How to charge for liquidity in your FTP framework

What the liquidity component charges for#

The liquidity component of a transfer rate is the charge a bank makes for the cost of being able to fund a position for as long as it may need funding, and for holding the buffer that covers what might be drawn or withdrawn unexpectedly.

It is separate from the term rate, which prices the shape of the yield curve, and separate from credit spread, which prices the borrower.

Where the liquidity component is set to zero — still the most common arrangement — the products that consume the most liquidity appear to consume none, and the ones that supply it get no credit for doing so.

The term liquidity premium#

The first component is the term liquidity premium: the additional cost the bank itself pays to borrow for a given tenor, above the reference rate for that tenor. It exists because a bank is not a risk-free issuer, and because the market charges more to lend to it for five years than for five days.

Deriving it means looking at the institution’s own funding: the spread over the reference curve on its issued debt, on wholesale term funding, and on term deposits sold at competitive rates.

Where a bank has no term issuance, the premium must be inferred from the rates it would have to pay to raise term money, which is a judgement — and one worth documenting, because the alternative in practice is assuming it is zero.

Contingent liquidity#

The second component is contingent liquidity: the cost of standing ready for outflows that are possible but not scheduled. Committed but undrawn facilities, overdraft limits, guarantees, letters of credit and collateral agreements all create the obligation to produce cash on demand without generating any funding in the meantime. So does a non-maturity deposit book, part of which could leave at any time.

The bank meets these possibilities by holding liquid assets that yield less than its lending — and the difference between what that buffer earns and what it costs to fund is a real, recurring expense that has to be allocated to whatever created the need.

Charging in proportion to the buffer#

Allocating that buffer cost is where frameworks most often stop short. The defensible approach is to charge each product in proportion to the buffer it obliges the bank to hold, using the same assumptions the liquidity assessment uses:

• a drawdown rate for undrawn commitments
• an outflow rate for each deposit segment
• a collateral-call estimate for derivative and guarantee exposures

The charge then falls on the business that wrote the commitment rather than being absorbed centrally, which changes how undrawn facilities are priced and, more usefully, how readily they are granted. Banks that allocate this cost properly are frequently surprised by how much undrawn capacity they had been giving away.

Regulatory ratios#

Regulatory ratios add a third dimension, because the same product can be cheap in economic terms and expensive in ratio terms, or the reverse.

A deposit that behaves stably but attracts a high assumed outflow rate under the standardised liquidity ratio consumes buffer the bank would not otherwise hold. A loan whose maturity falls just beyond a ratio boundary requires stable funding that a slightly shorter one would not.

Reflecting these effects in the transfer rate means computing the marginal ratio impact of each product and charging for it — which is what makes the front line’s pricing consistent with the constraint the treasury is actually managing to, instead of leaving treasury to absorb the difference quietly.

What a six-year-old current account is worth#

Behavioural maturity governs how much liquidity credit a deposit actually earns. A current-account balance that has sat undisturbed for six years is, in liquidity terms, closer to term funding than to overnight money, and a framework that credits it at an overnight rate undervalues the franchise that gathered it.

The reverse is also true: a large rate-sensitive corporate deposit that has never been tested by a competitive offer should not be credited as though it were core.

Splitting each non-maturity book into a stable core and a volatile portion, and pricing each part against a different point on the curve, is the standard treatment — and the split is a modelling judgement carrying real money, which is why it belongs under formal validation.

The options the bank has written#

Optionality is the component most often forgotten because it costs nothing until it is exercised.

A borrower who may prepay a fixed-rate loan holds an option the bank has written, and that option has value whether or not it is used: prepayments arrive precisely when refinancing is cheap, meaning the bank loses its highest-yielding assets in the market where replacements yield least. Depositors who may break a term deposit hold the mirror image.

Pricing these options into the transfer rate is what prevents a product that looks profitable at origination from becoming a systematic loss across a rate cycle.

One institutional view of the balance sheet#

Consistency with the liquidity assessment is not optional and is frequently absent.

The drawdown rates, deposit outflow assumptions and behavioural splits used to set liquidity charges in the FTP framework describe the same customers as the ones in the ILAAP, and where the two sets differ, at least one of them is wrong.

In practice the FTP assumptions tend to be the more optimistic, because they sit in a commercial process and the liquidity assessment sits in a prudential one. Reconciling them is a short exercise with a long payoff: it forces a single institutional view of how the balance sheet behaves, and it removes an inconsistency that reviewers find quickly once they think to look.

A three-product test#

A practical way to test whether liquidity is being priced at all is to look at what the framework says about three products.

1. An undrawn committed facility should carry a positive charge; if it carries none, contingent liquidity is unpriced.
2. A ten-year loan funded by overnight deposits should carry a term liquidity premium visibly larger than a one-year loan; if the difference is small, the premium is nominal.
3. A long-standing retail current account should earn a materially better credit than a rate-shopping corporate deposit of the same size; if they are treated alike, behavioural modelling is not feeding the mechanism.

Three answers, and most of the diagnosis is done.