Securitising banks cut lending harder after a tightening, not less
Serguey Asael Shinder2 min readECB Working Paper No 3289, September 2026, by Dorian Henricot and Enrico Sette, produced within the ChaMP research network, asks whether securitisation changes how monetary policy travels through banks. Using loan-level euro area data, it finds that banks actively engaged in securitisation adjust credit supply more strongly in response to monetary policy shocks than a matched sample of non-securitising banks.
The mechanism is the interesting part. Securitisation expands a bank's lending capacity, but it does so by increasing reliance on investors whose required returns and risk appetite are more sensitive to monetary conditions than a bank's own funding is. After a tightening those investors demand higher compensation and reduce their exposure to securitised assets - so the securitising bank contracts lending by more than its peers.
Three further findings sharpen it. Effects are stronger for the loans most likely to be securitised - safer borrowers, longer maturities. They are primarily driven by synthetic securitisations, which deliver additional capital relief through Significant Risk Transfers. And firms borrowing from securitising banks cannot fully substitute the tighter supply, either through existing relationships or by finding new ones.
The paper carries the standard disclaimer: the views are the authors', not necessarily the ECB's.
What it means
This inverts the usual argument for securitisation as a stability feature. The case normally made is that moving risk off balance sheets frees capacity and makes credit supply more robust; this paper finds the opposite at the margin that matters for policy - the funding it substitutes in is more rate-sensitive than what it replaces, so the amplification is a feature of the channel rather than a failure of it.
The synthetic-securitisation result is the one to hold onto. An SRT is capital relief, which is exactly the constraint that binds lending capacity, and it is purchased from investors whose appetite is priced off the same rates the central bank is moving. That makes the capital relief procyclical by construction: most available when conditions are easy, most expensive precisely when a bank would want to keep lending. Anyone modelling euro area credit supply through a tightening should treat an SRT-heavy bank as a higher-beta lender, not a better-capitalised one.
And the substitution finding is what turns this from a bank-level observation into a real economy one. If borrowers could switch lenders the aggregate effect would wash out; they cannot, so the amplification reaches firms. For the ECB the implication is uncomfortable in a useful way: as the securitisation market grows - which EU policy actively encourages - the transmission of a given policy move gets stronger and harder to calibrate.