Unconditional cash flow modelling
Regulatory Definition
Cash flow modelling under the assumption that the timing and amount of cash flows is independent of the specific interest rate scenario.
EBA GL/2022/14
What This Actually Means
Your cashflow projections are the same regardless of the rate scenario. Prepayment speeds stay constant, deposit volumes don't move, pipeline conversion is fixed. You calculate one set of cashflows and then just discount them under different rate curves.
Terminology nuance: strictly speaking, the definition makes all floating and managed rate instruments conditional — because changing rates mechanically alters their interest cashflows. But in practice the conditional/unconditional distinction is used more narrowly, to describe whether behavioural and strategic variables respond to the rate scenario: prepayment speeds, pipeline completion rates, early deposit withdrawals, deposit migration between products, rate-dependent new business volume forecasts, FTP forecasting, and balance sheet balancing assumptions. When someone says they're running unconditional cashflows, they almost always mean these behavioural variables are held fixed — not that floating rate cashflows are literally unchanged.
Where It Matters
Simpler to implement and explain, but misses the asymmetric risk. In practice, customer behaviour IS rate-dependent — that's the whole point of option risk. Using unconditional modelling for EVE means you're likely understating your tail risk in scenarios where behavioural optionality is most pronounced.
The terminology trap: because the strict definition technically makes floating rate instruments conditional, teams can end up in definitional arguments about whether their model is conditional or unconditional. The more useful question is which behavioural and strategic variables are being held fixed vs. rate-conditioned — prepayments, deposit migration, new business volumes, FTP forecasting, balance sheet composition. A model that conditions floating rate cashflows mechanically but holds all behavioural variables static is functionally unconditional in the sense that matters for risk measurement.
Hedgable vs. non-hedgable risk: much of the exposure generated by conditional cashflow modelling — behavioural prepayments, deposit migration, rate-dependent new business volumes — is non-hedgable in any practical sense. You cannot put on a swap to hedge the risk that customers repay early faster than modelled, or that deposit migration accelerates beyond your assumptions. There is therefore a real tension between model sophistication and actionability. A highly conditional model may produce a more accurate picture of total interest rate risk, but if most of that risk cannot be hedged, the metrics become difficult to act on and risk appetite limits become hard to set meaningfully.
Critically, these non-hedgable risks are not really IRRBB issues at all. The risk that prepayment speeds deviate from model, that deposits migrate faster than assumed, or that new business volumes disappoint in a given rate environment — these are business risks that originate in, and should be owned by, the originating business areas: the mortgage business, the retail deposits business, the product teams. Pulling them into an IRRBB framework not only creates unactionable metrics for the ALM desk, it can also obscure accountability by making a business performance problem look like a treasury risk management problem.
The practical design principle: if you build a complex conditional model, design it from the outset so that hedgable and non-hedgable exposures can be bifurcated. Report them separately. This allows the ALM desk to manage the hedgable component with precision while the non-hedgable component is returned to the business areas that own it — rather than everything being blended into a single metric that sits with the wrong team.