The DSRB Additionality Test: Against €418bn of European Defence Spending, Where a New Bank Can Still Add Capacity

A new bank only matters if it adds financing that would not otherwise exist. That is the additionality test, and it is the right way to read the Defence, Security and Resilience Bank. On 7 July 2026, at the NATO summit in Ankara, the leaders of Canada, Albania, Belgium, Greece, Latvia, Luxembourg, Romania, Türkiye and Ukraine announced a "shared intention to establish" the institution, with the ambition that it "commence operations as early as 2027". The question that decides whether it matters is not how large it is, but how much capacity it adds that Europe does not already have.

Reuters reported on 30 August 2026 that the project had secured about €5 billion in commitments, against a design contemplating roughly €20 billion of paid-in capital and €80 billion of callable capital. Those are the headline figures. Set them against European defence spending of around €418 billion, and the scale question inverts: €80 billion of callable capital is not large next to €418 billion of annual spending, so the bank's value cannot come from size alone. It has to come from doing something the existing system does not.

That existing system is no longer trying to finance rearmament through defence ministries alone. The European Investment Bank, the national promotional banks and the capital markets are already lending against the same demand. A multilateral bank adds capacity only where it reaches borrowers or carries risks the others will not: long-tenor lending, cross-border guarantees, callable capital that mobilises private money, or coverage of the countries and suppliers the incumbents quietly avoid. Everything else is re-badging.

There is also a statistical test hiding inside the political one. Whether a state's subscription counts as government spending or as a financial transaction turns on Eurostat's framework — the ESA 2010 manual, 2022 edition, which treats capital increases in multilateral development banks in a dedicated chapter and is conditional rather than permissive. If paid-in capital scores as deficit, political appetite shrinks; if it is a financial transaction, callable capital can be mobilised largely off balance sheet. The accounting is not a footnote here — it sets the ceiling on what the bank can actually do.

So the case for the bank rests on two joins that have nothing to do with its announced size. For a government, the institution is worth backing only to the extent it is additional. For a supplier or an investor, the signal is which financings it would do that no one else will, and whether its capital is classified in a way that lets it lend at scale. A €20 billion paid-in and €80 billion callable design is either a genuine new channel or a relabelling of capacity that already exists — and that difference is the entire argument.

The comparison that matters is with the lenders already in the field. The European Investment Bank has expanded its defence lending, the national promotional banks carry their own mandates, and private capital is chasing the same demand. A tenth institution is additional only if it finances beyond their combined reach — the projects too long-dated, too cross-border, or too politically exposed for the incumbents to take alone. If it lends against the same bankable projects everyone already wants, then €80 billion of callable capital is competition for deal flow, not new capacity for the system.

This analysis works through the additionality test deliberately, and leaves the reader with the questions that decide the money:

• Against €418 billion of European defence spending, what specifically would the Defence, Security and Resilience Bank finance that the EIB, national banks and markets would not?
• Does its paid-in capital score as deficit or as a financial transaction under ESA 2010 — and how does that change what it can lend?
• What has to happen for €5 billion in commitments to become €20 billion paid-in and €80 billion callable, and on what timetable?
• If the bank commences operations in 2027, which borrowers, tenors and countries define whether it is additional or duplicative?