Blended finance

Can public money pull private capital into risky sustainable projects?

Caroline Flammer, Thomas Giroux, and Geoffrey Heal open up the black box of blended finance in their paper "Blended finance".

They pair a conceptual model with deal-level evidence from 173 blended finance transactions arranged by the World Bank's International Finance Corporation (IFC) between 2018 and 2023.

They analyse concessionality, i.e. the subsidy embedded in each deal expressed as a share of total project cost, and relate it to two measures of sustainability impact: the IFC's impact score and the number of Sustainable Development Goals each project targets.

Their main conclusions include:

• A one-standard-deviation rise in IFC-measured impact per dollar invested is associated with a subsidy that is 2.0% to 2.6% of project cost higher, confirming the core prediction of their model.
• Using the count of Sustainable Development Goals per dollar instead of the impact score gives a very similar effect, with a subsidy higher by 1.9% to 2.3% of project cost.
• Higher country political risk lifts the subsidy by 1.7% to 1.9% of project cost, and greater data opacity adds a further 1.2% to 1.6%, both tracking what private investors demand before investing.
• A one-standard-deviation increase in political risk raises the likelihood of a first-loss guarantee or similar provision by about 15.6%, while expected impact does not predict it at all.
• Theory says riskier, more opaque settings should attract more equity-like, first-loss positions, yet neither predicts the debt-versus-equity choice in practice, which the authors read as a sign of suboptimal contract design.
• Set against 595 non-blended IFC deals, blended transactions cluster in higher-impact projects and riskier, more opaque countries, with 50.9% of them sitting in Africa (nearly three times more than non-blended deals).

The paper offers a benchmark for what concessionality actually buys and argues subsidy budgets should depend on measured impact and country risk rather than negotiated deal by deal.

It also speaks to the debate on additionality and over-subsidy: concessionality should be no greater than necessary and paired with a public and concrete impact-to-subsidy relationship.

The evidence covers one institution and is correlational, so it describes how the IFC sets concessionality, not what the subsidy causes, and realised outcomes after deal completion remain unknown.