The All-In Neutral Price

What every basis point in a loan is actually paying for — a complete decomposition of the break-even rate, from the risk-free reference to the last cent of capital

A bank quotes a corporate client 7.17% on a five-year fixed-rate loan. The client sees one number. But behind that single rate sit at least six distinct costs, each measuring something different, each owned by a different function, each responding to different drivers. The risk-free rate. The cost of keeping funding in place. The price of the customer's right to prepay. The expected credit losses. The operating cost of origination and servicing. The return shareholders demand for bearing the residual risk.

Strip any one of these layers away and the loan is underpriced. Double-count any one and the bank prices itself out of the market. The discipline of funds transfer pricing is to name every layer, measure it independently, and stack them into a single break-even rate that answers one question: what is the minimum this deal must earn for the shareholder to break even?

That break-even — before any commercial margin, before any relationship concession, before any cross-sell subsidy — is called the all-in neutral price. It is neutral because it contains no profit. It is all-in because it contains every cost. And it is the number against which every lending decision should be measured.

What follows is a walk through each layer, in the order it stacks. No formulas — just the intuition behind why each layer exists, what it charges for, and what goes wrong if you get it wrong.

The yardstick: the risk-free reference

Before measuring anything, you need a zero. In loan pricing, the zero is the risk-free rate — the return on an asset that carries no credit risk, no liquidity risk, no optionality. Typically it is the government bond yield or the overnight indexed swap (OIS) rate at the loan's repricing tenor.

It is tempting to call this the "foundation" of the price, but that is slightly misleading. The risk-free rate is not a cost the bank incurs — it is the yardstick against which every other cost is measured. If the five-year risk-free rate is 4.38%, that number enters the stack not because the bank pays it, but because every layer above it is defined as a spread over it.

Why does this matter? Because two common errors flow from misunderstanding the reference. The first is using the bank's own funding rate as the zero — which buries the liquidity premium inside the reference and makes it invisible. The second is using the wrong tenor — reading the three-month rate for a five-year bullet loan, which ignores the term structure entirely. The reference must match the loan's repricing tenor: the horizon over which the rate is locked.

Layer 1: The cost of funds — and the franchise that subsidizes it

The first real cost layer answers the most basic question in banking: where does the money come from, and what does it cost?

A bank funds itself from two broad sources. Deposits — sticky, behaviorally stable, priced below market because customers accept a lower rate in exchange for convenience, insurance, and the relationship. And wholesale funding — market-priced, term-dependent, rolling at whatever spread the market demands on the day.

The cost of funds for a given loan is the blended cost of these two sources, weighted by the funding mix that applies to new origination. Not the whole-balance-sheet mix (which includes legacy funding), but the marginal mix — what actually stands behind the incremental loan being priced today.

This blending produces a counterintuitive result. The deposit leg typically costs less than the risk-free rate — because depositors accept below-market returns. That difference is the deposit franchise value: the economic benefit of having a deposit-gathering network. When deposits fund part of the loan, their below-market cost pulls the blended cost of funds below the risk-free reference.

In our illustrative example, the risk-free rate is 438 bps. The deposit franchise credit is −99 bps. The wholesale leg adds +66 bps (the term liquidity premium). Blended together: 305 bps — well below the risk-free rate. The deposit franchise is not just a nice-to-have; it is a pricing weapon that lets the bank quote below where a pure wholesale funder could compete.

The term liquidity premium inside the cost of funds

The wholesale leg deserves its own mention, because it carries a cost that deposits do not: the term liquidity premium. This is the price of keeping market funding in place for the full life of the loan.

If the bank funds a five-year loan with three-month commercial paper, it must roll that paper twenty times. Each roll is a repricing event — a moment when the market might demand a wider spread. In a crisis, that spread can triple. The term liquidity premium charges for this risk: the expected cost of rolling, plus a tail charge for the possibility that the rolls happen during a stress.

Three dials control the size of this premium. The wholesale share of the funding mix (the premium only bites on the part that rolls at market). The turnover frequency (more rolls mean more repricing events). And the spread volatility (how wild the market can move between rolls). Change any one, and the premium moves.

There is an alternative: if the bank can issue a five-year unsecured bond at a quoted spread, that spread is the term liquidity premium — locked in, certain, no tail risk. But not every bank can print term debt at every tenor, at size, on demand. For those that fund short and roll, the premium must be built from the rollover simulation, including the fat tail.

The seam: where Treasury ends and the business begins

At this point in the stack, a line is drawn. Everything below it — the risk-free reference, the deposit franchise credit, the term liquidity premium — belongs to Treasury. It is the transfer price: the internal rate at which Treasury "lends" funds to the business unit. Treasury collects this rate, and its P&L closes against it.

Everything above the seam belongs to the business unit that originates and manages the loan. Credit risk, operating costs, capital charges — these are the business's burden, and the business's margin must cover them.

Why does this separation matter? Because it ensures that each function is measured against the risks it actually controls. Treasury controls funding and liquidity — it should be judged on whether it funds the balance sheet efficiently. The business controls credit selection, origination quality, and relationship management — it should be judged on whether its loan portfolio covers its costs. Without the seam, the two are entangled, and no one can tell whether a bad quarter came from expensive funding or bad credit.

Layer 2: Contingent liquidity — the cost of standing ready

The term liquidity premium pays for scheduled funding: the rollover of known wholesale tranches at known dates. But what about the unscheduled outflows? The deposit run that was not in anyone's forecast. The credit line drawn by a corporate client precisely because markets are stressed. The collateral call on a derivatives book when rates gap overnight.

These contingent drains do not follow a timetable. They arrive when least convenient and demand immediate liquidity. The bank must hold a buffer of liquid assets — government bonds, central bank reserves — before the drain happens, ready to liquidate at a moment's notice.

Holding that buffer has a cost: the difference between what those liquid assets earn and what the bank's funding costs. This carry cost is real, persistent, and must be recovered from the loans that create the contingent exposure. A committed credit line, for instance, carries a heavier contingent charge than a fully drawn term loan, because the undrawn portion can be pulled at any time.

In our example, this layer is small — 4 basis points — because a fully drawn corporate bullet creates relatively little contingent exposure. For a revolving credit facility, where the undrawn commitment is the primary liquidity risk, this layer can reach 30–40 bps. The product's DNA drives the charge.

Layer 3: Optionality — pricing the customer's rights

Every loan contract hands the borrower at least one right that the bank cannot refuse. A mortgage borrower can prepay — refinancing when rates fall, leaving the bank holding the lower-yielding replacement at exactly the wrong moment. A corporate with a committed facility can draw the full line in a crisis, forcing the bank to fund the drawdown on the worst possible day. A floating-rate loan with a contractual cap limits the bank's upside when rates rise.

These are options, and they have value. The borrower holds them; the bank has sold them, usually for free, embedded in the loan contract. The optionality layer recovers that cost.

What makes this layer tricky is that the option's value depends on future states of the world. The prepayment option on a fixed-rate mortgage is worthless if rates rise (no one refinances into a higher rate). It is extremely valuable if rates fall 200 bps. The charge must reflect the probability-weighted average cost across all scenarios — not just today's rate environment, but the full distribution of possible rate paths over the loan's life.

For a five-year BBB corporate bullet with limited prepayment optionality, this layer is 12 bps. For a thirty-year fixed-rate mortgage with free prepayment in a falling-rate environment, it can exceed 30 bps. For a revolving credit facility with a large undrawn portion, the drawdown option alone can dwarf the prepayment cost. The right the customer holds determines the charge.

Layer 4: Credit expected loss — the defaults you can see coming

Every loan carries the possibility that the borrower will not pay back in full. Some fraction of loans — predictable over large portfolios, even if unknowable for any individual borrower — will default. Of those that default, some fraction of the outstanding amount will be lost after recovery efforts.

This is expected loss: the actuarial cost of lending. It is not a surprise — it is a statistical certainty over a large enough book. The question is only how much, and the answer comes from three inputs multiplied together: the probability of default (how likely this borrower's rating is to fail), the loss given default (how much is lost after collateral and recovery), and the exposure at default (what the balance will be when the default happens).

For a BBB-rated corporate borrower over five years, historical through-the-cycle data suggests a cumulative default probability of roughly 1.5–2.5%, with a loss-given-default in the 40–60% range for unsecured senior debt. The annualized expected loss — 135 bps in our example — is not a worst-case scenario. It is the average outcome, the cost of doing business in lending.

Two mistakes are common here. The first is pricing on point-in-time default estimates that react to today's sentiment (tight in good times, wide in bad). This produces pro-cyclical pricing: too cheap when credit is easy, too expensive when it is tight — exactly backwards. The fix is to use through-the-cycle estimates that reflect the full economic cycle.

The second mistake is confusing expected loss with unexpected loss. Expected loss is a cost — it belongs in the price, just like any other cost. Unexpected loss is a risk — it belongs in the capital charge. Double-counting by putting both in the spread, or omitting the capital layer because "credit is already charged," misprices the loan in opposite directions.

Layer 5: Operating cost — the forgotten layer

Every loan must be originated, documented, monitored, serviced, collected, and eventually closed. Every one of those steps consumes people, systems, and overhead. This is operating cost — and it is the layer most often either omitted from the transfer price entirely or allocated so crudely that it distorts the relative pricing of small loans versus large ones.

Consider the difference between a $50 million syndicated corporate facility and a $250,000 small-business term loan. The syndication might require 40 hours of structuring, legal review, and credit analysis. The small-business loan might require 8 hours — less in absolute terms, but enormously more per dollar of exposure. If operating cost is spread evenly across dollars, the small loan subsidizes the large one. If it is allocated by hours consumed, each product carries its own burden.

The 60 bps in our example reflects a mid-market corporate deal: moderate origination complexity, standard servicing, no unusual monitoring requirements. A branch-originated small-business loan might carry 120 bps. A digitally originated personal loan might carry 40 bps. The channel matters, the product matters, and the size matters.

Operating cost is not glamorous. It does not involve stochastic models or tail distributions. But a pricing framework that ignores it is a framework that implicitly asks the shareholder to subsidize the cost of running the business — and that subsidy comes directly out of the return on equity.

Layer 6: Cost of capital — the return equity demands

Expected loss covers the defaults the bank can see coming. But what about the ones it cannot? A recession deeper than any in the historical record. A sector-wide collapse. A concentration blow-up. These unexpected losses are what equity capital absorbs — and shareholders do not lend their capital for free.

The cost of capital layer charges the equity return hurdle on the regulatory capital the loan ties up. It has two components:

Credit capital. Under the standardized approach, a corporate loan typically carries a 100% risk weight. If the bank targets a 10.5% CET1 ratio, the capital charge on $1 of exposure is 10.5 cents, and the pre-tax return demanded on that 10.5 cents is whatever the board has set as the ROE target. For our example, that works out to 152 bps.

Interest-rate risk capital. A fixed-rate loan creates a duration mismatch between the asset (which earns a fixed coupon) and the funding (which reprices more frequently). Treasury hedges most of this gap with swaps, but a residual rate exposure remains — and that residual ties up capital too. For Avelmont's five-year fixed bullet, the rate-risk capital charge is 49 bps.

Together: 201 bps. This is the largest single layer in the stack, and it is the one most sensitive to regulatory changes. A shift from standardized to IRB risk weights, a change in the CET1 target, a recalibration of the ROE hurdle — any of these moves the capital charge by tens of basis points.

Expected loss is a cost you absorb. Unexpected loss is a risk you carry. The credit spread covers the first. The capital charge compensates for the second. Conflating them — by arguing that "the credit spread already accounts for risk" — is one of the most persistent errors in loan pricing.

What "neutral" means — and where profit lives

The stack adds up to 717 bps. But notice what is not in the stack: profit. The all-in neutral price is a break-even. It is the rate at which the loan pays for every cost it creates and every risk it carries, with exactly zero left over for the shareholder beyond the minimum return on capital.

The commercial margin — the actual profit — lives above the break-even. If the bank quotes 7.50%, the margin is 33 bps. If it quotes 7.00%, it is 17 bps short of break-even, meaning the deal destroys value unless compensated elsewhere (cross-sell revenue, relationship retention, portfolio diversification).

This is the power of the decomposition: it turns every pricing decision into a visible, defensible trade-off. The relationship manager who argues for a below-break-even rate must explain which layer they believe is overstated — or acknowledge that the deal is a conscious subsidy, approved by someone with the authority to grant it.

How the same engine produces different prices

The stack is universal — every loan passes through the same layers. But the intensity of each layer varies dramatically with the loan's profile. A secured mortgage, a corporate revolver, and a short-term working-capital line each light up different parts of the stack.

The profile drives the price, not the other way around. A mortgage's cost is dominated by the long-tenor funding layer and the prepayment option. A revolver's cost is dominated by the contingent liquidity charge and the drawdown option. A short working line's cost is dominated by operating cost per dollar (because the loan is small) and capital efficiency (because the commitment rolls frequently).

This is what the pricing engine does: it takes the loan's characteristics — tenor, repricing, collateral, embedded rights, credit quality, size, channel — and maps them through the same six layers. The same discipline, the same decomposition, completely different intensities.

Three principles underneath the stack

The decomposition rests on three principles that, if violated, collapse the entire framework:

Principle 1: Discount at one rate, carry each risk as a named line. Every cash flow in the loan is discounted at the single risk-free reference. Every risk — credit, liquidity, optionality, capital — appears as its own explicit layer in the stack. This prevents risks from hiding inside the discount rate, where they become invisible and unmanageable.

Principle 2: Price at the margin, not the average. The cost of funds uses the marginal funding mix (what new money costs today), not the book average (which includes cheap legacy deposits that will not be there next year). The credit charge uses through-the-cycle default rates, not today's point-in-time estimate. Every input reflects the forward-looking economic cost, not the backward-looking accounting cost.

Principle 3: Separate expected from unexpected. Expected credit loss is a cost — it belongs in the price. Unexpected credit loss is a risk — it belongs in the capital charge. Expected funding cost is the center of the spread distribution. Unexpected funding cost (the tail) is the risk premium charged above it. The distinction is not semantic; it determines whether the bank is charging for what it will lose or for what it might lose, and each demands a different treatment.

Why this decomposition changes the conversation

Without the stack, a loan rate is a single opaque number. "We quote 7.50%." Why 7.50%? "Because the market is there." What costs does it cover? "All of them, presumably." Is it profitable? "We think so."

With the stack, the conversation changes entirely. The rate is 7.50%. The break-even is 7.17%. The margin is 33 bps. Of the break-even, 201 bps is capital cost, 135 bps is expected credit loss, 66 bps is term funding, and the deposit franchise is subsidizing the rate by 99 bps. If the client pushes for 7.00%, the bank knows exactly what it is conceding: 17 bps of margin erosion, taking the deal below break-even. That concession requires a name, a signature, and a reason.

This is what the all-in neutral price delivers: not a number, but a language. A way to discuss, negotiate, approve, and audit every lending decision against the specific costs and risks it creates. The relationship manager speaks the same language as Treasury, who speaks the same language as Risk, who speaks the same language as the board.

And the language starts with a simple question that turns out to be anything but simple:

What is every basis point in this loan actually paying for?

Now you can answer it. Layer by layer. 717 basis points. Every one of them traceable.

The illustrative numbers use Avelmont, a fictional institution, with parameters calibrated to public data. The all-in neutral price framework follows the decomposition discipline described by Castagna and Fede. The specific layer values will differ for every institution, product, and market environment — but the structure of the stack is universal.