IISS Military Balance: Intelligence Source Guide

The Military Balance is the annual open reference on the world's armed forces – personnel, equipment inventories, defence budgets and deployments, country by country, published by the IISS since 1959. It is the field's default order-of-battle baseline, and it is not free.

At a glance

Source	IISS Military Balance
Category	Conflict, Crime & Human Security › Military, Weapons & CBRN
Homepage	https://www.iiss.org/publications/the-military-balance/
Format	`HTML`
Access	Licensed — paid or institutional access The platform catalogue records this as open. That is wrong, and the correction is explained under Access, licensing and what you may do with it below.
Disciplines	Measurement & Signature Intel, Open Source Intelligence
Mission domains	Military & Defense

Authoritative order-of-battle reference. — as catalogued in the platform’s own source registry.

The Military Balance is a compiled assessment rather than a dataset. Each year the International Institute for Strategic Studies produces a country-by-country account of national military capability: active, reserve and paramilitary personnel strengths; equipment holdings broken out by service and by type, from main battle tanks and combat aircraft to submarines and transport helicopters; defence budgets and actual expenditure; force organisation; and deployments abroad, including contributions to multinational operations. Around the country entries sit regional chapters that describe trends in force development, procurement and defence economics, comparative tables ranking states on expenditure and holdings, and a set of explanatory notes that define every term used. Those explanatory notes are the most under-read section of the book and the one that determines whether an analyst uses it correctly. The IISS also maintains Military Balance+, an online database version offered on subscription that exposes the same underlying assessments in a queryable form. What is not on offer is an open API or a free bulk download; this is a commercial publication from a think tank that funds a substantial part of its research through it.

The analytical job the Military Balance does is standardisation. Every state publishes something about its armed forces, in its own categories, with its own definitions of what counts as active personnel or as an operational aircraft, in its own currency and fiscal year, and with its own incentives to inflate or conceal. Almost nothing is comparable as published. The IISS applies a consistent taxonomy across every country in the world, resolves the definitional differences, converts the money, and states what it has done in the explanatory notes. That is why it became the field's shared baseline: two analysts arguing about a regional balance can at least agree on what the numbers mean. For MASINT and OSINT work on force structure it is the reference against which imagery-derived counts and open-source order-of-battle reconstructions are checked. It is not a primary source and does not pretend to be. It is a carefully constructed secondary compilation whose value lies in the consistency of its judgements, and treating it as a measurement rather than as an assessment is the standing error in how it is used.

Who publishes it, and why that matters

The IISS is a British research institute founded in 1958, headquartered in London with offices in Washington, Berlin, Singapore and Manama, and it is a charity funded through publications, membership, conferences – it convenes the Shangri-La Dialogue and the Manama Dialogue – and research contracts including government and foundation funding. That funding mix has been the subject of public scrutiny, particularly around Gulf state support, and an analyst should hold the ordinary caution that applies to any policy institute with state funders: the factual compilation is far less exposed to that pressure than the analytical framing around it, and the two are printed in the same volume. The Military Balance itself is a commercial product and a significant revenue line, which explains the absence of open access and the subscription model for the online database. That commercial dependence is also a stability guarantee of sorts – the publication has appeared annually for over six decades and there is no sign of that changing. What it does mean is that your access is a procurement decision, renewable, and that an institution which lets its subscription lapse loses the current data and often the historical series with it.

Provenance is the first question to ask of any dataset and the one most often skipped. Who collects it, what their incentive is, whether they publish a methodology, and whether they correct the record when they get something wrong all bear directly on how much weight a finding drawn from it can carry.

What a record actually contains

The fields you will be working with, what each one means, and whether it is something you can pivot on. Read the meanings carefully — more analysis is wrecked by misreading a field than by failing to find one, and a field that looks like an observation is often an inference.

Field	Type	What it means	Pivot value
`country`	string	The state whose forces are described. Entries exist for recognised states and, in some editions, for entities with de facto armed forces; the editorial decision about who gets an entry is itself a political statement the notes explain.	Country dashboards; joins to any ISO-keyed geopolitical dataset.
`active_personnel`	int	Full-time serving military personnel. The figure is an assessment of actual strength where the IISS can establish it, and authorised establishment where it cannot – and the difference between those two can be large in states with recruitment problems.	Force ratio comparison; per-capita and per-GDP militarisation measures.
`reserve_personnel`	int	Reserve strength, which is the least comparable figure in the entire publication because reserve obligation means completely different things in a conscript system, a professional force with a trained reserve, and a state that counts everyone who ever served.	Mobilisation potential, with heavy caveats; never compare across systems without reading the notes.
`paramilitary`	int	Forces under military-style command that are not part of the regular armed forces – gendarmeries, border troops, internal security formations. Analytically crucial for regime security assessment and routinely omitted from headline comparisons.	Internal security posture; the coup-proofing structure of a state's security architecture.
`equipment_holding`	array	Inventory counts by equipment type and often by specific model. These are holdings, not operational availability – stored, non-serviceable and reserve equipment may be included, sometimes annotated as such.	Model identification for imagery correlation; procurement history and supplier relationships.
`equipment_category`	enum	The classification a holding falls under – main battle tanks, armoured personnel carriers, artillery, combat aircraft, principal surface combatants, submarines and so on. Categories are defined in the explanatory notes and the definitions are not intuitive.	Cross-country comparison within a category, which is the only comparison the taxonomy actually supports.
`defence_budget`	int	The announced or appropriated defence budget for the year, in national currency and converted. This is what was allocated, which in many states diverges substantially from what was spent.	Fiscal trend analysis; comparison against announced procurement programmes.
`defence_expenditure`	int	Assessed actual spending, which the IISS derives where budget figures are known to be incomplete or misleading. The methodology for this derivation is stated in the notes and differs from SIPRI's, which is why the two disagree.	Economic burden measures; comparison with SIPRI's series to expose methodological divergence.
`deployments`	array	Forces stationed or operating outside national territory, including multinational operation contributions, basing arrangements and advisory missions. One of the most useful fields and one of the hardest to keep current.	Theatre presence; coalition structure; host-nation relationships.
`force_organisation`	string	Order of battle description – the formations a service is organised into, their type and sometimes their location. Depth varies enormously by country according to what is establishable from open sources.	Unit-level analysis; correlation with imagery of garrisons and bases.
`edition_year`	int	Which annual edition a figure comes from. Essential metadata, because figures are revised between editions and a number without its edition is unciteable.	Longitudinal series construction; identification of revisions.
`data_confidence_annotation`	string	Notation used to flag estimated, uncertain, in-store or reported-but-unconfirmed figures. These annotations are the IISS telling you how much to trust a specific number and they are stripped out of almost every secondary citation.	Confidence weighting in any derived analysis.

Coverage — and what is not in it

Global by state, published annually, and the coverage is genuinely worldwide – the Military Balance covers small states and states of little interest to any particular analyst with the same taxonomy as major powers, which is precisely what makes it useful as a baseline. Depth is very uneven and varies with the transparency of the state in question. For NATO members and other states with published defence planning documents, the entries are detailed and well grounded. For closed states, the entries are careful assessments built from what can be established, and the explanatory annotations become correspondingly more important. Historical coverage runs to 1959, which makes the back series one of the longest continuous records of global force structure in existence, though comparing across decades is complicated by taxonomy revisions and by changes in how the IISS treats stored equipment and paramilitary forces. The publication cycle is the fundamental constraint: data is compiled to a cut-off some months before publication, so an edition describes a state of affairs that is already several months to a year old when it appears, and up to two years old by the time the next edition arrives. For a slow-moving question – fleet composition, structural reform, long-run expenditure trend – that is irrelevant. For an active conflict it is disqualifying.

Known blind spots

Absence of evidence here is not evidence of absence. These are the conditions under which IISS Military Balance will not show you something that is nevertheless real:

The publication cycle means an active conflict is described as it was at a cut-off months earlier, so attrition, emergency procurement and force reorganisation during a war are invisible until the following edition and sometimes the one after that.
Holdings are not readiness – an inventory count includes equipment that is stored, cannibalised, unserviceable or lacking crews, and the publication cannot tell you what fraction of a fleet can actually fly or move.
Sustainment is absent. Munitions stockpiles, spare parts, maintenance capacity, fuel and the industrial base that replaces losses determine how long a force can fight, and none of it appears in an order-of-battle table.
Closed states are assessed rather than reported, and the assessments inherit the limitations of the open sources behind them – for the most opaque militaries the entry is an informed reconstruction whose error bars are not published.
Non-state armed groups, militias and irregular formations are largely outside the scope even where they are the militarily decisive actors in a theatre, which makes the source structurally weak for exactly the conflicts most common today.
Qualitative capability is invisible in a count: two states with the same number of combat aircraft may differ by a generation in avionics, weapons, training hours and enabling capabilities, and the table gives no purchase on that.
Cyber, space and electronic warfare capabilities do not fit the equipment-count taxonomy and are covered narratively at best, so a domain that increasingly decides outcomes is under-represented by construction.
Defence expenditure figures depend on what a state chooses to put inside the defence budget line, and off-budget procurement, security service funding and state-owned enterprise subsidies escape the measure in the states where they matter most.
The data is a paid product, so any analysis built on it cannot be fully reproduced by a reader without a subscription, which is a real constraint on published research and on evidentiary use.

Write the blind spot into the product. A statement that something “was not observed in IISS Military Balance” is defensible; a statement that it “did not happen” is not, and the difference is what survives cross-examination.

Access, licensing and what you may do with it

Access model: Licensed — paid, or institutional access only

This is a commercial publication and should be treated as one despite any catalogue entry suggesting otherwise. The annual volume is sold as a book and as an electronic edition through the publisher, Military Balance+ is a subscription database, and institutional access is a procurement decision with a recurring cost. What the IISS publishes freely is a different and smaller body of material: launch commentary and press briefings around each edition, analytical articles and blog posts, and the descriptive pages on iiss.org. Those free items frequently contain the headline findings the press reports, so an analyst without a subscription can track the top-level narrative but not the underlying tables. University and government libraries commonly hold the series, and for historical work an institutional library is usually the more sensible route than a subscription. If your requirement is a machine-readable time series across many countries and years, be realistic: extracting it from the books is substantial manual work, the licence position on doing so is not obviously permissive, and SIPRI's openly published expenditure and transfer databases may serve the actual analytical need at zero cost.

Licence

All rights reserved commercial publishing, with terms set by the IISS and its publisher. Reading, citing and quoting limited extracts with attribution is normal scholarly use. Systematic extraction of the tables into a database, redistribution of substantial portions, or building a derived product that substitutes for the publication are not covered by any general permission and require a licence. Subscription access to the online database carries its own terms including restrictions on bulk export and on sharing credentials, and institutional subscriptions typically limit use to members of the subscribing institution. This matters more than it might seem because the temptation to build an internal order-of-battle database by transcribing the tables is strong and the licence position on doing so is at best unclear. Confirm terms directly with the IISS if you intend anything beyond citation – they license data for specific uses and a conversation is cheaper than a dispute.

Rate limits and fair use

Not applicable to the printed publication. For the subscription database, throughput is governed by the terms of your licence rather than by a technical limit, and bulk export is typically what those terms restrict. The practical throughput problem with this source is human: reading and transcribing country entries is slow, error-prone, and the single largest cost of using the Military Balance at scale. Budget analyst time accordingly, build a transcription check into any workflow that produces numbers from the tables, and resist the urge to automate extraction from a paid publication without first establishing that you are permitted to.

Licensing changes, and it changes without warning. A dataset that was free for research this year may not be free for commercial or evidential use next year. Confirm the current terms before you build a dependency on it, and record the terms you relied on alongside the data — the licence in force at the time of collection is part of the provenance.

Collecting it

How IISS Military Balance is actually pulled, in the order you would set it up. Prefer the bulk or export interface over per-item lookups wherever one exists: it is kinder to the publisher, faster for you, and gives a reproducible snapshot rather than a series of point-in-time answers you cannot reconstruct later.

Method	Format	Cadence	Notes
Annual edition	bulk	Annual, on publication	The core artefact. Acquire each edition and retain the back series – the historical volumes are the source of the longitudinal record and are not otherwise reconstructible.
Military Balance+ database	HTML	Per subscription; updated between editions	The queryable version for subscribers. More convenient than the book and subject to export restrictions; check what your licence permits before designing a workflow around it.
IISS free analysis and launch material	HTML	Weekly polling; concentrated around the annual launch	Openly published commentary carrying the headline findings. The accessible channel for anyone without a subscription and worth collecting even if you have one, because it states what the institute considers significant.
Manual transcription to a structured table	CSV	Per edition, per country of interest	The realistic route to a machine-readable series. Transcribe only the countries and categories you need, record the edition year on every row, and preserve the confidence annotations rather than discarding them.
Library access to the back series	bulk	As required	For historical work, institutional library holdings are usually more practical than subscription. Note that taxonomy changes between decades make naive longitudinal comparison unsound.

Ingesting it into the platform

Every step below is idempotent and cursor-based: interrupt one and it resumes from where it stopped rather than duplicating rows or losing progress. Collection is recorded per source, so a feed that quietly stops publishing shows up as a stale timestamp instead of silently thinning your coverage.

Register it honestly as a licensed reference — Add the Military Balance in sources.php as a commercial annual publication and record the access constraint. The catalogue entry in source-catalog.php should not imply an open feed, because an analyst planning collection against it needs to know there is a procurement step.
Stamp every figure with its edition — During import.php processing, carry edition_year on every record. A holding count without its edition is unusable, and this is the most common data hygiene failure when order-of-battle figures move between systems.
Preserve the confidence annotations — Carry the estimated, in-store and unconfirmed markings through into the record rather than normalising them away into a clean integer. The annotation is the IISS telling you how much of an assessment the number is, and stripping it manufactures false precision.
Resolve equipment models to entities — Map named equipment types to platform entities so that a specific aircraft or vehicle model becomes a node linking the states that operate it. This turns an inventory table into a supplier-and-operator graph and is where the compiled data starts doing work the book cannot.
Build the country capability view — Push personnel, holdings and expenditure into country.php and country-dashboard.php as dated attributes, so a state's force structure sits alongside its sanctions position, conflict exposure and economic indicators rather than in a separate reference silo.
Construct the longitudinal series deliberately — Assemble multi-year series in analytics.php only where the taxonomy is stable across the editions involved, and flag category definition changes explicitly. A series that silently spans a taxonomy revision will show a step change that is an artefact.
Cross-check against open expenditure data — Use correlate.php to compare IISS expenditure figures against SIPRI's independently produced series for the same country and year. Divergence is expected because the methodologies differ, and quantifying the divergence is more informative than picking one.
Constrain redistribution at the export boundary — Configure export.php so that licensed reference data is flagged and not shipped in bulk to external partners. Attribution and licence status should travel with the record, because the practical risk with this source is not accuracy but inadvertent redistribution of paid content.

Registered sources and their last-collected state are listed in sources.php, and the scheduled chain that keeps them current is in automation.php.

How it is wrong, and how to tell

Every dataset is wrong in characteristic ways. Knowing which ways is the difference between using a source and being used by one, and it is the part of source evaluation most often skipped because it is the part that takes work.

High, on a specific and limited definition of quality. The Military Balance is carefully compiled by specialists with access to defence attaches, national publications, industry reporting and imagery-derived analysis, and it is internally consistent in a way no other global compilation matches. Its taxonomy is stable, its definitions are published, and its uncertainty is annotated – which puts it ahead of almost every competitor on the criteria that actually matter for a reference work. Where it is weaker is exactly where every such compilation is weaker: it is secondary, it lags, it cannot see readiness or sustainment, and for opaque states it is an informed reconstruction whose sources are not shown. The most honest characterisation is that the IISS is very good at telling you what a state's force structure looks like on paper and cannot tell you what that force could do next month. Analysts who understand that treat it as a baseline to be adjusted; analysts who do not treat it as ground truth and are periodically embarrassed. The publication's own explanatory notes make the limits clear, which is the strongest available evidence of its quality and is read by almost nobody.

Characteristic false positives

Inventory counts are read as operational strength, producing capability assessments that assume every listed platform is available – an assumption that has failed publicly and expensively in recent conflicts.
Reserve figures are compared across states with incompatible reserve systems, generating mobilisation estimates that are meaningless because the underlying obligation differs by orders of magnitude in practice.
Defence expenditure comparisons ignore purchasing power, so states with low domestic costs for personnel and locally produced equipment appear far weaker relative to their spending than they are.
The publication lag is forgotten during fast-moving events, and figures compiled before a war are quoted months into it as though they described the current force.
Confidence annotations are dropped in transcription, converting an explicitly estimated figure into an apparently precise one as it moves from the book into a spreadsheet into a briefing slide.
Category definitions are assumed rather than read – what counts as a main battle tank, a principal surface combatant or a combat aircraft is defined in the notes and does not match intuition, so cross-source comparisons compare different things.
Longitudinal series are built across taxonomy revisions, producing apparent step changes in a country's holdings that reflect an editorial decision rather than any procurement or disposal.
Paramilitary and internal security forces are excluded from comparisons, which systematically understates the coercive capacity of states that deliberately place capability outside the regular armed forces.
Equipment listed as in store is counted as available, which in some states represents a large fraction of the nominal inventory and material in unknown condition.

None of these make the source unusable. They make it a source that requires corroboration before an assertion built on it goes into a product, which is true of every source and admitted by few.

Ageing

The publication ages on a predictable annual clock and its data ages faster than that, because compilation precedes publication by months. In peacetime this barely matters: force structures change slowly, procurement takes years, and a two-year-old holding count is usually within a few percent of the truth. In wartime the same figure can be worthless within weeks, and the gap between the printed inventory and the actual one becomes the central analytical problem. Defence expenditure figures age differently again – they are anchored to a fiscal year and remain historically valid indefinitely, but they get revised in later editions as better information emerges, so the figure for a given year is not stable across editions. A stale record here has a recognisable shape: a briefing that cites tank or aircraft numbers for a state at war, sourced to an edition published before the war began, with no adjustment for documented losses. The correction is not to abandon the source but to use it as a starting inventory and to subtract independently documented attrition, stating both.

What this source feeds

A source is only worth what it lets you conclude. These are the disciplines that collect through it, the mission domains it serves and the data points it yields — every one is a tag, so you can follow any thread from here into the rest of the library.

Collected by these intelligence disciplines

Serves these mission domains

Military & Defense

Yields these data points

Facility / Site

How each sector uses IISS Military Balance

The same dataset is worked very differently depending on who you are, what authority you hold, and what you are ultimately producing. A military analyst is supporting a commander’s decision; a journalist is meeting a publication standard; an NGO caseworker is protecting a person. The records are shared — the constraints, thresholds and outputs are not.

🎖 Military and defence

For planners this is the standing reference for force comparison, coalition contribution analysis and regional balance assessment, and its taxonomy is close enough to NATO conventions that translation into planning products is straightforward. Its greatest practical value is as a common baseline in a coalition setting where partners cannot share classified assessments but can all cite the same open compilation. The professional discipline is to treat every entry as a paper strength and to apply your own readiness, sustainment and training judgements on top – the publication explicitly does not model those and a plan that assumes it does is unsound. For a theatre with an active conflict, use it for the pre-conflict baseline and drive current strength from operational reporting and documented attrition, because the edition in your hand describes a force that no longer exists in that configuration.

🕵 National intelligence

The Military Balance is the open baseline against which classified order-of-battle holdings are sanity-checked, and the check runs both directions – a divergence between your holdings and the published figure is a question worth answering, sometimes about your data. For MASINT and imagery work it provides the expected inventory that makes a count anomalous, which is how a new deployment or an unreported acquisition becomes visible. Its expenditure and procurement narrative supports indications and warning at the strategic level, where budget shifts and structural reform precede capability changes by years. Two cautions matter for an intelligence user. The compilation draws on open sources and on defence attache reporting, so it is not independent of official narratives in the states that cooperate with it. And because it is a paid, all-rights-reserved product, its use in disseminable assessments raises a redistribution question that should be settled with your legal staff rather than assumed.

👮 Law enforcement

Direct law enforcement application is narrow but real. Investigations into arms trafficking, sanctions evasion and end-use violation frequently need to establish what a state's armed forces actually operate, because a transfer of equipment a country does not use is a red flag and a transfer consistent with its inventory is not. The equipment tables answer that quickly and citably. The publication also identifies paramilitary and internal security formations, which matters for cases touching state-linked violence, procurement by security services and human rights sanctions. Recognise that this is a reference work rather than evidence: it is a secondary compilation of assessments, it will not support a finding about a specific transfer, and its licence terms constrain reproduction in case material. Use it to frame enquiries and rely on primary records – customs, export licences, registry documents – for anything evidential.

🔍 Private investigation and corporate security

For corporate intelligence, defence sector due diligence and country risk work, the Military Balance answers questions clients ask constantly: what does this country's military actually consist of, how much is it spending, where is it deployed, and who supplies it. For firms advising on defence exports or on operations in militarised environments it is the standard reference and citing it carries credibility. The commercial constraint is straightforward and often ignored – a consultancy cannot lift the tables into client deliverables without regard to licensing, and a firm that builds a product on transcribed content is exposed. Cite, do not reproduce, and where a client needs an underlying dataset they can hold, point them at the openly licensed alternatives and be honest about what those do and do not cover.

📰 Journalism and OSINT media

For defence journalism this is the citable authority, and the annual launch is a fixed point in the reporting calendar because the IISS presents its own headline findings. The reporting errors are consistent: inventory counts described as operational strength, expenditure compared at market exchange rates without any acknowledgement of what that distorts, and figures from a pre-conflict edition used to describe a force in the middle of a war. Reading the explanatory notes once would prevent most of them. The publication's paywall is also a live issue for reporting practice – readers cannot check your figures, so state the edition, state that holdings are not readiness, and where the story is about an active conflict, say plainly that the numbers predate it. IISS researchers speak to press and are a better source for interpretation than the tables alone.

🌍 NGO, humanitarian and human rights

Arms control, human rights and peacebuilding organisations use the Military Balance for militarisation trends, for identifying the security forces implicated in a situation, and for grounding advocacy on defence spending in figures that governments recognise. The paramilitary and internal security entries are disproportionately useful for human rights work because they identify the formations that conduct internal repression and that sit outside conventional arms control frameworks. The access barrier is a real equity problem in this sector – the organisations closest to the affected populations are the least likely to afford the subscription – and SIPRI's open expenditure and transfer data covers a meaningful part of the need without cost. Where you cite the Military Balance, cite the edition and be careful not to reproduce tables in published material, because enforcement of publishing rights is not hypothetical.

🎓 University and research

The back series to 1959 is a genuinely valuable long-run record of global force structure and is used across security studies, international relations and defence economics. Its methodological transparency – published definitions, annotated uncertainty, explained expenditure derivation – makes it more defensible than most compiled sources, and the explanatory notes should be cited alongside any figure taken from it. Three problems recur in the literature and are avoidable. Researchers build panels across taxonomy revisions without checking whether the category definition changed. They merge IISS and SIPRI expenditure series, which are constructed differently and are not interchangeable. And they publish datasets derived from a paid all-rights-reserved publication without addressing whether they were permitted to, which is a growing issue as journals require data deposit. Settle the licence question at the design stage rather than at submission.

Playbook: working IISS Military Balance end to end

A repeatable sequence from first pull to finished product. Each phase states what you are trying to establish, not merely what to click — the objective is a defensible chain of reasoning, not a completed checklist.

Phase 1 — Read the explanatory notes before any figure

The definitions of every category, the treatment of stored equipment, the annotation scheme and the expenditure methodology are all stated there. An hour with the notes changes how you read every subsequent table. This is not a formality – the majority of misuses of this source are definitional and would be prevented by that hour.

Phase 2 — Fix the edition and the cut-off date

Establish which edition you are using and roughly when its data was compiled. Every number you produce inherits that vintage. Record it on the face of your working file, because the edition is the first thing a competent reviewer will ask for and the first thing an analyst forgets to carry forward.

Phase 3 — Separate paper strength from usable strength at the outset

State explicitly, in your own working assumptions, that holdings are inventory. Then decide how you will estimate availability – documented readiness rates, maintenance reporting, imagery of active versus stored equipment, operational tempo. That estimate is your contribution; the inventory is the input. Conflating them is how force comparisons go badly wrong.

Phase 4 — Build the baseline before the conflict, not during it

For any theatre with active fighting, take the last pre-conflict edition as the starting inventory and treat every subsequent figure as baseline minus documented attrition plus documented resupply. Open-source loss documentation and procurement reporting drive the adjustment. The next edition will eventually catch up, and comparing your running estimate against it is a useful check on your method.

Phase 5 — Handle expenditure with two methodologies, not one

Take the IISS figure and the SIPRI figure for the same country and year and understand why they differ – budget scope, exchange rate treatment, what counts as defence spending. Present the range rather than choosing. Where the question is real military capacity rather than fiscal burden, consider purchasing power adjustment and say that you have, because market exchange rate comparison systematically misrepresents states with low domestic cost bases.

Phase 6 — Pull the paramilitary and internal security picture forward

For any assessment touching regime stability, internal repression or coup-proofing, the paramilitary entries matter more than the regular force tables. States that fear their own militaries build parallel structures, and the shape of that parallel structure is one of the most informative things in the country entry. It is also the part most consistently omitted from comparative analysis.

Phase 7 — Use the deployments field as a relationship map

Foreign deployments, basing and advisory presence describe alliance structure and dependency more concretely than any diplomatic reading. Extract them into a relationship graph across countries. Patterns of who trains and bases whom persist across changes of government and are a better predictor of alignment than declaratory policy.

Phase 8 — Correlate equipment models to suppliers

Resolve named equipment types to their producers and export histories. This converts an inventory into a procurement relationship map, exposes dependency on a single supplier for sustainment, and identifies which states would be affected by sanctions on a given defence industrial base. This is the highest-value derivation available from the tables and it requires nothing but a lookup table.

Phase 9 — Cross-check counts against independent observation

Where an inventory figure is load-bearing, test it against imagery-derived counts, open-source loss documentation and reported deliveries. Agreement raises your confidence; divergence identifies either an error or a change the compilation has not yet captured, and both are worth knowing before you publish.

Phase 10 — Interrogate the qualitative gap

Two states with identical counts can be a generation apart. Before making any balance judgement, characterise generation, weapons fit, training levels, enabling capabilities and command arrangements from other sources. The count is the skeleton; the assessment lives in what you add to it, and a comparison presented as a table of numbers is not an assessment.

Phase 11 — Construct any time series with the taxonomy in view

When building a multi-year panel, verify that the category definition and the treatment of stored equipment were stable across the editions involved. Where they were not, break the series and say so. An unexplained step change in a published chart is nearly always an editorial revision and it will be found by someone.

Phase 12 — Settle the licence position before you build anything

If your work involves transcribing tables into a database, redistributing figures to partners, or depositing a derived dataset, resolve the permissions before you invest the effort. This is a commercial all-rights-reserved publication and the IISS does license data for defined uses. A conversation at the start is a great deal cheaper than a rebuild or a retraction.

The platform ships this as a step-checked workflow in playbooks.php, so progress is recorded against a case rather than held in someone’s head.

What to pair it with

No single source carries a finding. These are the datasets that corroborate, extend or contradict this one — and a source that contradicts is worth more than one that agrees, because it is the only thing that will tell you when you are wrong.

Source	Relationship	What it adds
SIPRI Military Expenditure Database	contradicts	An independently constructed expenditure series, openly published and free. It uses different definitions from the IISS and disagrees with it – which is why holding both is more informative than trusting either.
SIPRI Arms Transfers Database	extends	Transfers of major conventional weapons with a documented methodology and openly published data. Explains how the inventories in the Military Balance came to exist and who supplied them.
UN Register of Conventional Arms	corroborates	State-declared transfers in seven major categories. Voluntary and incomplete, but a state's own declaration is the strongest possible confirmation of an acquisition when it exists.
Janes	extends	Commercial defence intelligence with deeper platform-level and unit-level detail than the Military Balance provides. Different scope, similar commercial constraints, and the two are routinely used together.
NATO	corroborates	Publishes defence expenditure figures for member states on a common definition, which is the cleanest comparable expenditure data available for that group and a useful check on both IISS and SIPRI.
World Bank Open Data	extends	Military expenditure as a share of GDP and the wider economic series needed to normalise defence spending against national capacity.
Oryx	extends	Open-source visual documentation of equipment losses. The practical instrument for adjusting a pre-conflict inventory to a current one, with the caveat that it counts only what has been photographed.
CSIS	corroborates	Open defence analysis including force structure and budget work, useful as an independent read on the same questions and generally published without a paywall.
IISS	prerequisite	The institute itself – its freely published analysis, its funding disclosures and its explanation of how the Military Balance is compiled.

Legal, ethical and operational constraints

The dominant legal issue here is copyright, not collection. This is a commercial all-rights-reserved publication and systematic reproduction of its tables, redistribution to third parties, or construction of a derived product that substitutes for it are not permitted without a licence. That constraint bites hardest exactly where the source is most useful – building an internal order-of-battle database – and it is routinely ignored. If you intend to do more than cite, obtain permission; the IISS licenses data and the process is ordinary commercial negotiation. Subscription terms for the online database typically limit access to the subscribing institution and restrict bulk export, and credential sharing breaches them. There is a secondary consideration for government users: using open compiled material in a classified product raises a redistribution question when that product is later shared with partners, and it should be resolved with legal staff rather than assumed. Ethically, the source raises little difficulty – it is aggregate national capability data, not personal data, and its publication is a long-standing and accepted transparency contribution. The one place to be thoughtful is in publishing derived analysis of a specific state's vulnerabilities, where the reasonable line is analysis of force structure rather than production of anything resembling targeting material.

Operational security

Buying a book or holding a subscription reveals your institution's interest in defence matters, which for most users is unremarkable. The subscription database logs what you query, and a query pattern narrowly focused on one country over a sustained period is visible to the operator – a consideration for a small number of users and irrelevant to most. The more substantial exposure is internal and downstream: an order-of-battle database assembled inside your organisation is a concentrated statement of your analytical interests, and its access controls should reflect that even though every individual number in it came from a published book. Where your work compares a specific state's forces against another's, the existence of that comparison is often more sensitive than its content. Apply ordinary compartmentalisation, and do not assume that because a source is open the product built from it is.

Two rules that hold regardless of jurisdiction. Collection that is lawful is not automatically proportionate, and a dataset assembled for one purpose does not carry consent for another. Where the records concern identifiable people, the question is not only whether you may hold the data but whether holding it serves the purpose you are accountable for.

Is it earning its place?

Sources accumulate. Feeds get added during an incident and are never reviewed again, and a decade later the pipeline is carrying dead weight that nobody dares remove. These are the measures that show whether IISS Military Balance is contributing anything, and they are worth baselining now so the answer is available later.

Proportion of order-of-battle figures in your holdings carrying an explicit edition year, which should be all of them and frequently is not.
Retention of confidence annotations through transcription, audited by sampling records against the printed source.
Divergence between IISS and SIPRI expenditure figures for your priority countries, tracked as a measure of how much methodological uncertainty your analysis is carrying.
For active conflicts, the gap between your adjusted running inventory estimate and the next published edition, which measures whether your attrition tracking is working.
Number of comparative claims made across a taxonomy revision without acknowledgement, which should be zero and requires a deliberate check to keep there.
Coverage of equipment-to-supplier mappings across your countries of interest, since that derivation is where most of the additional value is created.
Licence compliance status of any derived internal dataset, reviewed at least annually rather than settled once at creation.

Beware of volume. Indicator counts rise easily and say almost nothing. Unique contribution — findings this source produced that no other source in your stack would have — is the measure that matters, and it is usually far lower than anyone expects.

Tradecraft notes

The distinctions that separate a competent analyst from a fast one:

The explanatory notes are the most valuable pages in the book and the least read; category definitions there do not match intuition and reading them prevents most misuse of the tables.
Holdings are not readiness, and the recent record of large armies performing far below their inventory is the most expensive demonstration of that distinction available.
IISS and SIPRI expenditure figures differ because the methodologies differ, and an analyst who can explain why they differ is more useful than one who has picked the larger number.
Market exchange rate comparison of defence spending systematically flatters states with high domestic costs and understates those with low ones, which is a first-order effect and not a refinement.
Paramilitary and internal security entries reveal a regime's assessment of its own military, and that structural fact predicts internal behaviour better than any regular force table.
The publication lag is a fixed property, not a defect, and the correct response is to treat each edition as a dated baseline and to run your own adjustment forward from it.
Reserve figures are the least comparable numbers in the volume because reserve obligation means fundamentally different things in different systems, and mobilisation estimates built naively on them are worthless.
Equipment listed as in store may be a large fraction of a nominal inventory and its condition is generally unknown, so an analyst who counts stored equipment as available has made a strong hidden assumption.
Because access is paid, published work built on this source is not fully reproducible by readers, which is a real limitation to state rather than to ignore.

Questions analysts actually ask

Is the Military Balance free?

No. It is a commercial publication and Military Balance+ is a subscription database. Some analysis and launch commentary is published freely on iiss.org, but the tables are behind a paywall. If your requirement is open data, SIPRI's expenditure and arms transfer databases are free and cover part of the same ground with published methodologies.

Can I use it to assess who would win a conflict?

Not on its own, and not close. It gives you inventories and personnel, and says nothing about readiness, sustainment, training, munitions stocks, command quality or the enabling capabilities that decide outcomes. Use it as one input to a force comparison and be explicit that the count is the least important part of the assessment.

Why do IISS and SIPRI give different defence spending numbers?

Because they define and construct the series differently – what counts as defence spending, whether the figure is a budget or an assessment of actual outlay, and how currency conversion is handled. Neither is wrong. Present both, explain the divergence, and treat the gap as the uncertainty you are actually carrying.

How current is the data during a war?

It is not. Data is compiled months before publication, so an edition describes the force as it was before the events you are interested in. Use the last pre-conflict edition as a baseline and drive current strength from operational reporting and documented losses.

Can I build a database from the tables?

Not without addressing the licence. This is an all-rights-reserved commercial publication and systematic extraction is not covered by any general permission. The IISS does license data for defined uses; ask before you build, because retrofitting permission onto an existing product is much harder.

Does it cover non-state armed groups?

Largely not, and this is a structural limitation rather than an oversight – the taxonomy is built for state armed forces. For conflicts where irregular formations are decisive, you will need conflict-specific sources, and the IISS Armed Conflict Survey addresses some of that ground separately.

How far back does the series go and can I use it longitudinally?

To 1959, which is a remarkable record. You can use it longitudinally with care: category definitions and the treatment of stored equipment and paramilitary forces have changed, so verify stability across your window and break the series where they did not hold.

What does an in-store annotation actually mean?

That the equipment is held but not in active service, with condition generally unknown. In some states this represents a large share of the nominal inventory. Whether to count it depends entirely on your question, and the decision should be stated rather than made silently.

Should I cite the book or the database?

Cite whichever you used, with the edition or the access date, because figures are revised between editions and between database updates. An order-of-battle number without a vintage is not a citable fact, and reviewers in this field will ask.

Standards, formats and interoperability

What this source speaks natively, and what it has to be translated into before a partner can consume it. Work that arrives in a recognised format is easier to defend, easier to hand over and easier to automate against:

The IISS explanatory notes and category definitions, which function as the de facto taxonomy for open order-of-battle work across the field.
UN Register of Conventional Arms categories, the seven-category scheme states use for voluntary transfer declarations and the nearest thing to an intergovernmental equipment taxonomy.
NATO defence expenditure definitions, which provide a common measurement basis across alliance members and a benchmark for evaluating other methodologies.
ISO 3166 country coding, the join key between force structure data and every other geopolitical dataset.
SIPRI trend-indicator values for arms transfers, which are not monetary and are frequently misread as though they were – know the difference before comparing them to budget figures.
NATO reporting names and standard equipment designations, the shared vocabulary that makes equipment identification comparable across sources.
STIX 2.1 identity and location objects for export into the platform's model, with the caveat that force structure data is a poor fit for a threat-intelligence schema and CSV usually serves better.

References

Primary documentation and authoritative references for this source. Publishers revise and retire material, so treat the retrieval date as part of the citation and re-check before relying on any of it in a formal product.

The Military Balance — International Institute for Strategic Studies. The publication's own page – scope, edition information and access arrangements. Start here to understand what you would be buying.
International Institute for Strategic Studies — IISS. The institute, its freely published analysis, and its funding disclosures. Worth reading the latter for the ordinary reason you read any policy institute's funders.
SIPRI Military Expenditure Database — SIPRI. The open alternative for defence spending, with a fully documented methodology. The first thing to check any IISS expenditure figure against.
SIPRI Arms Transfers Database — SIPRI. Open transfer data explaining how inventories came to exist. Read the trend-indicator value methodology before quoting any figure from it.
UN Register of Conventional Arms — UNODA. State-declared transfers. Sparse, voluntary, and authoritative where a declaration exists because it is the state's own admission.
NATO — North Atlantic Treaty Organization. Common-definition defence expenditure figures for members, the cleanest comparable spending data available for that group.
Janes — Janes. Commercial defence intelligence with deeper platform and unit detail. The usual companion to the Military Balance in institutions that can afford both.
World Bank Open Data — World Bank. Economic series for normalising defence spending against national capacity, including military expenditure as a share of GDP.
Oryx — Oryx. Visually confirmed equipment loss documentation – the practical tool for adjusting a printed inventory to a wartime reality.
Center for Strategic and International Studies — CSIS. Open defence analysis on force structure and budgets, useful as an independent and generally unpaywalled read on the same questions.

Link integrity: every reference above was verified with a live request when this page was generated. Where a publisher had moved or withdrawn a document, the link was repointed at a preserved copy in the Internet Archive and marked as archived. Anything with no reachable copy anywhere had its link removed rather than left to rot — the source is still credited, it simply cannot be linked.

Put it into practice

The Quantus Intel threat intelligence platform operationalises this source: it holds order-of-battle figures with their edition year and confidence annotations intact, resolves equipment models into a supplier-and-operator graph, keeps IISS and SIPRI expenditure series side by side rather than silently choosing one, and flags licensed reference data at the export boundary.. Browse the full source catalogue, or follow any tag above into the rest of the library.