The Twilight of Annexation

The Twilight of Annexation


For the most part, the history of the world is a history of states struggling against each other. From the Punic Wars to the partitions of Poland to WWII, this struggle has usually manifested as states seeking to annex territories from each other. Historically, annexation is a virtual constant — it was the basic currency of war and politics. This is why it is particularly strange that the Russian annexation of Crimea was such a shocking, norm-violating event; since when did annexation become unexpected?

Indeed, this presents us with a very substantial modern historical oddity: Very few states still try to directly annex territory. Many have argued that synthetic changes imposed on the anarchic international order have, for now, blunted the basic annexationist nature of states. Some have argued that America’s current global hegemony holds together a fragile international peace; others have argued that pacts like Kellogg-Briand or institutions like the UN check rampant expansionism. More contend that annexation has dropped off as a result of pure, hard-nosed power balancing.

Yet few have stopped to ask if there is anything internal to the recent precipitous decrease in annexation. Most assume that states have roughly the same broad interests that they did centuries ago: that they seek to survive, and that territorial expansion is a means of ensuring survival. External forces like hegemony, power balancing, treaties, or institutions provide an alternative survival method. I contend that these are not likely the primary reason that annexation has virtually vanished from the world stage. Annexation’s twilight is internal to states’ survival incentives, not external.

I argue that annexation has dwindled specifically because the nature of both states and warfare have changed drastically over the past two centuries. In the past, regime security relied directly on territorial expansion and annexation. However, modern nation-states rarely need to annex for security purposes anymore, as security is now a product of technological advantage, which has little relation to control over land.  Further, the idea of the nation has introduced significant new security disincentives to annexation, in most cases. Finally, there are robust explanations for why non-security-related annexations have declined, but are still extant in a substantially reduced capacity. In other words: annexation still exists, but only as edge cases operating under a fundamentally different logic, when compared to the rest of human history. The phenomenon of annexation is dying.

Part 1: Empires of Land

Before asking why annexation is vanishing, one should ask why annexation existed in the first place. Why would annexing part of another country achieve a country’s ends? For most of history, the answer is relatively simple: territory was the primary determinant of military power, and thus the determinant of security from outside threats. There are three reasons why land was the ultimate determinant for pre-modern[1] state security: wealth, men, and mobility.

Before we dive in, I would like to preface this first section by noting that this is going to sound rather reductionist. This is primarily because it is, in fact, rather reductionist. I would ask that the reader bear with me. I will largely ignore a great deal of nuance and complexity in pre-modern military tactics in order to outline why the primacy of land in external security was, in general, the case.

First, more arable land meant more revenue for the state, and more wealth within the state. For much of human history, regional GDP has been a virtually 1:1 correlation with productive farming[2] land. Urban centers represented a relatively tiny share of total human economic activity and wealth generation. In short, the more arable land one controlled, the more swords, cannons, mercenaries, and ships a lord and his lieges could purchase.

Second, more arable land meant more people. The more crops you could grow, the more people (potential soldiers) could be supported. In most cases, a larger mass army would historically usually win out against a smaller mass army. Similarly, a larger well-trained army would usually win out against a smaller well-trained army. The capacity to maintain a well-trained standing army was constrained by regime wealth, which, as noted above, was largely a function of land area.

Third, in the pre-modern era, mobility was largely limited to the marching speed of an army of foot soldiers.[3] The more land a state controlled, the more of a buffer it could maintain between its capital and its border with a foreign power. The more distance between a frontier and a capital, the more chances a state had to fend off a relatively slow-moving invader with extended supply lines headed for the seat of governmental authority. This is the same logic that dictates the maintenance of buffer states.

At the end of the day, two things went into the fighting effectiveness of an army: the number of men, and the amount of capital they are armed with. Both wealth and manpower were primarily a function of land, and rarely a function of anything else.[4] Therefore, improving a state’s external security was, in general, an endeavor to annex more land.

In a system where external security is a function of acreage of arable land, security is very clearly a zero-sum game: Any gain in land for one state must be taken from a different state. In this world, annexation pays double returns. Not only does a state improve its own security by annexing land from its neighbors; it also weakens the threat posed by said neighbors, as they now have less land to generate threatening troops and capital. When acreage is directly connected to external security in a zero-sum game, we would indeed expect everyone to constantly try to annex everyone else, where possible.

Part 2: What Changed?

So, if the precipitous decline in annexation is internal to a rational state’s changing security incentives, as I claim, what changed the incentives? In short, external security has shifted towards having the best, most technologically advanced weapons. How much land a state controls is mostly irrelevant in obtaining security through technological advantage.

One might ask, isn’t having more land still beneficial for security? It is, but, only in very particular special cases. These include access to warm-water ports, control of unique strategic geographic positions, control of water sources or other vital resources, and cases where a state’s capital is in artillery range of a potentially hostile power.[5]

The importance of these security-interest cases should not be underrated. Quite a few large wars could start over strategic control of the aforementioned concerns. However, those cases are nonetheless limited. They do not detract from the overall major shift in incentives for general annexation.

Outside of these cases, controlling more land is virtually useless, for much the same reasons that controlling it was once essential. In the modern era, the vast majority of virtually any given state’s population lives in cities, and a vast majority of a state’s wealth is generated in cities as well. Agriculture rarely makes up a double-digit share of a modern state’s economy, and it never does so amongst powerful industrialized nations. States no longer need much arable land to generate people to arm, and to generate the capital to arm them with.

Further, security is not much a matter of arming manpower with capital anymore. Multiple innovations in military technology have vastly reduced the importance of having more manpower in a great power conflict. Specifically: mechanized warfare, air power, and nuclear weapons all have helped to dissociate military strength from troop numbers.[6]

Mechanized warfare placed cavalry units (tanks and the like) at the forefront of ground combat, each of which is only operated by a handful of men. Air power took this much further, by enabling massive amounts of explosive yield to be delivered by a single man inside a jet fighter (or in an air base, remotely piloting a drone). A true modern war between great powers would likely be fought primarily with capital, which is generated at every level by comparatively land-efficient cities.

Nuclear weapons upended the calculus even further. If a state can obtain nuclear weapons, it has the ultimate deterrent to conventional war. No amount of increased land would protect an aggressor from a retaliatory nuclear strike, and no amount of increased land would improve the security of a nuclear-armed state beyond what it already has through nuclear weapons.[7] The incentive to annex is virtually non-existent from a rational security maximization standpoint when nuclear weapons are involved.

Today, the only plausible way to maximize one’s security is to obtain the most cutting-edge technology. The second half of the 20th century is replete with cases where a moderate technological edge gave an overwhelming advantage to one side. The Falklands War and the First Gulf War are both strong examples.

Generally speaking, having a massive amount of inferior technology staffed by a massive army is relatively useless in the face of a moderate technological edge. Even if one has the most extensive air defense network in the world, if said network cannot detect or hit an F-22, then four well-supplied F-22’s operated by four men can easily eradicate an air defense network comprising thousands of troops, given enough time.

Gaining such an edge comes from either having a robust military-industrial complex dedicated to developing new weapons of war, or being closely allied with a country that has one, such that they will sell said new weapons. Having the first is a product of capital investment, generated from economic activity in compact urban spaces. Having the second is a product of diplomacy and power-balancing. Neither involve control of arable land in any real respect. In general, maximizing for security no longer has much of anything to do with annexing and controlling more land.

Part 3: The Nation Problem

So far, we’ve established that states should be indifferent to acquiring more land, as it offers far less of an advantage than  obtaining a technological edge. However, states do not simply lack incentive to annex; currently, they have substantial disincentives to do so. In short, nationalism has made annexation a very dangerous prospect for the extremely low return I just established above.

Through the 19th century, the concept of the nation spread worldwide. People began to identify more strongly with groups that transcended old imperial boundaries. Simultaneously, the idea of the nation-state – that a nation should have its own self-governing state – helped destabilize the concept of empires of annexation. When a large group of people all believe that they should rule themselves, annexing a part of that group began to run a much higher risk of starting an extremely costly insurgency. Given the low returns on land, and the elevated risk that the people already on the land are willing to die for self-governance, a rational military planner would likely avoid annexation in the vast majority of cases.

It is also worth noting that the concept of the nation-state is predicated on having a government primarily for the nation, and at best secondarily for anyone else. The idea is definitionally exclusionary – and so, many nations are averse to expanding membership in their state by including people of other nationalities. The concept of the nation subtly asserts that only nation-members really belong within the state. Annexation is both increasingly risky due to the idea of the nation increasing the risk of insurgency, and it is increasingly distasteful to nation-states which seek to exclude non-members of the nation from the state.

The rise of the national ideal, however, has also added one new(ish)[8] reason for annexation. A nation with a share of its peoples ruled by another, different nation-state often has strong political incentives to annex the territory in which this share of people live. This is (part of)[9] the logic behind the annexation of Crimea, as well as Chinese ambitions for Taiwan. Thankfully, these ambitions are strictly geographically bounded, for the reason mentioned in the previous paragraph. A revisionist nation-state may make great efforts to include all members of the nation, by force if necessary, but will likely make similar efforts to keep non-members out. In other words, the rationale of Russia for annexing a Russian-speaking segment of Ukraine would certainly not hold for somewhere like Poland, without a sudden re-emergence of violently revisionist pan-Slavism.

Part 4: Why else annex?

Finally, we should ask: are there other internal reasons why states would try to annex each other, beyond security motives? In a purely realist model, not really; the pole star for every state in such a system is external state survival. If the security motive of annexation were to disappear, we would expect annexation to disappear as well.

Outside of realism 101, there are some plausible factors aside from security and wealth that could push annexation. Glory and honor is one; a deeply militaristic state could genuinely seek to annex other states’ territory, independent of any security motive. Relatedly, if a country’s political structure benefits leaders with glorious military victories, leaders could seek such victories as a means of preserving their regime.

Ideology is another major explanation. As mentioned above, people are quite willing to go to war over how they believe “our people” should be ruled. Postnational ideologies such as Leninist communism, Iran’s early-80’s revolutionary Islamism, and American neoconservatism advocate for war over how all other people should be ruled.

These factors challenge the theory that the decline in annexation is internal, and due to a change in security incentives. If we’ve seen a major, global decline in annexation attempts, then might there be an external force suppressing annexation from all internal motives? Why would we see a roughly concurrent decline in ideological or militaristic annexation?

First, on militarism. I would argue that we should not actually expect to see much purely militarist expansionism in general, and that we have indeed seen about as many attempts at annexation from militarist powers as we would expect in the past few decades.  Interestingly, authoritarian states face a paradoxical trade-off: leader-glory may be politically useful, but actually having a military competent enough to annex someone else’s territory is extremely dangerous.

Both Talmadge’s The Dictator’s Army[10] and Mesquita & Smith’s The Dictator’s Handbook[11] note that authoritarian regimes often actively seek to make their militaries less effective in battlefield combat. They do this because a military organized based on effectiveness, rather than loyalty, is a constant coup risk. A coup from a well-organized military poses a larger threat to an authoritarian regime that outside attack, especially when the incentives for attacking and annexing land in a regime are minimal in the modern era. Further, a general who actually wins in the field is a threat to an authoritarian leader’s power, as demonstrated repeatedly by Stalin’s purges of his generals before, during, and after WWII.

Therefore, it is rather unsurprising that few militaristic states have actually attempted to annex extra land, given the threat such efforts pose to a dictator’s goal of keeping his head attached to his neck. We have observed a few instances of fascist countries attempting to do so in the postwar era of mechanized and aerial warfare, primarily from Ba’athist Iraq in the 80’s and early 90’s. The practice is not completely absent, and there is substantial overlap in the motivations for these efforts with the motivations for annexing peoples ostensibly of the same nation under foreign control (read: Crimea, Taiwan, etc.). Nonetheless, the overall lack of militarist annexation seems to also be a result of internal factors.

Second, on ideology. In short, the decline in expansionist ideological annexation is purely incidental, and due to a (likely temporary) lack of a truly ideologically revisionist great power on the world state. Throughout much of the 20th century, a variety of different powers intervened forcibly in the internal politics of quite possibly a numeric majority of the countries in the world. These powers intervened both for the sake of geopolitical dominance over their rivals, as well as due to strong belief about how people of other countries should be ruled. The most extreme example of this was Nazi Germany; it did not simply want to change the way other people are ruled, but actively believed that genocide of people outside German borders, and thus military annexation, was necessary to their ideological project.

Today, there seem to be very few countries left who are interested enough in the type of government in other countries to engage militarily. Such a list may well only include the United States, if that. China is not backing Maoist insurgencies anywhere, Russia is not funding militant Marxist-Leninists, and Angela Merkel is certainly not occupying Poland anytime soon. Even the hard-right regimes in Poland and Hungary have shown zero indication of seeking to determine how people outside their respective countries are governed.

However, there is little reason to believe that such a regime could not emerge, due to either broad internal or external reasons. I cannot prove the negative claim, “there is nothing preventing the emergence of expansionist ideological powers”, in much the same way that one cannot disprove the existence of Russell’s teapot, but I am not aware of any plausible reason to believe as much. I suspect that no one today believes that history ended in 1991.

Part 5: Conclusion

In conclusion, there are robust explanations about why other reasons for annexation have not seemed to cause much annexation lately. Further, we have equally robust explanations for why the annexations that have happened, happened, that are consistent with our theory that a rational security maximizer will very rarely seek to annex more land. On the whole, a rational state will likely not seek to annex, both because doing so usually will not increase security, and because doing so will likely decrease security, if part of a different nation-group already lives there.

What does this mean for policymakers? First, policymakers should note that projects like Chinese neocolonialism likely do not seek to outright annex territory. Rival states do not need to be as aggressive as in the past in order to ensure their security. Second, policymakers should understand that the rise of one country does not mean that they have any reason whatsoever to threaten most other countries compared to when annexation was necessary. This is especially true in the case of America; there is no conceivable reason that a power like China could have to directly threaten America itself.

Third, policymakers should examine what their rational security interests actually are. The foreign policy establishment seems to operate on the belief that we must compete for every square inch of land on earth in order to be secure in America. In reality, two oceans and a second-strike-capable nuclear force likely maximizes actual American security entirely on its own.

The belief that we must maintain hegemony everywhere is a holdover from a geopolitics where annexation was a constant and concrete threat. The re-emergence of revisionist ideological powers, as noted, could change this. However, policymakers must acknowledge that it is a) entirely possible that we could live in a world where annexation is not a material threat, and b) that we live in such a world.  The persisting belief that we live in a system where control over land determines security could cause completely unnecessary wars entirely on its own. If policymakers incorrectly believe that a rising China will rationally seek to threaten American security at some point, we may fall into the much-vaunted Thucydides Trap and seek to check their expansion for literally no genuinely rational strategic reason.

For all these reasons, policymakers must re-evaluate what security means in a world where land control only rarely impacts a state’s prospects of preservation and security.

Follow Davis on Twitter at @lark_notes.


[1] Take “pre-modern” to mean, roughly speaking, prior to the 19th century. This definition is arbitrary, and used only as a convenient informal shorthand.

[2] Or pastoral. The same logic holds for these cases: more land meant more taxable economic activity.

[3] The steppe tribes of Central Asia sometimes produced confederations with overwhelming amounts of mounted troops; however, for our purposes, they represent an occasional edge case. Most pre-modern armies were constrained by the maximum speed of foot soldiers.

[4] One exception: mercantile trading empires. These were relatively rare. In such a case, a state could bypass the need for wealth via agriculture on arable land, and primarily focus on controlling strategic trading lanes to obtain the wealth needed to arm men (or buy mercenaries). These sorts of empires generally manifested only around aquatic trading lanes, of which there are certainly not enough of to change the incentive for annexation for the majority of pre-modern, land-oriented states.

[5] It is rather concerning that two out of four apply to modern day Sino-Indian relations, but that is a topic for another article.

[6] Manpower is still quite important in insurgencies or counterinsurgencies, but this would only become a concern after a state-state war has been fought.

[7] Excepting the special cases mentioned above.

[8] Catherine the Great’s Russia was “protecting the Christians” in the Ottoman Empire long before Crimea was even but a twinkle in Vladimir Putin’s eye.

[9] Naval basing in Crimea is also a major reason. The Russian Black Sea Fleet is based in Sevastopol, within Crimea. Russia was loath to give up its base infrastructure in Crimea to a newly antagonistic Ukrainian state.

[10] Talmadge, Caitlin. The Dictator’s Army: Battlefield Effectiveness in Authoritarian Regimes. Ithaca; London: Cornell University Press, 2015.

[11] Bueno de Mesquita, Bruce, and Alastair Smith. 2011. The dictator’s handbook: why bad behavior is almost always good politics. New York: PublicAffairs.