Cost of capital — private equity & credit

Your hurdle rate only looks conservative.

An 8% threshold sits comfortably above a 7.27% blended funding cost. Seventy-three basis points of headroom. Then you count the capital sitting idle, the facilities drawn early, and the fees accruing on both — and the headroom is gone.

Built by a treasury and fund finance practitioner with sixteen years across Oaktree, Citco and KPMG — not a consultancy pyramid.

Marginal ICC — worked example EUR 500mm fund
0 ↑ static hurdle 800
Funding cost
Costs the hurdle never sees
LP capital — 60% at 8.00%480 bps
Subscription line — 30% at 5.20%156 bps
NAV facility — 10% at 9.10%91 bps
Weighted funding cost727 bps

What the hurdle rate misses

Net idle capital drag63 bps
Management fee drag18 bps
Drawn capital opportunity cost59 bps
Invisible cost — never invoiced140 bps
Marginal ICC 868 bps
Static hurdle rate800 bps
Capital underpriced by68 bps

Illustrative, on the five-component v1.2 model. Every input — source mix, days idle, drawdown timing, fee basis — is yours to set. See the full calculation, or start with the plain-English glossary if the terms are unfamiliar.

The invisible 140

Three costs that never reach the investment committee.

We call them invisible because nobody invoices for them. A bank bills you for interest and it lands in the accounts; nobody bills you for the profit you failed to make while capital sat waiting. Both are real money. Only one gets counted — and the uncounted ones arise from timing rather than pricing, which is why a hurdle rate set once at fund formation cannot possibly capture them.

727 bps

Weighted funding cost

LP preferred return, subscription line and NAV facility, blended by their real share of the next drawdown. Conventional, well understood, and comfortably below a typical 800bps hurdle.

63 bps

Net idle capital drag

Preferred return compounds from the call date, not the deployment date. Ninety days of idle capital costs EUR 4.2m gross — EUR 2.5m after deposit interest, and only if the LPA permits placement at all.

18 bps

Management fee drag

On an AuM basis, fees accrue on capital that has been called but is doing nothing. On a commitment basis they do not. The model switches between the two rather than assuming one.

59 bps

Drawn capital opportunity cost

Facilities drawn ahead of deployment cost you the interest on the borrowing and the return that capital was not earning. Treasury systems capture the first because someone invoices for it. Nobody invoices for the second.

The discipline argument

What gets measured is the smaller number.

EUR 60m drawn seventy-five days ahead of deployment. The interest on that borrowing is invoiced, booked and visible — EUR 641,096. The return that capital was not earning while it sat there is neither invoiced nor booked, and it is nearly three times larger. Both are real. Only one shows up.

What treasury books
EUR 641,096
Facility interest at 5.20% for 75 days
What it actually costs
EUR 2,367,123
Facility interest plus forgone deployment return at 14.00%
Understated by 3.7×

Where ICC sits

WACC values the portfolio. ICC prices the decision.

ICC does not replace your valuation work. It sits earlier — a gate before IC approval, and an audit trail afterwards.

Anchor
WACC and CAPM derive from public market data and equity beta.
ICC derives from your LPA, facility pricing and drawdown log.
Frequency
Refreshed annually, sometimes quarterly.
Recalculated at the point of approval.
Basis
Average cost across the whole capital base.
Marginal cost of the next euro deployed.
Output
A framework requiring translation before use.
One line in the IC memo, with the workings visible.

Delivery

Six weeks, start to IC-ready.

The sequence is fixed because each stage depends on the one before it. Discovery sets the inputs; the build encodes them; the pilot proves them against live deals.

Weeks 1–2

Discovery

Capital structure mapping across LPA, credit agreements and hedging policy. Deployment pattern and idle capital audit.

Weeks 3–4

Build

ClarityStack™ calibrated to your fund. Capital source decision matrix integrated into the treasury playbook. IC memo template with an ICC field.

Weeks 5–6

Pilot

Live test across three to five deals. Variance analysis against your existing hurdle and against realised cost.

Ongoing

Institutionalise

Quarterly reporting, annual policy refresh, and LP-facing materials where capital allocation rigour is worth demonstrating.

Engagement

Priced against the drag it removes.

On the worked example above, the invisible 140 basis points represent roughly EUR 5.6m of cost across a EUR 400m deployment window. The pilot costs a fraction of one basis point of that.

Pilot

EUR 15–25k
4–6 weeks · capped scope

Calibrated ICC model, IC memo template, three-deal pilot against defined success criteria.

Full integration

EUR 35–50k
Includes annual refresh

Everything in the pilot, plus treasury playbook integration, scenario modelling and LP reporting materials.

Retained support

EUR 5–10k
Per quarter · optional

Policy updates, scenario work on live deals, and recalibration as facility terms change.

Fit

This works for some funds and not others.

Worth establishing early. A short call is usually enough to tell which side you are on.

Strong fit

  • EUR 1–10bn European private equity or credit, with a real treasury function
  • A CFO or COO willing to sponsor the work rather than delegate it downward
  • An investment committee that already applies quantitative gates
  • Active use of subscription lines, NAV facilities or multi-currency deployment
  • Preference for operational rigour over presentation

Poor fit

  • Treasury fully outsourced to the fund administrator
  • Investment committee process that is purely qualitative
  • Entirely evergreen LP base with no preferred return or call mechanics
  • Mid-fundraise and looking for optics rather than substance
  • Expecting a deck rather than a working model

Next step

Bring one live deal.

Thirty minutes, your numbers, no preparation needed. We calculate the marginal ICC on a deal you are actually looking at and compare it to the threshold you would otherwise have used. If the gap is not material, that is a useful answer too.

Book a diagnostic
CoverageUK & Europe
ResponseWithin one working day

ClarityStack™ v1.2 — five-component methodology

The calculation, shown in full.

Nothing here is a black box. Every input has a stated source, every component has visible logic, and the two judgement variables are labelled as judgement variables. This is the working model, run on an illustrative EUR 500mm fund.

868
Marginal ICC (bps)
800
Static hurdle (bps)
+68
Hurdle understates by
+552
Deal spread over ICC

Start here if the jargon is unfamiliar.

This page is written for finance specialists, but the underlying idea is simple and worth understanding without the vocabulary. Here is the whole thing in ordinary language, followed by definitions of every term used elsewhere on the site.

The idea in one paragraph. A private equity fund raises money from investors and borrows from banks. That money is not free — investors are promised a return and banks charge interest. Before the fund invests in a company, it should check that the expected profit beats what the money is costing. Most funds check against a fixed figure chosen years earlier, usually 8%. That figure ignores several real costs, so it flatters every decision made against it. This model calculates the true, current cost instead.
Where the money comes from The three funding sources, and what each costs
TermIn this exampleWhat it means
LP capital60% at 8.00%Money from the fund's investors — pension schemes, insurers, wealthy families. "LP" means limited partner. They are promised a minimum return, called the preferred return, before the fund manager earns anything. Here that promise is 8% a year. It is the most expensive money in the mix.
Subscription line30% at 5.20%A bank loan secured against investors' unfulfilled promises to pay. Cheap, short-term, and used to move quickly on a deal before asking investors for cash. Sometimes shortened to "sub line".
NAV facility10% at 9.10%A bank loan secured against the companies the fund already owns. "NAV" is net asset value — what those holdings are worth. More expensive than a subscription line because the bank is taking more risk.
The vocabulary Every term used on this site
TermWhat it means
Measurement
Basis point (bps)One hundredth of one percent. 100 bps = 1%. Used because in finance the difference between 8.00% and 8.68% matters enormously, and "68 basis points" is clearer than "nought point six eight percent".
AnnualisedA cost expressed as if it ran for a full year, so figures covering different periods can be compared. A 90-day cost is annualised by scaling it: 90 ÷ 365.
Marginal costThe cost of the next unit, rather than the average across everything so far. It matters because decisions are always about what to do next, and the next euro can cost quite a lot more than the average euro already spent.
The decision
Hurdle rateThe minimum return an investment must promise before the fund will consider it. Traditionally a fixed number set when the fund launched — typically 6% to 8% — and left unchanged for years.
ICCInternal Cost of Capital. What this model produces: the true, current, all-in cost of deploying the next euro. It replaces the fixed hurdle rate with a figure that reflects what is actually happening in the fund today.
IRRInternal rate of return — the annualised profit a deal is expected to make, as a percentage. If a deal's IRR is 14.2% and the ICC is 8.68%, the deal earns 5.52 percentage points more than the money costs.
SpreadThe gap between two rates. Here, the deal's expected return minus the cost of the capital funding it. A positive spread means the deal creates value; a negative one means it destroys value however profitable it looks in isolation.
Investment committee (IC)The group inside the fund that approves or rejects proposed investments. The ICC figure is designed to go into the paperwork they review.
The costs the hurdle rate misses
DragA cost that reduces returns without anyone deciding to incur it. It arises from timing — money sitting still, borrowed too early, or charged fees while doing nothing — rather than from the price of anything.
Invisible costOur term for drag that never appears on an invoice or in an accounting system, and so is rarely counted. Nobody bills a fund for the profit it failed to make while capital sat waiting, but that failure is as real as any interest payment. In the worked example these costs total 140 basis points — about a sixth of the fund's true cost of capital.
Called capitalMoney the fund has formally requested from its investors and now holds. From the moment it is called, the 8% preferred return starts accruing — whether or not it has been invested in anything.
Idle capitalCalled capital that has arrived but not yet been invested. It is costing 8% a year and earning perhaps 3.4% in a deposit account. The difference is pure loss, and it is the largest of the invisible costs.
DeploymentActually investing the money in a company. "Undeployed" means the cash is sitting in an account rather than working.
Commitment feeA small charge banks levy on the unused portion of a loan facility — payment for keeping the money available. Here 0.50% a year.
Facility interestThe interest paid to a bank on borrowed money during the period before it is put to work. Real, invoiced, and usually the only part of the waiting cost anyone records.
Carried interestConfusingly, "carry" in private equity means something else entirely: the fund manager's share of the profits, typically 20% of gains above the preferred return. It is not a cost of capital and plays no part in this model. Mentioned here only so the term is not mistaken for facility interest.
Opportunity costThe value of the thing you did not do. If capital sat in an account for 75 days instead of earning 14% in an investment, that forgone 14% is an opportunity cost. It never appears in the accounts, which is precisely why it gets ignored.
Fund mechanics
LPALimited partnership agreement — the contract between the fund manager and its investors. It sets the preferred return, the fees, and what the manager may do with cash that is waiting to be invested.
Management feeAn annual charge, here 1.50%, paid to the fund manager for running the fund. Charged either on the total investors committed, or on the value of what the fund currently holds.
Commitments vs AuM basisTwo ways of calculating that fee. On a commitments basis it is charged on the full amount investors promised, so it does not change with deployment. On an AuM basis (assets under management) it is charged on current holdings, which means the fund pays fees on capital sitting idle. Only the second creates drag.
GPGeneral partner — the firm managing the fund and making the investment decisions. The client for this work.
Hedging / FXProtecting against currency movements when investing abroad. A euro fund buying a US business can either borrow dollars directly or borrow euros and use a currency contract. The two routes cost different amounts, and the model picks the cheaper.
Why any of this matters. In the worked example the fund believes its capital costs 8%. It actually costs 8.68%. That gap is small enough to sound trivial and large enough to matter: on a EUR 400m deployment programme, 68 basis points is roughly EUR 2.7m of cost the fund had not counted. Deals approved in that gap looked profitable and were not.

What the model does, and what it will not do.

ClarityStack calculates the marginal cost of deploying the next unit of capital, for use as a pre-approval gate at investment committee. It replaces a static 6–8% hurdle rate with a live, audit-trailed threshold.

Marginal ICC (%) =  Weighted Cost of Next Unit Deployed
                + Net Idle Capital Drag — v1.2: nets off idle cash yield
                + FX Sourcing Cost (optimised) — v1.2: lower of direct draw vs hedged
                + Management Fee Drag Adjustment
                + Drawn Capital Opportunity Cost — v1.2: new component
Two gating conditions Confirm before switching either toggle on
Neither should be assumed. Both change the answer materially.
ConditionWhat must be confirmedWhere
Multi-currency draw The credit agreement supports drawing in the target currency, at transparent per-currency pricing. Credit agreement
Idle cash placement The LPA permits called-but-undeployed capital to be held in interest-bearing accounts. Limited partnership agreement
Why the toggles default to off. Both conditions are commonly assumed to be true and frequently are not. A fund that cannot place idle cash under its LPA has a materially higher net idle drag than one that can — in this worked example the difference is 168 basis points of gross drag. Assuming the permission exists produces a flattering number that does not survive diligence.
Hardcoded input — editable
Formula — do not overwrite
Cross-tab link
Judgement variable — confirm before relying on output

Every input carries its source.

Twenty-three inputs, each labelled with the document it comes from. Where a figure is a judgement rather than an observable rate, it says so. The worked example below is an illustrative EUR 500mm fund.

Inputs Edit blue cells only
InputValueUnitBasis / source
1  Fund parameters
Fund size (commitments)500,000,000EURLPA / final close documentation
Fund base currencyEURLPA
2  Capital source mix
LP capital — mix %60.0%%Deployment plan / treasury forecast
LP capital — cost (pref rate)8.00%%LPA preferred return
Sub line — mix %30.0%%Deployment plan / treasury forecast
Sub line — all-in cost5.20%%Credit agreement: SOFR/EURIBOR + margin
NAV facility — mix %10.0%%Deployment plan / treasury forecast
NAV facility — all-in cost9.10%%NAV facility agreement
Mix check100.0%%If this is not 100%, weighted cost is under- or overstated
3  Idle capital
Called capital sitting idle200,000,000EURTreasury cash position report
Days idle90daysCall date to expected deployment date
LP preferred return rate8.00%%LPA preferred return
Commitment fees on undrawn facilities0.50%%Credit agreement commitment fee
Overnight / term deposit yield3.40%%Treasury sweep rate at measurement date
Toggle: LPA permits idle cash placement?YESY/NMust be confirmed against the LPA. Set NO if unconfirmed.
4  Drawn but undeployed capital  — new in v1.2
Amount drawn, not yet deployed60,000,000EURFacility drawdown log vs deployment log
Days drawn and undeployed75daysDrawdown date to actual deployment date
Facility interest rate on drawn amount5.20%%Credit agreement: all-in rate on the facility actually drawn
Forgone deployment return14.00%%Judgement input: fund target IRR or realised portfolio IRR
5  FX sourcing
Foreign-currency deployment amount40,000,000EUR eq.Deal funding requirement in non-base currency
Proportion of deployment requiring FX0.0%%Set to 0% if the deal is base-currency only
Toggle: multi-currency facility available?YESY/NMust be confirmed in the credit agreement. Set NO if unconfirmed.
Path A — native-currency tranche rate5.65%%Credit agreement: target-currency tranche pricing
Path B — base-currency tranche rate5.20%%Credit agreement: base-currency tranche pricing
Path B — forward points (annualised)0.68%%FX forward quote; embeds cross-currency basis
6  Management fee
Toggle: fee basisAUMAUM/COMMITAUM = fees on invested and idle capital. COMMIT = on commitments.
Management fee rate1.50%%LPA management fee provision
7  Drag allocation basis
Total expected deployment over drag period400,000,000EURCapital expected across all deals in the measurement window
8  Deal assessment
Deployment amount for this deal75,000,000EURProposed investment size
Expected deal IRR (gross)14.20%%Deal team underwriting model
Required spread over ICC to proceed1.00%%House policy: minimum ICC clearance margin
The allocation basis is the input that moves the answer most. Fund-level drags — idle capital, fee drag, interest on drawn-but-undeployed capital — are spread across total expected deployment of EUR 400mm, not loaded onto the single deal being assessed. Loading them onto one deal would overstate its cost several times over and produce a threshold no deal could clear. Set this figure carefully, and set it consistently across deals.

Five components, built in the open.

Each component states its own logic. The two memo lines exist to show what the calculation would have missed under a simpler method — the gap between them is the argument for doing this properly.

1  Weighted cost of next unit deployed727 bps
2  Net idle capital drag63 bps
3  FX sourcing cost (optimised)0 bps
4  Management fee drag18 bps
5  Drawn capital opportunity cost59 bps
Marginal ICC 868 bps
ICC Calculation All figures derive from the Inputs tab
ComponentValueUnitLogic
Component 1  — Weighted cost of next unit deployed
LP capital contribution4.80%480 bps60.0% of funding × 8.00% preferred return = 4.80%. The largest single cost, because LP capital is both the biggest slice and the most expensive.
Subscription line contribution1.56%156 bps30.0% of funding × 5.20% all-in rate = 1.56%. Cheaper than LP capital, which is why sub lines are drawn first.
NAV facility contribution0.91%91 bps10.0% of funding × 9.10% all-in rate = 0.91%. The most expensive source per euro, but only a tenth of the mix.
Component 17.27%727 bps4.80% + 1.56% + 0.91%. An annual rate: the blended cost of holding this funding mix for a year.
Component 2  — Net idle capital drag  · measured over 90 days
Gross idle drag4,191,781EUREUR 200m idle × (90 ÷ 365) × (8.00% preferred return + 0.50% commitment fee) = EUR 4,191,781. The preferred return accrues to LPs from the day capital is called, not the day it is invested — so 90 days of waiting costs real money.
Less: interest earned on placement(1,676,712)EUREUR 200m × (90 ÷ 365) × 3.40% deposit rate = EUR 1,676,712. Earned only where the LPA permits idle cash to be held in interest-bearing accounts. If it does not, this line is zero and the drag more than doubles.
Net idle drag2,515,068EUREUR 4,191,781 − EUR 1,676,712. What 90 days of idleness actually costs the fund.
Component 20.63%63 bpsEUR 2,515,068 ÷ EUR 400m total expected deployment = 0.63%. A one-off cost per euro deployed, not an annual rate.
Component 3  — FX sourcing cost (optimised)
Path A — direct native-currency draw5.65%%Native-currency tranche rate
Path B — base-currency draw + hedge5.88%%Base tranche rate + forward points
Selected path cost5.65%%Lower path, only where multi-currency draw is confirmed
Memo: saving vs hedged path0.23%23 bpsWhat the optimisation is worth when FX is in play
Component 30.00%0 bpsSelected path × 0% FX proportion — nil on this deal
Component 4  — Management fee drag  · measured over the same 90 days
Fee drag on idle capital739,726EUREUR 200m idle × (90 ÷ 365) × 1.50% management fee = EUR 739,726. On an AuM basis the fee is charged on capital that has been called but is not yet working, so the fund pays to manage money that is doing nothing. On a commitment basis the fee does not vary with deployment and this line is zero.
Component 40.18%18 bpsEUR 739,726 ÷ EUR 400m total expected deployment = 0.18%. A one-off cost per euro deployed.
Component 5  — Drawn capital opportunity cost  · measured over 75 days  — new in v1.2
Combined annualised rate19.20%1,920 bps5.20% facility interest rate + 14.00% forgone deployment return. You pay interest on the borrowing and lose the return the money would have made if invested.
Drawn capital opportunity cost2,367,123EUREUR 60m drawn × (75 ÷ 365) × 19.20% = EUR 2,367,123. The 75 days run from the drawdown date to the date the money was actually deployed.
Memo: facility interest alone641,096EUREUR 60m × (75 ÷ 365) × 5.20% = EUR 641,096. This is the only part the bank invoices for, and the only part most treasury systems record.
Component 50.59%59 bpsEUR 2,367,123 ÷ EUR 400m total expected deployment = 0.59%. A one-off cost per euro deployed.
Marginal ICC  — total
Marginal ICC8.68%868 bpsSum of components 1–5
Typical static hurdle (reference only)8.00%800 bpsThe proxy this framework replaces — shown for contrast, not used in any calculation
ICC vs static hurdle+0.68%+68 bpsPositive = the static hurdle understates true cost of capital
Two different measurement bases, and why. Component 1 is an annual rate — the cost of holding the funding mix for a year. Components 2, 4 and 5 are one-off costs — a specific euro amount incurred over a specific number of days, then divided by the capital expected to be deployed in that window. They are added together because the question being asked is a single one: what does it cost, all in, to get the next euro deployed and hold it for a year. If your measurement window is much shorter or longer than a year, say so explicitly when presenting the number, because the two bases only reconcile cleanly when the window is roughly annual.
The memo line on component 5 is the point of the whole exercise. Facility interest alone on drawn-but-undeployed capital is EUR 641,096. Add the return that capital was not earning while it sat there and the true cost is EUR 2,367,123 — nearly four times larger. Treasury systems capture the first number because someone invoices for it. Nobody invoices for the second.

One block, straight into the IC memo.

The output tab is written to be pasted. It gives the committee a recommendation, the attribution behind it, and an audit trail of every condition that was assumed — so the assumptions can be challenged in the room rather than discovered later.

IC Output — decision block Every figure traceable to Inputs
MeasureValueNote
Deployment amountProposed investment size
Marginal ICCLive cost of deploying the next unit of capital
Expected deal IRR (gross)Deal team underwriting model
Spread over ICC5.52%Deal IRR less marginal ICC
Spread over ICC (bps)552Same figure in basis points
Required spread (house policy)Minimum ICC clearance margin
RecommendationPROCEEDProceed only where spread over ICC meets or exceeds house policy minimum
Component attribution Where the cost is coming from
ComponentContributionShare of total ICC
1  Weighted cost of next unit deployed83.8%
2  Net idle capital drag7.2%
3  FX sourcing cost (optimised)0.0%
4  Management fee drag2.1%
5  Drawn capital opportunity cost6.8%
Total8.68%100.0%
Audit trail — conditions confirmed Challenge these in the room, not afterwards
ConditionSettingNote
Drag allocation baseFund-level drags spread across this, not this deal alone
Multi-currency facility draw available?YESMust be confirmed in the credit agreement
LPA permits idle cash placement?YESMust be confirmed in the LPA
Management fee basisAUMAUM-basis funds incur fee drag on idle capital
Forgone deployment return14.00%Not a market-observable rate. Defaults to fund target IRR; confirm the house basis.
The audit trail is not decoration. Four of the five conditions above change the answer if they are wrong, and two of them are permissions that must be read out of legal documents rather than assumed. Publishing them alongside the recommendation means the committee is approving a number it can see the shape of.

Next step

Run it on a live deal.

Thirty minutes, your numbers. We set the inputs from your LPA and credit agreement, calculate the marginal ICC on a deal you are actually looking at, and compare it to the threshold you would otherwise have used.

Book a diagnostic
VersionClarityStack™ v1.2
ComponentsFive
Inputs23, each sourced
Worked exampleEUR 500mm fund
FormatExcel, no add-ins