June 27, 2018

This is an excerpt from the draft of my upcoming book on great founder theory. Learn more here.

Continuing from Part 1: Competitive Landscape.

Power classes are a useful typology for players in an empire, because each group is subject to consistent incentives. As a result, there are consistent patterns of interaction between these groups. Understanding these patterns enables a deeper understanding of the strategic landscape and the crafting of superior strategy. In this essay we will explore these dynamics in detail.

The Dynamics of Power

Coordination and power go hand in hand. To understand both the opportunities for cooperation and under what conditions competition makes sense we have to take a look at key facts about power.

1. Power is a convergent instrumental good

Power can be used to accomplish a very broad range of goals. As such, many kinds of actors will aim to acquire power in the pursuit of their goals. The more effective they are and the better their understanding of reality is, the likelier they are to seek power.

There are two interesting consequences of this fact. First, those aligned on ultimate goals and values might still choose to compete over power, if they have different ideas as to how to achieve those goals.

Second, even those that aren’t aligned on ultimate aims can still choose to cooperate for a time to acquire power together. Those that accurately understand the instrumental value of power recognize each other and cooperate in ways that are not available to the less savvy.

To miss out on the usefulness of power is to miss out on a mechanism of coordination with the powerful, while failing to protect yourself from competition by the savvy.

2. Power is Pareto-distributed

The most powerful players are orders of magnitude more powerful than all other players. This distribution is observed in many, many domains vital to gaining and maintaining power, ranging from land ownership to income to political contacts to personal effectiveness.

3. The competitive nature of reality

Everyone is locked in a state of de-facto competition against all others trying to access the same scarce resources as they are (e.g. companies in Silicon Valley competing for talent). Power is a scarce resource, and, as noted above, it will be pursued by many actors. Thus pursuing power successfully can quickly result in reaching high levels of competitive difficulty.

4. The difficulty of coordination

Coordination is a troublesome problem. It takes a large amount of skill and resources to successfully coordinate large numbers of people. If you’ve ever tried to organize a group of volunteers or run a company you know just how true this is.

5. The insufficiency of inherited models

Society doesn’t equip people with correct ideas about how the social world works. A lot of political and social common sense is wrong or contradictory. For example, many people talk about decision making through consensus, but many people also say that committees are utterly ineffective. Inherited models are insufficient for effective action.

6. The deceptive side of society

Sometimes rather than merely being insufficient, the models people are equipped with are actively deceptive.

In most modern cultures vicious competition is not socially acceptable. There are carved out exceptions to this, such as in business or entry to prestigious educational institutions such as the Ivy League universities.

Even there, the competition is claimed to be limited to only a few domains. Further, the justification for these partial exceptions is prosocial and ultimately cooperative.

There are a few possible justifications for competition. One of them is the notion of a meritocratic society, one where positions of privilege are distributed in accordance with merit—that is, talent and skill. Everyone should be as excellent as they can be — ultimately competition is supposed to produce relative rankings for the distribution of positions, rather than an absolute standard.

In the example of elite universities, the justification is applied to admissions tests of various kinds. SAT scores and the like limit attendance at the universities to the talented, rather than using some other key such as say family ties.

Sometimes this prosocial story is correct and other times it isn’t. The deceptiveness of the societal story and the attempts to obscure competition are especially visible as low, mid, and high form secret alliances to attack other players and claim power for themselves.

The Dynamics of Power Classes


As we said in Empire Theory, Part 1: Competitive Landscape, high is the central power and cause of coordination in an empire.

High is generally concerned with maintaining its power in the empire; since high is already in the most powerful position, high has a lot to lose and less to gain locally. Due to its preoccupation with maintaining power, high will consistently be concerned about mid players growing strong enough to overthrow and replace high. As such, high will seek to control mid, usually through distribution or denial of resources.

High will also seek to expand its empire as a means of securing its position within the empire. High will seek both to increase the direct power imbalance between high and mid, as well as to acquire more resources in order to buy off certain mid players and play them against others. There is an important difference between resources high directly owns versus resources in the empire. While high can benefit from having powerful middle players with a lot of resources, high cannot directly use these resources. The total power of an empire is always larger than the power of high. High will try to steer growth with the priority of benefit to its internal position as the first priority; the overall growth of the empire is a secondary goal. Security and the ability to produce other kinds of effects in the world are usually not at a trade-off; when they are, however, security takes priority.


Mid is the group of players that can challenge the high power.

Mid will often fight with other mid players, both to destroy competitors and to add those mid players’ resources to their own empire. Mid will also often make alliances with high by specializing to perform services which high cannot or will not provide. Businesses, banks, and universities are good examples.

Mid players, in pursuit of increasing their own power, will be strongly incentivized to challenge high, since high has the most obvious concentration of resources. As such, mid needs to receive something very valuable from high in order to not challenge it. The tense interaction between mid and high is the most important thing to focus on when trying to understand an empire.


Low players can challenge mid players.

Low usually matters little as an independent force within an empire, although it will sometimes contest mid players. Instead, low is important because it will very often be used as a proxy by both mid and high players for their own purposes. As such, it will be commonplace to observe low powers being picked up and discarded by stronger powers. Low players will rarely demonstrate agency in their strategic moves.


Outside is the group that is not within high’s empire. Outside is composed of all empires and players outside of high’s zone of coordination. As such, outside will include competitors of high, as high will be competing with other empires for expansion.

Sometimes outside empires will invade and try to take over an empire in their quest for growth. These takeover attempts might include alliances with players inside the empire so as to subsume or disintegrate it. Mid powers are often interested in leaving empires, and might accept aid from outside to break off from high. Low powers might be interested in rising to mid in the new empire led by the former outside. A negotiated surrender is an example of such an alliance between high and outside.

It is possible to further divide outside into near and far. In the context of an empire, near can be considered as the direct competitors to the empire, who are primarily external competitors to high. In addition to active competitors, near will also include potential or likely competitors. Far can be considered as the outside players that are not direct competitors to high. This categorization is useful because near players will often try to undermine high by allying with mid players, and vice versa. High, in contrast, will be more interested in allying with far players against mid and near.

Outside can also aim for opportunistic collaboration to achieve a particular end without aiming to merge with their collaborator. An example would be the cooperation between the French company Sud Aviation and the British Aircraft Corporation to develop the world’s first supersonic passenger airline, the Concorde. The alliance is narrow, with the intent to produce a particular piece of technology.

The Dynamics of Interactions

The following sections will discuss all pairwise interactions between high, mid, low, and outside players.

In this discussion, there is an important distinction between degrees of cooperation. When two players are cooperating, they are working together to achieve a particular goal, but they are not necessarily generally aligned. Two players can cooperate in one domain while battling in a different domain. I call this a narrow alliance. When two players are coordinating to achieve most of their goals and no longer contest one another, I call it a broad alliance. Narrow alliances are the default between most players in an empire, whereas broad alliances are unusual.


High can be made up of many individuals. Each of these individuals will seek to expand their own power and increase the size of their personal empire. As we described in Part One, empires are fractal, and high is frequently best modeled as an empire in itself. High/high alliances will emerge when individual high players discern that the best way to grow their personal empire is if high can act in a unified manner. High can do things that no other player can do, because of the large pool of resources available to its members. As a result, there will often be especially large rewards for high acting in a unified manner. For example, in many countries, the only organization that can successfully execute large engineering projects is the central government, because they are the only group with sufficient resources and coordination power. The construction of the U.S. highway system beginning in the 1950s is an excellent example of this.

What does the unification of high look like? In considering the dynamics within high (when it is composed of multiple individuals), it can be useful to model high as an empire unto itself, yielding low high, mid high, and high high players. High is in a state of unification when high high and mid high are broadly allied. If high high and mid high are not broadly allied, then high is disunified.

High will tend to be unified when it has the ability and opportunity to expand its empire. In this circumstance, individual high players will perceive that the best way to grow their personal empires is to help the larger empire to expand. If these opportunities dry up, high will often become disunified, because the best strategy available to individual high players is to contest the other high players’ power. External threats to high are typically a subpar unifying force compared to the opportunity to expand. There is a nice story to be told about a dangerous external threat unifying a group of people, who then win against all odds; but more often in history, an external threat provides an opportunity for one high player or group of high players to win a local battle with another high player at the expense of the empire as a whole.

High disunity is especially problematic when considered in the context of the problem of local focus. When high is disunified, high players will contest each other’s personal empires. The focus of each high player will be the defense of his or her personal empire. In order to transition back to a unified high, the attention of high players needs to return to expansion of the broader empire. This transition can be very difficult to achieve, because all high players will need to simultaneously stop contesting each others’ empires such that their attention can focus on the larger empire. As a result, high disunity is an equilibrium that is extremely difficult to break out of. High unity, then, is unstable, because any outbreak of internal strife can lead to stable disunity.


Mid players usually gain by participating in the empire’s domain of coordination. For example, two dukes can resolve a border dispute by going to the king instead of having to resort to violent conflict. Similarly, national governments can enforce contracts for mid players in modern states. As seen in these two examples, when mid establishes a narrow alliance with high, high can resolve problems that are outside the reach of either mid player: high provides a coordination service.

Likewise, high gains from having mid players, because there are goals high cannot achieve without the cooperation of mid. For example, consider a startup in which the founder is the sole high. Since the founder’s time is scarce, he cannot personally manage each programmer once the company grows beyond a certain size. As such, he will cooperate with mid players (say, programming team managers) to manage the lower-level employees.

There is an interesting asymmetry in what has been described so far. The coordination services provided by high are insufficient, as they are merely making interaction with other mid players smoother. On the other hand, the delegation services provided by mid are frequently sufficient to justify the cost of the coordination service and more from the perspective of high. Providing arbitration and other means of coordination in exchange for delegation is almost always a worthwhile trade for high. Mid, however, appears incentivized to leave the empire and only opportunistically ally for such services when needed.

Absent further action from high, this incentive is often followed, leading to cascades of mid players leaving being one of the common causes of the downfall of empires.

Furthermore, since mid players will always seek to expand their personal empires and high has the most resources in the empire, mid successfully challenging high is among the most rewarding possible resource acquisition strategies.

To establish a broad alliance between high and a mid player, high must provide mid with something that both offsets the cost of delegation services as well as the temptation of seeking to challenge high. As a result, high will usually control the distribution of resources in an empire, thereby incentivizing mid players not to challenge high. For example, a central government can bribe mid players to not challenge it by distributing industrial contracts. A totalitarian state can coordinate mid players by giving them the opportunity not to be sent to a prison camp.

In general, if an empire is not expanding, broad alliances between high and mid will be fragile. A high player coordinating mid players primarily with threats will usually not be able to coordinate the mid players long-term.

Providing and denying opportunity are asymmetrical. You only have to occasionally provide positive opportunities for collaboration to be worthwhile. If you are merely denying opportunities to force cooperation, you have to carry this out always. For example, a CEO that is constantly threatening to fire his managers due to the company’s poor performance will not be able to stably coordinate those managers. It would be much better for the CEO to set up incentives such that all the managers want to stay on to grow the pie and get a piece of it. As such, the most stable high and mid broad alliance is one in which mid is receiving resources from high (e.g. colonies, subsidies, commissions, etc.). High can give its own resources to mid in exchange for cooperation, or high can get resources from outside the empire and give some of these to mid. A somewhat ingenious high can even create resources from nothing using superior knowledge. An example might be the British honor system, with an endless number of titles to receive and orders to be knighted into. Social resources can be created de novo by high and then distributed by high. This type of coordination is limited by the skill and knowledge of high, so it is not an infinitely usable hack.

The strategy of resource distribution is much more stable than the strategy of threats, as it allows high to maintain the relative distribution of resources to high’s advantage, while the former does not. For high to stably distribute resources from the outside, however, the empire must be expanding.

High and mid achieving broad alliances, like those described above, is important for handling the problem of local focus. If both high and mid players do not need to focus on defense of their personal empires against adversaries within the broader empire, more effort can be put into expansion of the empire. An allied high and mid is an extremely effective internal structure for empire expansion.

High and mid can also ally to attack other mid players. High will often narrowly ally with a mid player to attack a more threatening mid player. For example, consider a university in which an influential tenured professor is rallying other professors to question the budget decisions of the administration.

The administration can ally with a different set of professors, who will usually be weaker or less politically savvy, to challenge the original professor. The professors allying with the administration can get pay increases, promotions, desired policy changes, or departmental budget increases in return for their cooperation.

If high is undertaking such an alliance, we can infer that it is already notably weakened. After all, it chose a mid player rather than a low player, which already means it required or desired the assistance of someone well-positioned. We can predict that the alliance will be short lived as the mid player might in turn become threatening.

Frequent alliances like this are not a good sign for an empire. It means that, for some reason or another, high is chronically finding difficulty in aligning with mid powers. It suggests that the only means available to it to preserve its domain, is undermining the powerful members of this domain, rather than, for example, distributing external resources to mid to preserve high’s power and mid’s loyalty. The limit of the empire’s power has been reached.

Finally, high will sometimes scrap mid players to add their resources to those under high’s direct control. We have previously mentioned the important difference between resources that are at high’s direct disposal versus resources that are in direct control of other players in the empire. One way high can increase the amount of resources at its direct disposal is to take a mid player’s resources. For example, a government can nationalize a particular industry as a legally held monopoly.


As we have previously said, low players are mostly irrelevant to high players. They don’t have enough power to effectively attack high, and they don’t have enough resources to be worth scrapping. They are also more difficult to usefully coordinate with than mid or outside players. Since they are individually weak, a large number of them must be coordinated in order to make it worthwhile. Coordinating such large numbers can be prohibitively difficult. For example, if the CEO of a tech company is working to launch a big new feature, it is much easier for him to work with three lieutenants to manage the project than manage fifty programmers himself.

Given the difficulty of usefully coordinating low players, why would high ally with low? High will ally with low because low can be weaponized against high’s adversaries. A common offensive move for high is to ally with a low player to attack a mid player. Low players are strong enough to attack mid players but are not strong enough to be dangerous to high, making this alliance very safe for high. For example, say the CEO of a tech startup wants to get rid of one of his managers but doesn’t have sufficient legal ground to fire them. The CEO could ally with one of the lower-level programmers managed by this person who has been doing poorly on recent work performance reviews. The programmer is tasked with filing a harassment complaint against the manager with HR in exchange for leniency in work reviews.

There are two important observations about this common type of offensive alliance. First, it helps explain the seemingly irrational paranoia that can be found among strategically savvy individuals. Attacks by powerful players will often appear to be random harassment by low players. Second, all alliances between high and low are very asymmetrical. Since low cannot challenge high, the relationship is almost completely in high’s control. The low player is disposable in high/low alliances, something important to keep in mind if engaged in an alliance with high as a low player.

High will also often ally with low players to avoid empowering mid players. For example, say the president of a university has to choose a professor each year to give a speech in front of the entire school. The president may pick an obscure professor so as to avoid giving a notable and powerful professor, a mid player, resources (in this case, public acclaim), since the president considers such professors a threat to his influence over the university. High/low alliances can appear extremely puzzling, because it will seem like high either has poor judgment or is wasting time with low players. In reality, though, it may be a prudent maneuver against mid.

It is useful to be aware of high’s predisposition to ally with low if you are a low player within an empire. Low players can position themselves to ally with high in order to destroy a mid player and achieve mutually beneficial aims.

“Grassroots” movements are an example of this. Take, for example, the Little Rock Nine. After the historic Brown v. Board of Education Supreme Court case in which the racial segregation of schools was declared unconstitutional, the governor of Arkansas deployed the Arkansas National Guard to physically prevent black students from attending previously all-white schools.

In response, President Eisenhower nationalized control of the Arkansas National Guard and sent the 101st Airborne Division to enforce the racial integration of the schools. One way of describing this event is that the grassroots civil rights movement won a major victory against segregationists. An alternate description is that high (the federal Government) took a resource (the Arkansas National Guard) from a mid player (the governor of Arkansas), using a conflict between low and mid which had been incited by high (the desegregation of schools, incited by the Supreme Court verdict) as justification.


There are four major ways in which high interacts with outside players. First, high can attack them to expand and gain resources. Second, high can use them to fight internal political battles. Third, high’s empire can be invaded by them. Fourth, high can ally with them to attack other outside players. 

High is incentivized to expand the empire as a means of increasing its own power and as a means of coordinating mid players through the dispersal of resources. Sometimes high will expand by acquiring an outside empire. Consider Google acquiring a startup. Google will often acquire a start-up because there is something that the outside empire can do which it cannot do (similar to how mid players specialize to coordinate with high). When an empire is acquired, it usually retains its original structure and some power, but becomes coordinated by and subordinate to high. In this case, the acquired company might maintain its internal structure and some powers like hiring, but what it produces will be owned by Google. Acquisition can also be less cooperative, like military conquest, for example.

Similar to how high can ally with low or mid players to defeat opponents in the empire, high can also ally with the outside to defeat its internal opponents. Take a tech startup in which the CTO and CEO disagree about strategy and the board is split on which to support. The CEO might hire a prestigious, supposedly unbiased consulting firm to rubber-stamp his decision in the hope of swinging the board. Another example is the hiring of foreign mercenaries by rulers to quell local rebellions. These sorts of alliances are basically always narrow alliances. It will rarely be the case that high and outside enter a broad alliance. These types of outside players are also likely to be far.

High inviting outside players into the empire carries a significant risk: the outside players might turn on high. This situation is particularly dangerous; outside players will learn a lot about high and the rest of the empire when they are invited in, because they need that information in order to coordinate with high. However, high won’t necessarily learn very much about the invited player. This information asymmetry can be extremely dangerous for high—it turns a far outside player into a near outside player.

For example, in 1169, the King of Leinster invited Norman mercenaries to help settle a rebellion in his kingdom. Instead, the Norman mercenaries ended up seizing the territory for themselves, deposing the king. When inviting players from the outside, it is easy to misjudge their power due a lack of information about that player. Even a single, highly persuasive individual can be dangerous to invite into an empire if he or she cannot be controlled.

Just as high can attack other empires to gain resources, other empires can attack high’s empire. Because there is intense competition for power, outside attacks are common and empires must defend against them. Competition in a market is one example. If your company locates a previously unserved market, you shouldn’t expect to be alone for long if you see any success. Other companies will soon seek to chip away at your empire. For example, Apple’s success with the iPhone rapidly led to many copycat competitors like the Samsung Galaxy.

High can supplement its strength against an external enemy by building a narrow alliance with a third player from the outside. The strategy is particularly apt when the aim is defeating an external empire rather than acquiring its resources; because high needs to spend fewer of its resources to acquire an outcome, and isn’t concerned about possible spoils, the ally can then be paid from the spoils. Successive Chinese dynasties relied on this policy heavily over the centuries, to the point of it being artfully captured in an idiom: “use foreigners to subdue foreigners; let the barbarians fight it out among themselves” (以夷制夷).


Mid players will often behave antagonistically towards one another because other mid players are their primary competitors for gaining power. For example, the U.S. government often offers competitive contracts for construction projects. Mid players (large construction companies) will have to battle one another for the contract. That said, there are two ways in which mid players will sometimes coordinate. First, mid players will ally to create an anti-high coalition. This is the only common mid/mid broad alliance. Second, mid players will narrowly ally to attack other mid players.

There are four common types of anti-high coalitions: conservative coalitions, coup coalitions, secession coalitions, and dissolution coalitions.

A conservative coalition is when mid players coordinate to oppose the actions of high in an empire. For example, if the federal government is trying to pass a law curtailing the power of state governments, state governors might ally to oppose the legislation. If the CEO of a startup tries to push for the adoption of particular code testing policies, the engineering team leaders might collectively reject the CEO’s policy. In both cases, the mid coalition may succeed; conservative coalitions can block attempted changes by high, but will often succeed only at slowing high rather than halting them altogether.

A coup coalition is an alliance in which mid players coordinate to depose high with the aim of having the group become high themselves. A classic example is when a king’s ministers depose the king and install a patsy as the new king. When this sort of transition happens, the empire will usually remain intact but with a new high. Such coalitions are most viable when a small number of mid players are notably more powerful than the rest. When this isn’t the case, the new high will not have sufficient advantage to keep the empire intact.

The third anti-high coalition is the secession coalition. Mid players will often have their own empires within the larger empire. If the benefits of being coordinated by the high power are not worth the costs, then mid powers will be incentivized to exit the empire. Sometimes mid players will simply leave the empire, although frequently this move will be blocked by coordination mechanisms, by which we mean social technology that incentivizes coordination and usually doesn’t require live players to pilot. Examples would be military force in the context of a local government breaking off from a national government, or social pressure in the context of a manager leaving a tech startup. In these cases, mid players can ally to aid each other in breaking away from the empire. The American Civil War is a classic example of this.

The fourth anti-high coalition is the dissolution coalition. Sometimes, instead of mid players coordinating to leave the empire, they will simply destroy the empire. If a collection of state governments collaborates to destroy the national government, then sovereignty will devolve the individual states. This process drove the collapse of the Soviet Union, with the individual Soviet Republics cooperating to reduce the legal and political role of the union, and eventually helping legitimize its dissolution as well.

The risk of mids creating a dissolution coalition provides a strong motivation for high to distribute valuable resources to mid players so as to make the empire’s continued existence preferable to them. Such distribution is viable if high’s domination of resources persists, through some kind of growth; otherwise high is merely giving away its own advantage. Such generosity might slow down a particular dissolution attempt, but will make success more likely when dissolution is attempted.

Since redistributing resources to the most powerful mid player trades off against the risk of them initiating a coup, buying off weaker participants in a dissolution coalition first staves off dissolution without increasing the risk of a coup.

These scenarios often preoccupy high’s attention and determine what actions are viable. Only the largest and most skilled mid players can fruitfully pursue them. In most circumstances continued cooperation with high is the best option.

An easier and more common option for cooperation among mid players is that of joining together against other mid players. Mid players compete for power, since they benefit from influence over the commons and possible allies in the empire. As a result, it is sometimes viable to create a narrow alliance to defeat a particular mutual mid competitor. Fewer strong competitors means more resources available for the remaining players.


There are four main ways in which mid and low players interact.

First, low players can be weaponized by high to attack mid. See the section on high-mid dynamics for a discussion of this.

Second, low players can be weaponized by mid players against mid opponents. This takes the form of a low player supporting or protecting a low player that is frustrating their common opponent. One might step into an existing conflict of interest between the two and support the weaker side to prolong it, or one can even incite the conflict to begin with.

Third, mid players will sometimes ally with low players in order to expand their own empires. For example, the manager of a team of programmers might notice a talented programmer on another team. The manager could befriend that programmer and convince him to join her team as a means of improving her own team’s performance.

Fourth, low players will sometimes aim to ally with mid players in hopes of becoming mid players themselves. This alliance usually occurs either by low riding on mid’s coattails as mid increases in power or by low directly gaining power through their alliance with mid. A good example of coattail-riding occurs in U.S. presidential elections. Campaign staffers (low) ally with a presidential candidate (mid) in the hopes that that mid player will win the election and then become high. If this occurs, the new president will repay the campaign staffers by delivering White House appointments, making the staffers mid players. An example of low directly gaining power through a mid alliance is mentorship. A mid player invests in a low player in the hopes that the low player becomes a tightly coordinated mid player.


Interactions between mid and outside are often tense, because it is risky for mid players to interact with outside players. Mid will primarily interact with outside in two contexts: when outside is attacking their empire and when mid is going outside of the empire for resources. In both of these cases, mid is more likely to interact with near than far.

Aggressive outside empires will often try to ally with mid players in an empire they are invading. Mid players can be extremely valuable to an invading empire because they will often have useful information on the target empire. Also, stealing them both increases the invader’s power and decreases the target’s power. For example, consider two website-builder tech startups competing with one another. It is very useful for one company to steal a highly skilled manager from the other company, because it gains a highly skilled manager, the opposing company loses a highly skilled manager, and the manager brings with her detailed knowledge of the opposing company’s strategy and internal dynamics. Due to the damage defection can cause, punishments are usually harsh. In the context of competing states, treason is punishable by death. Defectors are usually completely socially ostracized after being discovered. Even between competing companies, defecting to the opposing company will often result in total social ostracism from the first company. Defection of mid players is a rare and destructive event.

After an empire has conquered another empire, they will attempt to ally with the conquered mid players in order to preserve the basic working order of that empire. Much of the value of an empire comes from the local players’ ability to coordinate with one another. Setting up the structures necessary for effective coordination is very difficult. As such, when an empire is conquered, the conquering empire will often simply reuse the coordination structures that have already been set up by the previous leaders of that empire. Mid players are also incentivized to ally with the new regime, as the alternative is usually destruction—although, sometimes mid players will attempt to break off from the empire during the chaotic period of high’s replacement. This pattern of reusing existing coordination structures leads to such structures being surprisingly durable, usually lasting far longer than any single empire. Examples might include the Roman Catholic Church, which outlasted the Roman Empire, or the trade network of the Silk Road that outlasted the Mongol Empire. We might imagine the strategic landscape of history as a huge number of unimportant lows, dotted by a smaller number of important mids, who are constantly being recombined by competing highs in new empires.

Venturing outside the empire is an interesting challenge for the mid player. In a space of many somewhat coordinated players, it is ideal to achieve growth with the help of said players rather than going against their designs. Opposition is costly.

In the British Empire of the 18th and even 19th century, great fortunes and energies could be absorbed by political struggles in the capital. However, one of the best routes for influence in London was making your fortune by expanding Britain’s colonial holdings and then bringing that capital to bear. The returns were often better than fighting in the system. Examples of this were the military career of Sir Robert Clive, who conquered Bengal for the British East India Company, and the business ventures of Cecil Rhodes that drove expansion into Africa.

When high is coordinating mid players and distributing patronage from the common effort to grow the empire, there are few reasons for mid to pursue additional projects. Alexander the Great’s generals are best served by staying with his army and carrying out his orders; their prospects for wealth and fame against a still-standing Persian empire were miniscule. In such contexts a high player is staking their position on their ability to continue providing patronage rather than on the ability to defeat mid players.

In this example, Alexander demonstrates the ability to win battles against the Persian empire and acquire more and more provinces. He is overwhelmingly incentivized to maintain such growth. He takes on most of the cost of failure, but will share in the spoils of his success. An independent venture by a mid player means they are spending their own resources and also directly bearing the risk of failure.

Given these expenditures and risks, mid players should pursue outside growth when high is not offering sufficient resources for the mid players’ growth. An even more dire circumstance is when outside expansion is attempted to circumvent a high power actively trying to starve a large mid of resources.


Low is generally unimportant except for when being used by mid and high. As such, low/low interactions are mostly unimportant at the empire level. That said, there is one circumstance worth mentioning. A low player will sometimes assemble a cluster of other low players into a local empire, making the organizing low power a new mid power. The famous slave rebellion of Spartacus in ancient Rome serves as an example. For a more recent example, consider the situation in which a town is passing new zoning laws setting a minimum size for plots of land in a county. Low-income residents of the county would be hurt by this law, because plots of land would be notably more expensive if they could not be further subdivided. One low-income resident might rally other low income residents to fight the zoning law, with the organizer becoming the group’s leader. In this case, the organizer has suddenly risen from low to mid by coordinating low players using a new coordination mechanism (the personal incentive of low-income homes to oppose the zoning law). As the primary difficulty among low players is the cost of coordination, it is common to see the creation of new mid players when the strategic landscape changes and there are newly available coordination mechanisms for low players.


Low/outside dynamics are usually unimportant, but there are a few worth mentioning. First, low will sometimes coordinate with an invading empire by being weaponized against mid players—after the conquest they may coordinate with the new high. Second, low will sometimes leave the empire. It will usually be easier for low players to leave the empire than mid players, because an empire losing a mid player is both costlier and riskier than losing a low player, so the coordination mechanisms tend to be weaker in the case of low players.

Growth and Decay in Empires

We can use our understanding of the dynamics of power classes to determine whether empires are healthy. Since power is always at least somewhat insecure, there is always a need to import resources previously not in the empire, even if only to maintain the status quo.

When thinking about coordinated groups (i.e. empires), health and growth are synonyms. How is this growth achieved, or how is the scarcity managed? Depending on how centralized the empire is, these dynamics of growth and decay play out differently.

For any empire, we can ask how centralized it is: to what degree is high coordinating, and coordinated with, the rest of the empire, specifically mid? Though the level of centralization is a continuum, we can draw a line somewhere in the middle and say that an empire on one side is centralized, and an empire on the other is decentralized. We can also ask whether the empire is expanding—that is, gaining resources from outside—or declining. Combining the answer to both questions yields four empire types. Typing empires in this way allows us to rapidly understand the basic internal dynamics of a given empire.

Centralized expanding empire

In a centralized expanding empire, the central power (i.e. high) is broadly allied with the middle powers, often by buying them off with resources acquired from outside of the empire. The coordination thus bought is then directed towards keeping the empire growing. Growth can take the form of captured provinces, new trade routes, acquired competitors, new technologies, and so on. Since high is driving the growth, the empire tends to expand decisively in one direction at a time. This type of empire can usually be discerned by its decisive manner of expansion. An example of a political entity of this kind is the Ottoman empire, where the Sultan’s forces either focused their campaigning in Anatolia and the Middle East or in the Balkans and Europe, never both at the same time. 

Centralized declining empire

In a centralized declining empire, the central power is keeping the mid players coordinated by denying them resources and preventing them from acquiring resources from outside. Such an empire will either shrink gradually or suddenly and catastrophically implode. This type of empire can be discerned by observing a tightly coordinated empire that has shrunk over time, but hasn’t had any major parts of the empire break off and become independent. Few civilizations embodied this kind of decline as well as the Western Roman Empire in the fourth century AD. The extraction by high was not only evident in the high taxation of the population, but in how the rise of successful generals such as Flavius Aetius was seen as a threat rather than an asset by their emperors: often they were assassinated or executed.  

Decentralized expanding empire

In a decentralized expanding empire, the central power isn’t strong enough to prevent middle powers from going outside for resources. High maintains its position by acquiring resources for its direct control from the outside without the help of middle powers and by occasionally scrapping weaker mid players. In this state, the empire is growing. It grows in multiple directions in a patchy manner, due to the unsynchronized actions of mid and high. This type of empire can be discerned by its multi-directional expansion pattern. In the 18th century British Empire, many lower elites such as Robert Clive of the British East India Company made their fortunes and proved the ability to command in colonial ventures around the world, but eventually returned to London to gain honors and networks only accessible there.

Decentralized declining empire

In a decentralized declining empire, the central power is failing and isn’t strong enough to keep middle powers coordinated. In particular, it isn’t strong enough to prevent their growth. The empire is fragmenting, with no clear successor to the dying high. This type of empire can be discerned by observing an empire that is shrinking and has significant parts breaking off and becoming independent. The Zhou Dynasty of ancient China maintained ritual importance long after it lost effective control; as coordination between their nominal vassals who each grew more powerful broke down so too did the unity of China, giving way to the Warring States Period, where these states would contest each other seemingly without limit.

Empires need growth

The landscapes of power and coordination are intimately interwoven. Patterns of alliances and rivalries come through necessity to define who we coordinate with and why. No matter what goals we pursue, we face this reality of power, and so must understand and account for it.

We explored a classification system for the power of individuals and institutions, pairing it with an overview of the dynamics that play out between these classes. This analysis allows us to take several important steps. We can diagnose the current landscape of an institution and the state of coordination dynamics within it, and predict with reasonable confidence the effects of various actions and strategies on a given empire. Finally, we can track relations between the powers in an empire to accurately predict where in the life cycle of institutions the empire falls.

The importance of growth for the health of empires stands out. In the analysis, growth seems indispensable for harmonizing the interests of relevant high and mid stakeholders. Empires coordinated through cooperative ventures—by carrot rather than stick—will plateau and decay later than those coordinated by coercion, translating into more coordinated allies and resources.

The best way to win at adversarial encounters, then, is to focus energy on building out cooperative ones. In the long run, acquiring power and empowering others is mutually reinforcing rather than mutually exclusive. Something to keep in mind.