The Arsenal Illusion: Seven Fallacies in Measuring Military Power

A country may top military power rankings, yet emerge from war unable to achieve the objective it fought for. It may be worn down by a less heavily armed adversary that lacks the capacity to defeat it outright. This gap between possessing military power and achieving its intended purpose is where the most common error in assessing armed forces lies: measuring them by what they possess rather than by what they can actually do with it.

That gap has rarely been more apparent than it is today. In 2025, global military expenditure reached approximately $2.89 trillion, marking the eleventh consecutive annual increase, while the global military burden rose to 2.5% of world GDP, its highest level since 2009. Yet this spending did not buy the ability to achieve decisive outcomes. In the Red Sea, the world's most powerful navy fired nearly 400 defensive munitions, worth more than $1 billion, over fifteen months against an armed group with neither a fleet nor an air force in the institutional sense. Egypt, meanwhile, paid the price for a war it did not fight: the Suez Canal lost more than 60% of its revenues in 2024, equivalent to nearly $7 billion. Then came the war with Iran in 2026, which, according to estimates by the Center for Strategic and International Studies (CSIS), depleted roughly two-thirds of the United States' stockpile of Patriot interceptor missiles in less than six months.

These facts do not suggest that money and weaponry are of little value. Rather, they point to something more troubling: the prevailing approach to measuring military power counts what is easy to quantify while overlooking what ultimately determines outcomes. The rankings widely cited by the media, invoked by governments in their deterrence rhetoric and used to shape expectations about wars begin and end with the arsenal, as though the path from the stockpile to the battlefield, and from the battlefield to the negotiating table, were a straight line with no twists or turns.

This analysis therefore begins from a more rigorous definition: military power is what a state can do with what it possesses, against a particular adversary, in a particular mission and for a particular duration. Between possession and action lie three conversion processes, none of which occurs automatically: converting resources into ready capability, capability into operational effect, and operational effect into political outcome. What we call a "fallacy" here is any line of reasoning that skips one of these processes, turning a partial indicator into a comprehensive judgement.

The idea itself is not new. As early as 2000, the RAND framework distinguished between national resources, national performance, and military capability, emphasising that institutions and innovation convert the first into the last. What is new is that the past three years have provided an open laboratory for observing this gap. From Ukraine to the Red Sea and the confrontation with Iran, the divide between power on paper and power actually brought to bear has become measurable in hard numbers, rather than through intuition alone.

The first fallacy arises when the number of platforms on record is treated as a direct measure of military capability. A hundred aircraft on the books do not mean a hundred aircraft that can take off tomorrow to perform a specific mission. Nor does that figure reveal how many sorties can be generated each day, how much suitable ammunition is available, or the distinction between an operational unit and one held in storage.

The US F-35 fighter, the most expensive weapons programme in history, offers one of the clearest examples. In October 2024, the US Government Accountability Office (GAO) reviewed eleven tactical aircraft fleets on which approximately $57.2 billion had been spent on operations and maintenance between 2018 and 2023, concluding that their readiness had “generally not met the Department of Defense’s goals for several years”. A June 2026 report then revealed that the proportion of F-35s capable of performing all their assigned missions had fallen from 38.1% in 2021 to 24.6% in 2025, while the Air Force variant achieved just 28.5%, against a stated target of 65%. Put differently, for every four aircraft on the books, only one was fully mission capable.

These figures come with two important caveats. They do not demonstrate that US air power is weak overall, as the sheer size of the fleet and the depth of its reserves remain genuine advantages. What they do demonstrate is that the right question is not, “How many aircraft does a state possess?” but rather, “How many can it bring to bear, where, for how many days, and with what support?” More importantly, a state that publishes its readiness figures with such transparency may appear weaker on paper than states that disclose nothing at all, without this necessarily meaning that those states are in better shape.

A military budget measures what a state spends, not what it gets in return. Expenditure may rise because of personnel costs, inflation or the expense of maintaining ageing equipment, while investment in research and development may take years to translate into operational capability. The Stockholm International Peace Research Institute (SIPRI) therefore explicitly cautions against treating military expenditure as a direct measure of capability. It also notes that dollar-based comparisons are affected by exchange rates and that annual expenditure represents a financial flow, which is fundamentally different from the stock of capabilities accumulated over time.

The 2025 figures illustrate the scale of this distortion. The United States spent $954 billion, China $336 billion and Russia $190 billion, meaning that the three countries together accounted for 51% of global military expenditure. Yet Russia, whose military spending is less than one-fifth that of the United States alone, produced between 2 million and 2.3 million artillery shells in 2024, while combined US and European production amounted to only around 1.7 million shells in 2025. The key, then, lies not in the size of the budget but in its composition: a war economy allocating 7.5% of GDP to the military, state-owned defence industries, lower labour and material costs, and a clear prioritisation of ammunition over expensive platforms.

The paradox is even more striking in the Arab region. Military expenditure in the Middle East reached $218 billion in 2025, with Saudi Arabia ranking eighth globally at $83.2 billion, while Iran’s reported military spending stood at just $7.4 billion, a figure that SIPRI suggests is likely to understate the true level. Yet the missile and drone network built by Iran and its proxies has imposed defensive costs on its adversaries many times greater than Iran’s declared military budget. This is not evidence of Iranian military superiority. Rather, it demonstrates that the relationship between money and military power is mediated by institutional, doctrinal and industrial choices that headline budget figures alone cannot capture.

Technology confers genuine advantages in detection, precision, protection and communications. Yet a platform’s performance under test conditions does not translate intact to a battlefield shaped by jamming, deception, attrition and an adversary that learns and adapts. The fallacy here is twofold: assuming that the more expensive system is necessarily the better one, and that newer technology renders what came before it obsolete.

The first dimension shows up in the cost equation. A Shahed-136 drone costs around $50,000, while a Patriot interceptor costs approximately $4 million, eighty times as much. In the Red Sea, between October 2023 and January 2025, the US Navy fired around 120 SM-2 missiles, costing up to $2.5 million each, and 80 SM-6 missiles, at roughly $4.3 million apiece, in addition to SM-3 missiles costing between $12.5 million and $28.7 million each. And in just six weeks in the spring of 2025, seven US MQ-9 drones, each worth approximately $30 million, were shot down.

Yet reading these figures as evidence of a “victory of the cheap over the expensive” would be an inverse fallacy. Using an expensive missile to intercept a low-cost drone may be entirely rational if that drone is heading towards a destroyer carrying three hundred sailors. The real question is not the cost of a single interceptor, but whether that level of protection can be sustained against the expected intensity of attacks, and whether the defensive system can match the appropriate countermeasure to each threat. This is precisely what has prompted the US Navy itself to return to gun-based defences and seek lower-cost interceptors.

The second dimension is evident in Ukraine. Although Ukraine’s National Security and Defense Council estimates that FPV drones now account for around 60% of Russian military losses, and Ukraine’s annual production capacity has exceeded eight million drones, Ukrainian officers from some of the country’s most effective drone units told researchers at the Royal United Services Institute (RUSI) that they still need artillery, and that drones are far more effective when employed alongside it rather than as a substitute for it. New technology, therefore, cannot be assessed in isolation, but only as part of an integrated combat system.

Ownership of a platform may pass to a state while critical elements required to operate and sustain it remain distributed among foreign suppliers, governments and companies: munitions, spare parts, technical data, software updates and intelligence support. The real question, then, is not who owns the weapon, but who controls the decision that keeps it operational.

Afghanistan provides the archetypal case. By March 2021, the United States had spent $88.3 billion building the Afghan security forces, whose nominal strength exceeded 300,000 personnel, compared with an estimated 55,000 to 85,000 Taliban fighters, according to UN estimates. Yet the number actually available for duty on any given day was no more than around 180,000, while Washington financed roughly 75% of an annual military budget of between $5 billion and $6 billion. When US-funded contractor maintenance support was withdrawn in 2021, aircraft availability declined rapidly in a force whose doctrine had been built around air support.

However, the Special Inspector General for Afghanistan Reconstruction (SIGAR) did not attribute the collapse to a single technical factor. It also identified weak leadership, corruption, declining morale and the impact of the decision to withdraw. The lesson, therefore, is not that external dependence is inherently a weakness. Alliances and burden-sharing can be legitimate sources of strength. What matters is assessing “critical dependence”: what ceases to function immediately if a particular form of support is withdrawn? What alternatives are available? And how long would it take to bring them into operation?

Comparing stockpiles provides only a static snapshot at a particular moment, whereas wartime military capability is dynamic. It erodes as munitions are expended, equipment is lost, and personnel become casualties, and is replenished only to the extent that production, repair and training can sustain it. An initially wide advantage may therefore narrow over time if the stronger side cannot replace its losses as quickly as it consumes its resources.

Recent figures reveal the pace of this attrition. In just fifteen months, the US Navy fired more air-defence missiles in the Red Sea than it had over the previous thirty years. During the war with Iran, the THAAD interceptor stockpile fell from 452 missiles before the conflict to between 234 and 278. In contrast, the Patriot interceptor inventory declined to roughly one-third of its pre-war level. All of this occurred in less than six months, while replenishing those stocks is measured in years.

By contrast, Russia increased its artillery-shell production from around 1.25 million rounds in 2022 to between 2 million and 2.3 million in 2024, with estimates suggesting that annual output could reach 4 million. A RUSI study of the third year of the war in Ukraine documented how the relative advantages of the two sides shifted from one phase to another as each adapted to the other’s tactics. In a protracted war, then, the decisive question is not “Who has more?” but “Who can replenish faster and learn faster?”

Nor should stated production capacity be equated with actual output or deliveries, since these remain constrained by the availability of components, labour and financing. A RUSI analysis of the drone industry shows that control over critical subsystems, from engines to electronics, determines how quickly systems can be upgraded and adapted. The existence of an assembly plant alone, therefore, says little about the true depth of a country’s industrial capacity.

Ranking states along a single scale assumes that widely divergent military missions can be reduced to one aggregate score. Yet defending nearby territory, securing a maritime corridor, conducting a distant military campaign, occupying territory and deterring a strategic strike each require a different configuration of capabilities. A state may therefore possess a decisive advantage in one mission while facing severe constraints in another.

For example, the Houthis do not feature in any serious ranking of military power. Yet, using relatively low-cost missiles and drones, they reshaped international shipping routes and contributed to a loss of more than 60% of Suez Canal revenues in a single year. This does not represent a shift in the overall balance of power. Rather, it illustrates a simpler principle: the ability to disrupt a particular mission is fundamentally different from the ability to dominate a theatre, and a weaker actor overall may still achieve sufficient superiority at the specific point that matters to it.

Every comparison should therefore begin by specifying the mission, the adversary, the duration, the geography and the assumptions regarding allied participation. Aggregate rankings remain useful for providing an initial picture of available resources, but become misleading when used to predict the outcome of a confrontation whose specific conditions they do not capture. The same applies to nuclear weapons: they are a fundamental component of strategic deterrence, but they have not prevented limited wars involving nuclear-armed states, or wars in which such states have themselves come under attack, from Ukraine to the India-Pakistan confrontation in May 2025. Nor can nuclear weapons substitute for the capabilities required to prosecute each limited conflict.

The ability to destroy a target is one thing; the ability to compel an adversary to accept a political demand is quite another. An adversary may be prepared to absorb enormous losses if it regards concession as a threat to its survival. At the same time, the destruction of its forces may fail to produce a stable alternative authority. Assessing coercive power therefore requires understanding what the targeted actor seeks to protect, what alternatives remain available to it, and whether the demand itself can be achieved through force.

In March 2025, the United States launched a major air campaign against the Houthis. Less than two months later, the campaign ended with an Oman-brokered agreement to halt reciprocal attacks involving US vessels, while the Houthis declared that the agreement did not extend to Israel. In other words, the world’s most powerful military emerged having secured the safety of its own shipping, rather than reopening the maritime corridor the campaign had been launched to protect.

Four years earlier, twenty years of US military presence in Afghanistan ended with the return to power of the very movement Washington had intervened to overthrow. The conditions for success, however, differ fundamentally between the two sides. For the defender, denying the attacker its objective may be sufficient; the attacker, by contrast, must achieve lasting change on the ground. A defender’s ability to hold out does not mean that it has become the stronger military actor. It may simply mean that its threshold for success is lower, while the cost of achieving the attacker’s objective has risen beyond what the attacker is willing to bear.

Care is also needed to avoid the opposite error: treating every political failure as evidence of military weakness. The problem may instead lie in the design of the objective itself, in political constraints on the use of force, or in the absence of any coherent vision for what follows military operations. Equally, the fact that a war does not occur may itself be an effect of deterrence, although demonstrating this requires evidence of the adversary’s calculations rather than assumptions about them.

The seven fallacies demonstrate that assessing military power requires tracing the relationship between resources, capabilities and outcomes. Material and financial resources provide the means to build military power; military institutions determine its readiness and how effectively it can be employed; while the conditions of a particular confrontation and the political objectives at stake shape the outcomes that can ultimately be achieved. An arsenal therefore remains a necessary starting point for assessment, but it must be understood within the broader system that determines its effectiveness.

A rigorous assessment should begin by defining the mission, the adversary, the timeframe and the geographical context. It should then examine readiness, force integration, the sustainability of supply and replenishment, the speed of adaptation, the reliability of support chains, and the extent to which available capabilities suit the objectives at hand. Taken together, these dimensions help explain differences that purely numerical comparisons cannot capture.

Such an assessment also requires a clear acknowledgement of the limits of available knowledge. Data gaps increase uncertainty, while differences in the degree of disclosure among states demand caution when making comparisons. Results from testing, exercises and repeated operational performance are generally more revealing than any single episode. This approach also gives defence planners a clearer basis for setting priorities and maximising return on investment.

A particular mission may require a larger force, while another may call for greater investment in training, maintenance, ammunition stocks or supply-chain resilience. The value of each investment therefore lies in how much it contributes to an integrated capability that can be employed and sustained under the conditions in which it is expected to operate.

On this basis, military power rankings are most valuable as a starting point for comparison, with a clear understanding of their limitations in explaining performance or predicting conflict outcomes. The question that should guide any assessment, therefore, is: “What do these capabilities enable a state to do, under what conditions, and for how long?” Answering this question shifts the debate from ranking militaries to understanding their actual capabilities, and gives armament figures their strategic meaning.