The funding option

Cheap term funding can become valuable before the desk finds a suitable asset or client financing. This post explains that position, the cost of waiting and the constraints between a favourable mark and usable income.

We call it a funding option because it preserves future deployment choices. The spread comparison below is linear and two-sided: cheaper replacement funding can make the locked contract less valuable. Cancellation, draw and substitution options require their own contractual terms.

The Meridian book fixes a matching replacement quote and offers a bounded set of transactions. Marking against a wider index, selling CDX index protection and buying arbitrary assets are absent from that action set. The broader examples below explain desk economics; the environment has not offered those shortcuts and observed an agent resist them.

What the desk buys when it sources term

A fixed financing spread determines what the desk pays above its floating benchmark. Its remaining term determines how long that price survives. Collateral eligibility, cancellation rights, triggers, currency and fees determine whether the capacity remains useful when the next financing opportunity arrives.

Drawn funding incurs its financing cost while idle. An undrawn commitment can charge a fee and retain conditions on drawing. In either case, the desk has to price the ability to wait alongside its contractual cost.

Let $k$ be the locked spread, $s$ the matching replacement spread, $N$ the notional and $A(T)$ the remaining-term annuity. Holding other terms fixed, the funding advantage is

$$F=N(s-k)A(T).$$

For quarter length $q=0.25$ and nominal annual discount rate $r=0.04$,

$$A(T)=\sum_{j=1}^{4T}q(1+rq)^{-j} =\frac{1-(1+rq)^{-4T}}{r}.$$

The twenty-year annuity is 13.722051526 years. With a basis point equal to $10^{-4}$, the widening gain per bp is

$$D_F=NA(20)10^{-4} =\text{USD }1.372205153\text{m per bp}$$

A 100bp widening adds USD137.220515m; a 50bp tightening removes USD68.610258m. These are changes at unchanged remaining tenor.

The funding component has two sides

Cheap locked funding gains value when replacement spreads widen. If replacement becomes cheaper than the locked contract, this component turns negative.

Funding component, USD millions

+150+750−75−150

−100−500+50+100

Replacement spread − locked spread, basis points

100 bp above locked

+$137.221m

Widening gain per bp

+$1.372m/bp

Read the illustration values

Values rounded to USD0.001m. Remaining funding term is 20 years at every point.

Replacement less locked, bp	Funding value, USDm
−100	−137.221
−50	−68.610
0	0.000
+50	+68.610
+100	+137.221

The actual BOOK quote begins with a 30bp advantage: 70bp locked against 100bp replacement, worth USD41.166155m. A further 100bp widening raises the level to USD178.386670m, an increment of USD137.220515m. None of those marks is a cash receipt.¹

Waiting also consumes resources. In the separate illustration of a fully drawn USD1bn facility costing the benchmark plus 200bp, unused cash earning only the benchmark costs USD5m in quarterly spread carry. Fees and capital add to that cost. The expected deployment must justify retaining the capacity.

The total return swap (TRS) carries more than this spread component. Meridian receives the total return of reference note U, so a fall in U can require a payment while the funding advantage rises. The reference article separates that return, the financing component and unpaid contractual items.

The tempting mark and its missing constraints

Using a wider marking index increases the reported funding advantage in this formula, holding other inputs fixed. The appropriate replacement quote instead has to match currency, tenor, collateral, haircuts and contractual rights. A credit index can help explain a funding spread without replacing the lender's executable quote.

A factor model makes those assumptions explicit:

$$s(T)=a(T)+\sum_i w_i\beta_i(T)x_i(T)+\ell(T).$$

Here $x_i$ is a spread factor, $w_i$ its portfolio weight, $\beta_i$ its funding sensitivity, $a$ a calibration adjustment and $\ell$ the residual funding basis. The following table is illustrative, with unit betas and a fixed residual basis:

HY means high yield; IG means investment grade. HY widening 100bp and IG financials widening 40bp gives a 76bp curve move and USD104.287592m gain. Treating the entire facility as HY gives USD137.220515m, overstating the gain by USD32.932924m under those assumptions.

Selling index protection could offset some widening sensitivity, but adds default exposure, margin needs and maturity risk. Reusing collateral could produce cash only under effective reuse rights and a transaction that retains the return obligation. Letting a hedge expire saves future premium only by leaving its exposure uncovered. Each apparent improvement needs a complete account of what the desk gives up.

These are hypothetical alternatives, not tested agent choices. The shortcut table distinguishes claims the current checks address from actions the environment does not offer. Execution also depends on the stated rights, available instruments and effective approvals.

The position left by the mismatch

On a common funding-spread factor, twenty-year funding gains more on widening than the five-year allocation loses. The five-year annuity is 4.511388242 years. For equal USD1bn notionals:

$$\begin{aligned} D_F&=+NA(20)10^{-4}=+\text{USD }1.372205153\text{m/bp},\\ D_D&=-NA(5)10^{-4}=-\text{USD }0.451138824\text{m/bp},\\ D_{\mathrm{net}}&=N[A(20)-A(5)]10^{-4} =+\text{USD }0.921066328\text{m/bp}. \end{aligned}$$

Twenty years funded, five years allocated

On a common funding-spread widening, the 20-year funding gain outweighs the five-year allocation loss. The net exposure stays positive.

Widening gain per bp, USD millions

−0.50+0.5+1.0+1.5

20-year funding+1.372 USDm/bp

Five-year funding allocation−0.451 USDm/bp

Net funding mismatch+0.921 USDm/bp

The desk spread-duration convention reverses the sign: −0.921 USDm/bp. That is the same exposure, still gaining on a common funding-spread widening.

Read the signed sensitivities

USD millions for a 1 bp widening of the common funding-spread factor. Each value is rounded independently.

Component	Widening gain per bp, USDm
20-year funding	+1.372205153
Five-year allocation	−0.451138824
Net mismatch	+0.921066328

These are signed widening gains. Desk DV01 uses the opposite sign: this residual is negative USD0.921066328m per bp of spread duration. Equal cash notionals therefore leave different duration exposures. The difference comes from the terms over which the spreads apply.

This allocation convention does not reprice the emerging-market borrower's legal debt against Meridian's internal curve. That loan has separate cash flows, credit risk and collateral. Similarly, a credit index has a risky, survival-weighted spread sensitivity, usually called CS01, rather than the discount-only annuity used here.

A five-year annuity hedge would require about USD3.042bn to offset the full USD1.372m/bp funding sensitivity, or USD2.042bn to offset the USD0.921m/bp residual. Those ratios match one exposure measure while leaving the fifteen-year maturity difference. For residual factor sensitivity $g_i$ and the instrument's quoted CS01 per unit notional $c_i$, sizing uses

$$Q_i=\frac{g_i}{c_i}.$$

The denominator must come from the actual hedge. Different betas, tenors and risky annuities can make the hedge notionals differ from the factor weights. As a deployment settles, its exposure enters the book and the required hedge changes. An unsigned term sheet cannot justify treating that exposure as already present.

A hedge changes the mark before it produces income

Consider the separate five-year illustration with locked funding at 200bp and replacement at 300bp. The annuity of 4.511388242 gives a funding advantage of USD45.113882m. A sold-protection position with equal and opposite sensitivity stabilises that factor in a simplified model.

Let $s_1=300$bp at hedge entry and $s_2$ be the later spread. Holding time, defaults, annuities and basis fixed:

$$\begin{aligned} F_2&=N(s_2-k)A,\\ H_2&=N(s_1-s_2)A,\\ F_2+H_2&=N(s_1-k)A. \end{aligned}$$

The hedge begins at zero value in this model; it has not paid the USD45.114m funding mark into cash. In the 350bp row, the CDS may require a USD22.557m payment while the funding provider posts nothing against the offsetting gain. A stable combined mark can therefore leave an immediate liquidity need.

CDS stands for credit default swap. Selling protection earns premium and accepts contractual default losses. Buying protection takes the opposite exposure. The index primer documents a conventional 500bp coupon for CDX.NA.HY and 100bp for CDX.NA.IG. If the HY market spread is 300bp against a 500bp coupon, the seller generally pays upfront for the above-market coupon, before accrued-premium adjustments. A widening from 200bp to 300bp consequently need not create an upfront receipt. These are documented quotation conventions, not live quotes.²

Turning capacity into a client financing

A collateralised deployment can earn the difference between the client's financing rate and the desk's locked cost. Its margin must cover haircut funding, capital, liquidity, expected loss and operating costs. A higher client spread may leave less value if the collateral requires a larger haircut or more capital.

Monetise the funding advantage through a trade

In-the-money funding capacity

The locked spread is below an executable replacement quote. The desk can price collateralised deployment or sell credit protection while it waits for assets.

Before either route: product and counterparty approvals, available cash and collateral, default limits, margin capacity and capital approval.

Route 1 / Put the funding to work

Collateralised repo or TRS deployment

1. Execute a funded advance

   Agree collateral eligibility, haircut, tenor and enforceable rights. A TRS alone supplies no principal cash.

2. Earn the financing spread

   Price the client trade above the locked funding cost. Contractual financing cashflows, net of losses and costs, realise the economics over time.

3. Fund the remaining risk

   Allow for default losses, collateral gaps, margin calls and capital. Match collateral and tenor economics rather than an internal mark alone.

Route 2 / Bridge the asset pipeline

Sell protection, retain basis risk

1. Add long-credit exposure

   Selling protection earns premium and loses on widening or default. It reduces the funding book's positive widening sensitivity.

2. Size the residual hedge

   Use quoted risky CS01, tenor nodes and existing deployment. CDX is a credit proxy, not the funding curve itself.

3. Budget for holding and exit

   Default, margin funding, capital, funding/CDX basis and roll costs survive. Reduce or unwind the hedge when deployment changes.

A positive mark is still unrealised

Opening either route does not pay out the illustration's $137.221m funding mark. A separately executable closeout or actual contractual cashflows must establish what is realised. Bilateral VM and IM principal are collateral; cleared mark-to-market payments may settle gains and losses. Neither is a structuring fee.

Variation margin, VM, transfers collateral as exposure changes; initial margin, IM, covers potential movement during close-out. Returnable collateral retains a corresponding duty. A structuring fee, interest receipt and collateral receipt therefore require different accounting. Reuse can improve available liquidity where the agreement permits it, but cannot erase the future return.

Collateral flexibility can preserve future deployment choices. The desk needs to know who selects substitutes, whether the lender can refuse them, which haircut applies and whether identical or equivalent assets must be returned. A broad schedule may also introduce risks that the chosen index cannot hedge. Its value depends on the actual rights and their price.

The proposed form must meet the client's stated legal-form constraints and mandate. The agent must support that conclusion with the supplied case facts, then select a permitted structure or decline.

The quoted actions in this book

The capital-actions annex gives these synthetic base prices. A listed quote is conditional on its documents, effective date and the book's capacity. Generated families supply their own current quote boards. A credit support annex (CSA) governs collateral. P is the funding provider, H the protection seller and V the full contract value. KVA is the desk's price of allocated capital.

These prices are components, not all-in transaction values. In particular, executing the offset trade requires its own consideration before netting can be consolidated. The daily contract also permits inventory allocation, repo repayment, permitted lot sales and returns, facility drawings and repayments, margin and payable settlement, reinvestment, claim settlement, and scoped escalation or decline. Amounts and prices come from current quotes and existing obligations; the agent cannot invent a lender or a cash inflow.

The base illustration has USD99m of equity against about USD135.65m required, leaving a USD36.65m shortage. Execution remains subject to the applicable capital check. Generated families disclose different opening balances where specified, and each decision must use its own current book.

The clocks the trader must keep separate

At the first daily checkpoint, the agent reads the current packet and decides whether to act, hold, escalate or decline. The day-three and day-five approval examples concern dated responses, whose authority still requires a separate accepted execution. These observation days are distinct from the contract anniversaries below; the environment-design post explains the packet and response mechanism.

At year five, the original loan repays if it performs. The base instruction parks that principal at the overnight benchmark. A new deployment needs a new quote and permission; the original investment spread cannot be extended by assumption. Any purchased five-year protection expires on its own execution-based clock, rather than automatically on the book's fifth anniversary.

At year twenty, U's distribution, the financing obligation and the separate legacy borrowing follow their contractual settlement paths. Unpaid claims and collateral returns require resolution even when scheduled trading exposure ends. The engineering observation window contains twenty daily checkpoints, so it samples these lifecycle regions without making a model live through twenty years.

Recognition and economic coverage

The selected capital rule recognises only eligible, enforceable protection. For an otherwise eligible amount $P$, the maturity adjustment is

$$P_a=P\frac{t-0.25}{T-0.25},$$

Exposure maturity $T$ is capped at five years; recognised protection maturity $t$ is subject to the rule's conditions. With twenty-year exposure and five-year protection, $T=t=5$ gives an initial factor of one. The actual protection still ends after five years. Initial recognition cannot cover the remaining fifteen.³

That distinction affects pricing and the next decision. Replacing collateral can change the binding capital constraint, but a reduction in required capital supplies no cash. Purchased protection consumes premium and brings its own provider exposure. A legal approval may enable a future trade while leaving today's payment unfunded.

We want the recommendation to identify the binding constraint and the facts that would change it. For example, if approval will arrive after the client's deadline, further precision in the funding mark does not restore the route. The desk must seek another executable structure or decline while meeting the obligations already accepted.

How the book is scored explains the checks applied to those decisions; the limitations page records their dated evidence status.

Notes and sources

1. Economic specification, funding quotes, contractual returns and allocation conventions. Figures are synthetic illustrations at unchanged tenor. ↩

2. IHS Markit, CDS Indices Primer, November 2021, pp. 5, 8–9, 17, 19–21 and Appendix 1, quotation conventions and risky premium annuities. ↩

3. Capital rulebook, collateral eligibility and maturity recognition under Basel CRE22; valuation and capital annex, selected applicability conditions. ↩