Although generally positive, the 2025 OECD Development Assistance Committee (DAC) peer review of Australian aid, discussed on this blog by Cameron Hill, made several pointed observations about the Department of Foreign Affairs and Trade’s (DFAT) reliance on a “small pool” of managing contractors to run its largest, most complex programs, which could come at the expense of strategic, innovative programming, value for money, proportionate risk management and direct engagement with local partners. In the language of Australian aid, that’s a warning about “facilities”.
A facility, in this context, is a large, multi-year delivery platform under which a single managing contractor is engaged to run a country or sector program on DFAT’s behalf. Rather than DFAT contracting many discrete projects, it hands a broad mandate and a budget, often in the hundreds of millions of dollars over five to ten years, to a commercial firm that stands up an in-country team, commissions activities, sub-contracts local partners and reports on results. The model is meant to buy scale, continuity and flexibility; in practice, it also concentrates delivery in a small number of firms.
Facilities are not a new subject for this blog; the contractor model has been examined many times (here, here and here, for instance). Commercial contractors delivered around 28% of Australia’s official development assistance (ODA) in 2024-25, up from roughly 16% a decade earlier, absorbing functions once performed in-house by DFAT. The market that delivers them is concentrated, as the Centre’s own procurement analysis has shown — that’s the least contested part of this story. The facility is no longer one delivery option among many; for large, complex programs, it has become the default.
That makes the peer review’s questions the right ones to ask. Do facilities deliver value for money, or does scale mask inefficiency? Do they enable adaptive programming, or lock DFAT into a narrow supplier base? Does the model advance DFAT’s own commitment to “locally led” development, or run counter to it by routing local partners through a foreign-owned contractor?
Here the evidence is genuinely mixed. The most authoritative independent assessment remains the Australian National Audit Office’s 2020 audit of value for money in facility arrangements, covering two of the largest facilities in Indonesia and Papua New Guinea, which found design, procurement and partnering largely sound, but inconsistent evidence that value for money was assessed at activity level, and real gaps in evaluation, reporting and cost transparency. At the time, DFAT ran 20 commercial facilities worth close to AUD2.8 billion. Four providers delivered around 97% of that value, and administrative costs consumed an estimated 14-21% of expenditure. All this is money that could otherwise have funded direct grants to local partners, grants to Australian government and non-government agencies to support their overseas partnerships, government-to-government support, or multilateral channels.
Academic work points in the same direction, but for different reasons. Meg Taylor and Solstice Middleby found Pacific aid has been steadily privatised through a handful of large corporations, marginalising local firms and expertise, while Lucy Pennington’s study of PNG infrastructure delivery found facility-run programs pursue Australia’s strategic and diplomatic interests as much as partner-country goals. Mark Opoku Amankwa and colleagues’ systematic review of Australia’s localisation efforts found that even where facilities reference localisation, implementation stays donor-driven and resistant to local agency. The Publish What You Fund’s May 2025 study found that DFAT channels just 6.2% of project-type funding directly to local organisations, though this is a little above the average among five comparable donors (5.5%). Together, value-for-money claims are based on thin evidence, localisation delivers less than claimed, the national interest does quiet work within the development architecture, and facilities have become the default partly because DFAT staff know how to procure and manage them and because they shift the risk onto a single contractor, not because they’ve outperformed alternative delivery modalities.
The record also holds clear counterexamples. The Market Development Facility (MDF) in Timor-Leste was found by independent evaluation to have delivered real gains in the coffee and pig market systems it worked in, and is known for the rigour of its results measurement and capacity to adapt. Nor is MDF isolated; the independent evaluation of the Pacific Labour Facility (PLF), completed in late 2023, found PLF delivered effective and efficient services to Pacific Australia Labour Mobility (PALM) scheme stakeholders in culturally appropriate ways, across a rapidly growing, complex program — and its scope specifically assessed the appropriateness of PLF’s monitoring, evaluation and learning (MEL) arrangements. Two facilities, different countries and sectors, different evaluation teams, similar verdicts.
My view, shaped by working within or evaluating facilities for more than a decade (including the Australia-Africa Partnership Facility, the Australia-Mekong Water Facility, the Pacific Labour Mobility Facility and the Kiribati Facility), is that MEL is the hinge on which facility performance turns. Facilities with credible evaluation systems that support flexible yet accountable delivery balance learning and adjustment; facilities where MEL is bolted on late or abandoned once senior management stops treating it as a priority drift, generating the failures the peer review has in mind.
That’s not just my own impression. The recent independent evaluation of Prospera, DFAT’s AUD277 million economic governance facility in Indonesia, found much the same pattern. The program’s MEL framework was, by the evaluators’ own account, barely used — “many staff members, including senior leadership, were unfamiliar with the [MEL] Framework and its intended purpose” — and strategic decisions were struck verbally and left undocumented, leaving evaluators reliant on staff self-assessment. The shortfall wasn’t cosmetic: the evaluation found it had “reduced the relevance” of the program’s own outcomes and left it unable to demonstrate what its advisers were contributing. The variation across DFAT’s aid portfolio is not random; it correlates with the extent to which MEL is designed and implemented.
If my reading is correct, the policy conclusion is to treat facility-or-not as a genuine choice, tested on a case-by-case basis against realistic alternatives. Where a facility is the best option, fix the conditions: invest seriously in MEL rather than treating it as overhead; protect genuine flexibility rather than contracting it away; push localisation beyond local content towards real agency for local partners; and hold value-for-money claims to a higher evidentiary standard.
None of this lets the model off the hook. The evaluation base remains thinner than it should be, so the peer review’s call for more systematic evaluation is welcome. DFAT’s 2024-25 Performance of Australian Development Cooperation report, published this year, includes its inaugural perceptions survey — though perceptions are no substitute for the rigorous independent evaluation the model still needs. The facility model in itself is neither the problem the peer review fears nor the all-purpose solution its defenders sometimes claim it is; the real problems are treating it as the default without considering whether alternative implementation modalities would be more appropriate, and not sufficiently integrating MEL from the outset.
