The ESG Industry's Military Blind Spot — How "Sustainable" Funds Quintupled Their Defense Exposure

If you put $10,000 into a fund marketed as sustainable in Europe today, there is a meaningful chance some of that money is buying shares of major defense contractors. The probability depends heavily on which sub-category of "sustainable" fund you chose — and that distinction is itself the piece of architecture worth understanding. Funds tracking strict EU Paris-Aligned Benchmarks generally exclude Lockheed Martin and similar names because of their nuclear-weapons exposure. Funds in the broader, much larger SFDR Article 8 ("light green") category often do not. The result is a sustainability-finance world in which the same word — sustainable — covers products with materially different defense exposure, and where the world's third-largest emitter is, across the universe as a whole, structurally undercounted. A fighter jet that burns ten thousand gallons of jet fuel an hour is, from the perspective of MSCI's industry materiality model, an operationally efficient manufacturing facility assessed for the carbon footprint of its assembly line — not for the emissions of the product it builds.

This is a piece about why that's the case. Three structural facts make the ESG industry's military blind spot possible: the way climate-disclosure frameworks define Scope 3, the way ESG ratings define industry materiality, and the way the EU's sustainable-finance regulation classifies "Article 8" funds. Each of those decisions, taken on its own, has a defensible technical rationale. Stacked on top of each other, they produce a system in which a fighter-jet manufacturer can hold a higher ESG rating than most consumer-goods companies while their products burn through the planet's remaining carbon budget.

The Scope 3 problem

Greenhouse gas accounting since the 1990s has used three "scopes":

• Scope 1 — direct emissions from a company's own operations (its factories, its trucks, its boilers).
• Scope 2 — indirect emissions from purchased electricity.
• Scope 3 — all other indirect emissions across the value chain, including the emissions a company's products generate when customers use them. (This is called "use-phase" Scope 3, and is category 11 in the GHG Protocol.)

Scope 3 is where the action is. For most product-making companies, Scope 3 is more than 70% of total emissions. The Science Based Targets initiative — the gold standard for corporate climate target-setting, adopted by ~96% of validated targets — explicitly requires Scope 3 inclusion when value-chain emissions exceed 40% of a company's total. Use-phase Scope 3 is the largest single category for sectors like oil & gas, automotive, aviation, and home appliances. ExxonMobil reports the emissions of the gasoline it sells. Toyota reports the emissions of the cars it manufactures over their operational lifetime. Boeing reports the lifetime emissions of its commercial aircraft.

Defense contractors do this inconsistently. Most don't disclose use-phase emissions at all. Some — most strikingly Airbus — disclose a comprehensive figure that combines their commercial and military aircraft into a single use-phase Scope 3 number. Lockheed Martin's CSR reports include partial product-use disclosures; the Conflict and Environment Observatory has noted the unevenness directly: "a company like Lockheed Martin lists the emissions of its product use, while others provide incomplete data." The carbon footprint of a Eurofighter Typhoon over its operational lifetime is not in BAE Systems' reported emissions. Most missiles, ammunition, and ground systems are nowhere. The disclosure landscape is not empty — it is patchy, inconsistent, and increasingly available, but it is not yet being incorporated into the headline ESG ratings and indexes that drive capital allocation.

The technical reason this exception exists is a piece of GHG-Protocol guidance that says use-phase Scope 3 should be reported when the manufacturer can reasonably project use patterns. Defense contractors argue — and ESG providers have largely accepted — that they cannot project how often a customer will fire a missile or fly a fighter, so use-phase reporting is impractical. This is, at minimum, an interpretive choice. Toyota cannot project exactly how many miles each Camry it sells will be driven, but it does report fleet-average lifetime emissions. ExxonMobil cannot project exactly when each barrel of crude it pumps will be combusted, but it counts the carbon as if it will be. The defense-contractor exemption exists not because the math is unusually hard, but because nobody has decided to do it.

The MSCI materiality problem

MSCI's ESG Ratings — the dominant rating product used by global asset managers, indexed to the GICS sub-industry classification — are explicitly industry-relative. A company is rated AAA (leader) to CCC (laggard) based on how it compares to peers in the same sub-industry on the issues MSCI's materiality model deems financially relevant for that sub-industry. The materiality weights are set annually based on what kinds of risks have historically destroyed shareholder value in that industry.

For Big Oil, the model has correctly identified climate transition risk as financially material — stranded-asset risk, regulatory risk, litigation risk all show up in the weights. For automotive, the same. For aviation, the same. For aerospace and defense, MSCI's industry materiality map flags energy efficiency and emissions, but specifically the operational emissions of the manufacturing process, not the use-phase carbon of the finished product. The reasoning is consistent within MSCI's framework: governments buy weapons, governments are not subject to the same climate liability as private fleet operators, the financial-materiality channel for product-use carbon is therefore weak. MSCI is not pretending the carbon doesn't exist. They have decided it doesn't yet show up in shareholder-value math.

To put those ratings in perspective: Apple's MSCI rating is BBB. Lockheed Martin's is AA. A company whose primary product is a fighter jet that burns 10,000 gallons of fuel an hour, used in a war that has incinerated 5.9 million barrels of oil at fuel depots, sits two grades above the company that makes the iPhone — because MSCI assesses Lockheed against other defense companies on the management-of-defense-relevant ESG risks, not against the universe of all companies on the absolute carbon footprint of its products. That is a defensible technical position. It is also, when you say it out loud, faintly ridiculous.

The "sustainable funds" problem

The European Union's Sustainable Finance Disclosure Regulation (SFDR), in force since 2021, classifies funds into three buckets:

• Article 6 — funds with no specific sustainability claim.
• Article 8 — funds that "promote environmental and/or social characteristics." Often called light green.
• Article 9 — funds with sustainable investment as their explicit objective. Often called dark green.

Article 8 is the bigger, vaguer, and more popular category — and the scale is genuinely large. Per Morningstar's Q4 2025 SFDR review, combined Article 8 + Article 9 assets reached €7.1 trillion at the end of 2025, representing 56% of the entire EU fund market. Article 8 alone accounts for the overwhelming majority — Article 9 is just 2.7% of total EU fund AUM. Globally (including funds sold outside the EU), MSCI counts 14,126 Article 8 funds with combined AUM of $14.1 trillion as of mid-2025. These are the products that retail investors, pension funds, and family offices generally believe are doing some kind of climate-aligned, sustainability-screened investing. The term has become the de-facto definition of "sustainable investing" in the EU retail market. The scale visualization at the top of the piece breaks down this universe and shows where defense exposure actually sits within it.

One important regulatory note for any reader following this in real time: the European Commission proposed a major SFDR overhaul in November 2025 that would replace the Article 8 / Article 9 framework with three new categories — "Sustainable," "Transition," and "ESG Basics." The Article 8 label may not exist in the same form by 2027. The structural critique below applies to the framework as it operates today and through whatever transition period the Commission ultimately adopts; it is also a piece of evidence in favor of the SFDR 2.0 reforms.

What has happened to defense exposure inside those funds since the start of the Ukraine war is striking — and is the actual scandal the ESG-reform discourse has been slow to confront.

Per Morningstar Sustainalytics' tracking, defense exposure in active EU Article 8 funds has risen from 0.6% to 2.5% — a 4× increase. In passive Article 8 funds, the rise is from 0.2% to 1.2% — a 6× increase. As of mid-2025, only 31% of Article 8 funds had any military-contracting exclusion policy at all. (Article 9 dark-green funds are stricter — 48% have exclusions — but they represent a much smaller pool of money.) MSCI's own analysis found that Article 8 funds, applying their broadest screening criteria, had up to 36% exposure to conventional weapons. The funds explicitly marketed to retail investors as sustainable have, in three years, materially increased their stake in the defense industry.

The justification offered by sustainable-finance practitioners is mostly about "defending democracy" — that supplying weapons to Ukraine, or to NATO allies more broadly, is consistent with social goals like peace and territorial integrity. That argument has real merit at the geopolitical level. It is not, however, an argument about emissions. A Rheinmetall artillery shell may or may not be politically justified depending on where it is fired and at whom; the carbon released by its manufacture and use is the same either way. The conflation of geopolitically defensible with climate-aligned is the move ESG practitioners have collectively made over the last three years, and the move that has allowed sustainable funds to quietly become significant defense investors without breaking their marketing materials.

What the major providers are actually saying

The original framing of this critique — that the ESG industry is silent on military emissions — turns out to be only partly accurate, and increasingly less so. The picture is more nuanced. Each major data and ratings provider has, by mid-2026, said something publicly about defense, controversial weapons, or Scope 3 disclosure gaps. None has yet made the structural connection between use-phase military emissions and headline ESG ratings the central reform issue. What follows is a structured map of where each provider stands today.

MSCI ESG Research

The most publicly engaged of the major providers. MSCI has published "Material, But Missing: Gaps in Scope 3 Reporting Persist" (May 2025), with the headline finding that only 13% of MSCI ACWI IMI constituents reported on the two Scope 3 categories most material to their business. Their "Rethinking Defense Exposure in Sustainable Funds" (June 2025) is the source of the Article 8 quintupling data, and "Interpreting Controversial Weapons" (February 2026) lays out the methodology choices that determine whether Lockheed Martin, Boeing, and similar names get excluded under broad versus narrow interpretations of Italy's Law 220/2021. MSCI's PAB Select index family explicitly excludes nuclear and conventional weapons; the standalone MSCI ESG Ratings (the AAA-CCC product) do not. In MSCI's "War and ESG" podcast, ESG Researcher Bentley Kaplan was unusually candid: "ESG is not a label of morality, it's a market signal." The position is internally consistent — and explicitly disclaims moral assessment of defense.

Morningstar Sustainalytics

Has the deepest defense-sector content of any major provider. "Defense: Assessing New Investment Opportunities Through an ESG Lens" (September 2025) analyzes ~860 of their ~23,000-company universe for military involvement and explicitly states that "the total military carbon footprint is estimated to exceed 5% of global emissions, a figure likely to rise". Their controversial-weapons screen is broader than the EU's standard list — it includes nuclear weapons, white phosphorus, and depleted uranium, where the EU PAB list does not. "EU ESG Funds' Exposure to Defense Continues to Increase" is the source of the 0.6%→2.5% Article 8 trajectory data. The framing throughout: defense ESG is a "case-by-case" screening question, with the practical advice that "firms that can evidence robust ESG safeguards and measurable outcomes" can attract sustainability-aligned capital. They have named the 5% military emissions figure publicly. They have not yet built it into a use-phase Scope 3 estimate that flows into their headline ESG Risk Rating.

ISS ESG (now ISS STOXX)

The strictest of the majors on weapons screening, the quietest on emissions analysis. ISS's Controversial Weapons Research covers approximately 27,700 issuers — the broadest weapons-screening universe of any major provider. They include nuclear-weapons programs in all five recognized nuclear-weapon states (US, Russia, UK, France, China), which is genuinely stricter than competitors. Their methodology aligns with the Mine Ban Convention, the Convention on Cluster Munitions, the Treaty on the Prohibition of Nuclear Weapons, and the NPT. The granular flagging system uses traffic-light signals based on involvement and verification depth. What ISS does not yet appear to have published is research connecting military fuel-burn or use-phase emissions to its ratings methodology. Their leadership is on the values-based weapons-exclusion side, not the climate-accounting side.

S&P Global Sustainable1 (CSA)

Aerospace & Defense (industry code "ARO") is one of S&P's 62 industry-specific Corporate Sustainability Assessment questionnaires. The methodology uses "double materiality" — both financial materiality and impact on society and environment — which is theoretically broader than MSCI's pure financial-materiality framing. The ARO questionnaire weights Scope 1 and Scope 2 emissions and operational eco-efficiency. The S&P Aerospace & Defense Select Industry Index, which tracks the GICS sub-industry, explicitly footnotes that its emissions metrics cover "operational and first-tier supply chain greenhouse gas emissions" — meaning it does not include downstream use-phase emissions of finished platforms. S&P's CSA does allow defense companies to submit non-public data privately, which is structurally different from MSCI's and Sustainalytics' public-only methodology. So far, no public S&P research has called out the military use-phase Scope 3 question as a sector reform priority.

Bloomberg

The broadest carbon-emissions data coverage of any provider — Scope 1, 2, and 3 estimates for over 130,000 companies, with an ML model leveraging 800+ features. Bloomberg has published "Closing the Scope 3 GHG Emissions Data Gap" noting that only ~20% of companies disclose Scope 3 voluntarily. The critical limitation: Bloomberg's Scope 3 estimation model only covers oil & gas and mining sectors. Defense is not in the Scope 3 estimation universe. Bloomberg sees its role as data aggregator and tooling provider, not methodology advocate. The data infrastructure is there to support a defense-sector Scope 3 estimate; the institutional decision to build one has not been made.

Clarity AI

The newer, AI-native entrant. Coverage of 70,000+ companies and 420,000+ funds, with climate analytics structurally powered by CDP data for emissions — meaning their climate position effectively inherits CDP's coverage decisions. Strong on regulatory compliance products (SFDR, EU Taxonomy, TCFD, TNFD) and on ML-based estimation and gap-filling. Their value proposition is data quality and regulatory operationalization, not structural reform advocacy. No defense-sector research is publicly available. They are positioned to follow whatever the market consensus becomes rather than lead it.

Civil-society and adjacent academic work

The actual analytical work on military emissions is happening outside the formal ESG-data infrastructure. The Conflict and Environment Observatory (CEOBS) runs the Military Emissions Gap project and has been writing letters to index providers asking for defense as a standalone tracking category. CEOBS's November 2025 analysis: military emissions reporting is getting worse, with the top three military spenders (US, China, Russia) either not submitting data or submitting incomplete data. The Initiative on GHG Accounting of War (Lennard de Klerk) produces the cumulative Ukraine war emissions assessments now used in formal climate-reparations claims at COP31. Brown University's Costs of War Project (Crawford et al.) publishes peer-reviewed analyses of US military and Gaza war emissions. Scientists for Global Responsibility in the UK has been publishing military emissions estimates since 2019. BCG projects defense emissions could rise from 2% of global CO₂ today to 25% by 2050 if contractors don't act. The methodologies exist. The integration into formal ESG products does not.

Why this matters now

The reason this is becoming a research priority is that the wars currently being fought are starting to make the omission untenable. Russia's invasion of Ukraine, Israel's war on Gaza, and the US-Israel war on Iran have generated, between them, about 575 Mt CO₂e in cumulative emissions. None of those emissions appears in any country's national greenhouse gas inventory or in any defense contractor's Scope 3 disclosure. Lockheed Martin's stock is up 40% YTD on the war. Within the broader SFDR Article 8 universe — the "light green" category that dominates EU sustainable retail flows — defense exposure has quintupled since 2022. Strict PAB-tracking products, to their credit, mostly do not hold Lockheed and similar names. But the same EU asset managers running PAB-strict products often run much larger Article 8 books that do, and the marketing language to retail clients rarely makes the distinction precise.

The carbon-budget arithmetic from the previous piece in this series — ~3 years until the 1.5°C budget is exhausted at current rates — does not allow the world's third-largest emitter to remain functionally undercounted for much longer. The data infrastructure that built the climate-aligned investing universe over the last decade was, in retrospect, built around a specific set of corporate emitters: oil majors, automotive, utilities, heavy industry, aviation. The defense sector was treated as a separate question — as weapons-screening rather than emissions-accounting — for reasons that made sense at the time and now make less sense every quarter.

The actual reform is mechanical rather than ideological: incorporate the disclosure that already exists, and produce estimates where it does not. Airbus discloses 443 Mt of use-phase Scope 3. Lockheed discloses pieces. CEOBS and IGGAW have published methodologies. SBTi has a Scope 3 threshold rule. CSRD is tightening EU disclosure. The pieces are increasingly in place. What's missing is institutional willingness — at the major ESG-data providers — to publish a use-phase Scope 3 estimate for a defense contractor and put it inside the headline rating that asset managers actually use.

Whether the providers choose to lead on this, or wait for the contradictions to become so obvious they're forced to follow, is the live question. As of mid-2026, on current evidence, the field is moving — but slowly, and at the edges of methodology rather than at the center of headline ratings. That is a position. It is also a choice. And like every choice in climate accounting, it can be unmade.