Can the conduct of sovereign wealth funds be attributed to the State? More specifically, can a State be held responsible under international law when a State-owned investment vehicle holds stakes in, controls, or enters into strategic partnerships with companies linked to human rights abuses, including extrajudicial killings, spyware, cluster munitions and nuclear weapons production, or fossil fuel extraction?
Sovereign Wealth Funds (‘SWFs’)—defined broadly here to encompass State-owned investment vehicles, administered separately from official foreign exchange reserves, and responsible for the management of a diversified portfolio of domestic and international assets—sit uneasily within the State responsibility framework. Since the first SWF (the Kuwait Investment Board) was established in 1953, the number of SWFs has grown to over 100, managing over $15 trillion dollars globally, and operating across multiple jurisdictions with varied governance structures that can blur the line between public and private actors and complicate legal characterisation.
At the same time, there are growing examples of these funds being involved, through direct investment, in human rights abuses. One notable example came to light in November 2024, when Human Rights Watch published a 95-page report on rights abuses linked to Saudi Arabia’s Public Investment Fund. According to that report, the Fund facilitated serious human rights violations through companies it owns and controls, including the 2018 murder of Saudi journalist Jamal Khashoggi, whereby two planes owned by a company under the Fund’s control were used by Saudi agents to travel to Istanbul, where the murder occurred.
Leaving aside issues concerning the primary rules engaged by such conduct, this post examines whether the customary rules of attribution, as codified by the International Law Commission’s (ILC) Articles on State Responsibility (ARSIWA) can adequately address these challenges. To do so, it identifies three broad (and non-exhaustive) structural models of SWFs.
SWFs as Autonomous Entities
One organisational model places the State as (often sole) shareholder in the SWF but establishes the fund as a legally and financially autonomous entity operating at arm’s length from the government. Share ownership may also be divided between organs of the State. For example, the Malaysian Minister of Finance Incorporated owns almost all of the shares in the Khazanah Nasional Berhad, with the exception of one share held by the Federal Land Commissioner.
These funds are typically established by statute and governed by a separate board responsible for investment policy and administration. Board composition can vary across funds. Some are made up entirely by independent professionals, others combine civil servants and external experts, while some are composed entirely of senior government officials. Board members are typically appointed, and may be removed by, the Head of State or a senior minister (though there are exceptions).
What prospects are there to attribute the conduct of such funds to the State? Article 8 ARSIWA – under which conduct is attributable where an entity acts “on the instructions of, or under the direction or control of, [the] State” – may seem the natural starting point, given the presence of ministerial oversight. However, aside from the difficulty of establishing the “effective control” standard required by the ICJ (particularly with the Court’s recognition of the legal separateness of corporate entities), the ILC decided that the conduct of State-owned entities (of which SWFs form part) “prima facie […] is not attributable to the State unless they are exercising elements of governmental authority within the meaning of Article 5” (Article 8 commentary, para 6).
Turning then to Article 5, the conduct of an entity will be attributable to the State where: (1) that entity is “empowered by the law of [the] State to exercise elements of the governmental authority”; and (2) the entity is “acting in that capacity in the particular instance”. The Article 5 inquiry assesses the powers of the entity, the way those powers are conferred onto it, the purposes for which they are exercised, and the extent to which the entity is accountable to the government (see Article 5 commentary, para 6). The ILC clarifies that an exercise of “governmental authority” will arise where an entity performs “functions of a public character ordinarily exercised by State organs” (Article 5 commentary, para 2) but otherwise does not define the term.
Though SWF objectives can vary, many SWFs are established by statute to stabilise the budget and economy from volatility in revenues, or to build up savings for future generations. Both functions can be considered governmental in character, but their pursuit through commercial means (such as portfolio management, which can also be performed by a private person) may complicate the assessment (Petrochilos, §14.41; §14.45). Though funds within this organisational model are typically subject to some degree of State oversight (such as through appointment powers or reporting obligations), the level of oversight differs and will ultimately depend on the specific governance structure. Accordingly, the application of Article 5 will be heavily dependent on the circumstances.
SWFs Within the Central Bank
A different organisational model places the SWF within the structure of the State’s central bank, though the degree of integration can vary. Some funds operate as relatively autonomous investment arms using the bank’s infrastructure, others are subject to the bank’s day-to-day operational control, with broader policy oversight by the State. A further variant, illustrated by China’s SAFE Investment Company, involves the fund operating as a subsidiary of an agency of the People’s Bank of China.
These differences are important in determining whether conduct is attributable under Article 4 ARSIWA, or whether the analysis must proceed under Article 5.
Under Article 4, the conduct of “any State organ shall be considered an act of that State”, whether executive, judicial, or legislative, and regardless of its functions, hierarchical position, or the commercial character of its activities (Article 4(1), and commentary paras 5-6). Whether the entity is an organ under the State’s internal law is irrelevant (Article 4 commentary para 11). An initial question is therefore whether the central bank within which the SWF operates qualifies as a State organ, turning on the precise relationship between the bank and the government (Crawford, p. 128, fn 83). Where the bank is sufficiently integrated into the State apparatus, and the SWF operates as a structural component of it (rather than a functionally independent subsidiary), the SWF’s conduct may be attributable to the State through the bank’s organ status without the need to establish direction or control (cf Trendtex).
What about where the SWF is functionally autonomous from the central bank, though it operates within the overarching framework? If the central bank qualifies as an organ of the State and knew (or must have known) about SWF investments giving rise to potential human rights abuses, it may be possible to hold the State responsible under Article 4 for its omission to act, per the ICJ in Corfu Channel (Crawford, pp. 117-8). In that case, the Court found that it was a sufficient basis for Albanian responsibility that it knew, or must have known, of the presence of mines in its territorial waters and did nothing to warn third States of their presence. Applying broader rules of State responsibility, therefore, where a central bank: (1) has organ status, and (2) has institutional oversight of a fund engaged in activity giving rise to rights abuses, the bank’s failure to exercise its supervisory authority may itself constitute conduct for which the State is responsible.
Where Article 4 is not satisfied, Article 5 may offer an alternative: central banks are an example of the State outsourcing governmental authority to a body formally separate from the government (Crawford, p. 128), though it will still need to be shown that the SWF forms part of the bank.
SWFs Under Control of the Ministry of Finance
A third organisational model places the fund under the direct control of the Ministry of Finance, which typically owns the fund, sets its mandate, and is formally responsible for its administration. Day-to-day management may nonetheless be outsourced to private asset managers operating under ministerial oversight—for example, Kiribati’s Revenue Equalisation Fund is managed by firms BlackRock and Northern Trust. Attribution will therefore in the first place depend on whether the conduct can be attributed to the Ministry of external managers.
Where the Ministry itself makes or directs the relevant investment decision, attribution under Article 4 can be straightforward: the Ministry of Finance is an executive body of the State.
More difficult are cases where operational management is delegated to private actors and conduct cannot be traced to a ministerial decision. Article 4 is unavailable. Article 5 is also unlikely to apply: private asset managers operate under commercial mandates, with authority deriving from a contractual relationship rather than a statutory conferral of public functions (in Kiribati’s example, external managers were appointed through a service legal agreement and investment management agreement). Moreover, portfolio management decisions do not obviously constitute an exercise of governmental authority.
That leaves Article 8, where attribution turns on the degree of ministerial control. General oversight or policy guidance will be insufficient. It is likely that investment decisions by sovereign entities will only be deemed to be at the direction or control of the State where there is an unequivocal, specific, and targeted factual basis to reach this conclusion. Where the Ministry sets broad parameters but leaves execution to managers, Article 8 is unlikely to be met. But where ministerial oversight extends to detailed investment decisions, there might be a clearer case for Article 8 attribution, though this will depend on close scrutiny of governance arrangements (which are rarely available for SWFs due to a lack of transparency).
For all the above cases, it is of course possible that the State may expressly acknowledge and adopt the specific conduct of the Fund as its own under Article 11, ARSIWA (see commentary, para 8). However, it is unlikely that State approval of the overall investment strategy will suffice, requiring instead an express acknowledgment and adoption of the particular transaction as the State’s own, which is rare.
Conclusion
Applied to the three structural models identified, it is clear that the customary rules on attribution are capable, in principle, of applying to SWF conduct, though the route varies significantly depending on how the fund is constituted and governed. But it is also clear that the current rules leave attribution gaps. SWFs offer only one example of where these gaps may be exploited through structural design. Attributing SWF conduct to the State may also be difficult due to a lack of evidence required for an attribution analysis. With six of the ten largest SWFs currently operated by authoritarian regimes (see also here), examples of harmful conduct slipping through the cracks may continue to arise in years to come.
Daisy Peterson is a DPhil Candidate in Law at the University of Oxford and Senior Research Officer at the Oxford Institute of Technology and Justice. Her doctoral research, supervised by Professor Philippa Webb KC, focuses on sovereign wealth funds as instruments of accountability in international human rights violations and is supported by the Arts and Humanities Research Council through the Open-Oxford-Cambridge DTP [ref. AH/R012709/1].