Why Blended Finance Underperforms: A Structural Diagnosis of the Catalytic Claim

Blended finance is endorsed across development and impact capital as the mechanism that mobilises private investment toward public goods, using concessional capital to de-risk commercial returns. Yet after a decade of institutional uptake, mobilisation has stagnated and the central catalytic claim remains largely unverified. This paper argues the underperformance is structural rather than operational: blended finance assembles its instruments from single-cycle, liability-bearing components — tranches, guarantees, first-loss layers — and applies them to objectives that are inherently multi-cycle, realised across a sequence of transactions rather than at the resolution of any one. Three failure modes follow from that single mismatch. Horizon collapse consumes concessional capital within the cycle rather than recycling it across cycles. The de-risking paradox biases deployment toward the transactions that least need concession, because subsidy and catalysis are distinct acts and only the first is observable within a cycle. Measurement at the deployment boundary renders the catalytic claim unfalsifiable in the field's own terms. Each is endemic to the instrument design and each would persist under flawless execution. âš ï¸ Two of the three name phenomena the field has already reported, and the contribution is the derivation rather than the observation. What this paper supplies is why both follow from one property: only subsidy is recorded within a single transaction cycle, so an instrument resolving within that cycle leaves catalysis with no trace to be estimated from. âš ï¸ The claim is about the record, not about identification. Catalysis is a counterfactual, and no quantity of elapsed cycles identifies one by itself. What a single-cycle instrument forecloses is the data — it produces no sequence of outcomes against which any design could estimate. So no improvement in evaluation discipline can move a boundary set by what the instrument records. The paper closes by specifying the conditions any non-extractive, multi-cycle alternative would have to satisfy, each derived by inverting a failure mode.