The Cash Sweep: What a Dollar of Prepayment Is Actually Worth to Equity
A cash sweep takes a dollar that would have been distributed to equity and uses it to prepay debt. Equity is not deprived of the dollar — it gets the benefit back, as saved interest and as principal it no longer has to repay later.
But it gets it back later, and equity discounts later cash at a higher rate than debt costs. With a 12% equity discount rate, a 6.5% debt rate and ten years remaining, a dollar swept returns about 69 cents of present value. Thirty-one cents disappears into the spread.
On the worked project below, a 50% sweep costs equity $11.9m — 10.3% of the present value of its entire distribution stream — to retire the loan six years early.
That is the whole negotiation, and it is rarely expressed that way. This post covers the mechanics, the definitions that decide how much gets swept, what the percentage is actually worth, and the structures that make a sweep tolerable.
ℹ️ Note: All figures are labelled assumptions. The per-dollar result depends on the spread between the equity discount rate and the debt rate, and on the remaining tenor.
What Is a Cash Sweep?
A cash sweep is a contractual application of defined excess cash or other specified proceeds to debt prepayment. It can be continuous, triggered by coverage or leverage, linked to a mini-perm or PPA tail, or applied to specific receipts. The executed definition controls both the sweep base and the percentage.
In the worked model it sits after scheduled debt service and reserve funding, so it reallocates residual cash without changing the contractual DSCR numerator. Other financings can define excess-cash-flow triggers or cure mechanics differently, so do not assume every sweep is economically “below CFADS.”
The Mechanics
Three parameters do all the work.
The excess cash flow definition. The amount available to sweep is whatever the credit agreement defines after the specified deductions, reserves and adjustments. A sponsor should model that definition directly; a headline sweep percentage is meaningless without the base it applies to.
The percentage. It is whatever the credit agreement specifies. Filed financings include 100% excess-cash trapping during a DSCR-triggered sweep period, regular cash sweeps sized to repay debt by the end of a PPA term, and elective prepayments from dedicated excess-cash accounts.
The release / step-down mechanics. Some sweeps end only after specified ratios are satisfied for consecutive test dates; others continue until a target balance or maturity condition is met. Model the actual release test rather than a generic 100%/75%/50% ratchet.
Why Do Lenders Want One?
Because it converts good years into permanent risk reduction.
The clearest statement of the rationale is the volatile-revenue case: where a borrower's cash flow is uncertain — a merchant power plant, for instance — applying the project's excess cash in good years to reduce the debt acts as a buffer against years when revenues may be lower.
That is a genuinely sound argument, but the mechanism is deal-specific. A public Talasol financing, for example, required regular cash sweeps so the term loans would be repaid by the PPA termination date, with an option to place sweep amounts into a reserve that could later be released if a satisfactory new hedge or PPA was entered into.
The second reason is refinancing risk. Where a facility has a bullet, a sweep reduces the amount that must be refinanced, and does so faster if conditions are good — which is precisely when the refinancing would have been easiest anyway. The sweep is therefore somewhat counter-cyclical in its effect, which is unusual and valuable.
What Does It Cost Equity?
The spread, applied to every dollar swept for the remaining tenor.
The per-dollar arithmetic is worth doing once, because it generalises. A dollar prepaid saves interest at the debt rate for the remaining years and avoids a principal repayment at the end. Discounted at the equity rate:
| Years remaining | $1 swept returns | Destroyed |
|---|---|---|
| 5 | $0.802 | 19.8¢ |
| 10 | $0.689 | 31.1¢ |
| 15 | $0.625 | 37.5¢ |
The longer the remaining tenor, the more a sweep costs — because the cash comes back further away. That is the opposite of most intuitions, which assume an early sweep is cheap because the loan is retired sooner.
Now the whole-project version. Assumptions, labelled as such:
Debt at start of operations $150,000,000
Debt rate 6.5%
Scheduled tenor 15 years
Level debt service $15,952,917
CFADS $30,000,000
Excess cash before sweep $14,047,083
Asset life 20 years
Equity discount rate 12%
| Sweep | Equity NPV | vs no sweep | Debt repaid |
|---|---|---|---|
| 0% | $115,430,150 | — | year 15 |
| 25% | $108,324,672 | −6.2% | year 12 |
| 50% | $103,523,661 | −10.3% | year 9 |
| 75% | $100,039,316 | −13.3% | year 8 |
| 100% | $97,388,329 | −15.6% | year 7 |
Two observations that should change how this is negotiated.
The cost is front-loaded in the percentage. Going from 0% to 25% costs 6.2 points of equity value. Going from 75% to 100% costs 2.3. The first quarter of the sweep is nearly three times as expensive as the last — so a sponsor conceding "only 25%" has given away a disproportionate share of what the whole clause is worth.
The benefit is back-loaded in the tenor. A 50% sweep shortens the loan by six years, at a cost of about $2.0m of equity NPV per year of early repayment. Whether that is worth paying depends entirely on what the sponsor thinks those six years of leverage are worth — which is a view about refinancing conditions in year nine, not about the sweep itself.
There is a third observation worth adding because it cuts against the usual sponsor instinct. The marginal cost of sweep percentage falls as the percentage rises, which means a sponsor negotiating hard from 50% down to 25% recovers 4.2 points of equity value, while one negotiating from 100% down to 75% recovers only 2.3. The clause is worth fighting over at the bottom of the range, not the top.
The reason is mechanical: a high sweep retires the loan quickly, so there are fewer years over which the sweep operates at all. A 100% sweep repays the facility in year seven and then stops; a 25% sweep runs for twelve years. The total swept is far more similar than the percentages suggest, and most of the difference in equity value comes from the first few years when the balance — and therefore the remaining tenor over which each swept dollar is discounted — is largest.
What Makes a Sweep Tolerable?
Four structures, roughly in order of how much they are worth.
A step-down that actually triggers. A sweep ratcheting from 100% to zero as leverage falls is only valuable if the tests are reachable on the base case. A sponsor should model the step-downs explicitly and confirm the year each one is projected to bite. A ratchet whose first step-down never triggers on the sponsor's own numbers is a 100% sweep with decoration.
A cap on cumulative sweeps. If the documents permit one, a cumulative cap bounds how much residual cash is diverted to prepayment. Its availability is a negotiation point, not a market assumption.
A retained portion that is genuinely distributable. The unswept share should not be subject to a separate distribution test that blocks it anyway. Sponsors occasionally negotiate a 50% sweep and discover the retained 50% is trapped by the same lock-up conditions, which makes the percentage irrelevant.
Excess cash flow definition with real deductions. Since the sweep operates on a defined amount, deductions for committed capital expenditure, permitted acquisitions and reserve top-ups reduce the base. This is where the arithmetic is quietly won or lost, and it attracts far less attention than the headline percentage.
What Else Triggers a Sweep?
The recurring excess-cash sweep is only one mandatory-prepayment mechanism. Asset sales, insurance proceeds, tax-credit proceeds, refinancings or other receipts may have separate treatment under the agreement.
Asset sale proceeds. The agreement may require prepayment, permit reinvestment for a defined period, or apply another threshold / exception. Extract the actual proceeds mechanics.
Insurance and condemnation proceeds above the reinstatement threshold, as the waterfall post described.
Tax credit transfer proceeds. This is the one worth flagging, because it is new enough that many precedents do not address it. As the hybrid structures post set out, a §6418 transfer can produce sixty or seventy million dollars of cash arriving as a single receipt in a structure built to distribute periodic operating cash flow. If the credit agreement's mandatory prepayment provisions capture it — as a receipt outside the ordinary course, or as excess cash flow — the entire amount sweeps to the debt. If they do not, it falls to the residual and is distributable.
Neither outcome is wrong, and both are defensible. What is unacceptable is not knowing which applies, and the sum involved can exceed a year's entire distribution. A sponsor negotiating a facility on a project expecting to sell credits should name those proceeds explicitly in both the waterfall and the mandatory prepayment provisions, and decide their treatment deliberately.
Change of control. This may trigger mandatory prepayment, an Event of Default, lender consent or another consequence. Check the definition against any planned sell-down or ownership change.
The general instruction is that the recurring percentage sweep is the clause everyone negotiates and the event-driven triggers are where the large numbers actually arise.
Sweep and Lock-Up: Where They Overlap
Worth separating clearly, because the two mechanisms can be made economically identical by drafting.
A lock-up traps cash. A sweep applies it to the debt. They are different — until the lock-up provides that trapped cash is swept rather than released, at which point the lock-up is a sweep, triggered by a coverage test rather than running continuously.
the previous post quantified that difference: on the worked figures a one-year lock-up event cost equity $2.5m if the cash was released and $7.4m if half of it was swept. That is the same distinction appearing from the other direction.
The practical consequence is that a sponsor assessing total sweep exposure must read both clauses together. A facility with a modest 25% ongoing sweep and a lock-up that sweeps 100% of trapped cash is, in a bad year, a full sweep facility. One with a 50% ongoing sweep and full release on lock-up exit is considerably gentler in exactly the circumstances that matter.
The interaction also matters for modelling. A model that implements the sweep in the waterfall and the lock-up in the covenant tests, without connecting them, will double-count or omit depending on which one it applies first — and both errors are silent.
Sweep or Sculpt?
A question worth asking directly, because the two are alternatives and are rarely compared.
A sweep and a faster amortisation schedule both deleverage. The difference is that the schedule is certain and pre-agreed, while the sweep is contingent on performance.
For a lender, the sweep is better: it gets deleveraging when cash is available and does not create a default risk when it is not. For a sponsor, a faster fixed schedule is often better: it is predictable, it can be modelled and financed around, and it does not subordinate the entire distribution stream to an ongoing test.
The comparison to run is a sweep at the offered percentage against a shorter tenor with no sweep. Where the sweep's expected deleveraging path and a shorter fixed schedule produce similar balances, the fixed schedule is worth more to equity because the residual is predictable — and predictable distributions support a higher valuation in a sale than contingent ones.
That is a real negotiation and it is available more often than sponsors assume, because the lender's objective is the balance profile rather than the mechanism that produces it.
Why Sweeps and Merchant Tails Travel Together
Because a sweep can reduce the balance exposed to merchant or refinancing risk. Whether it is linked to the merchant tail, PPA expiry, leverage or a DSCR trigger depends on the financing.
The merchant tail analysis established that lenders credit only three to five years of post-PPA revenue, heavily haircut, at a raised coverage ratio — and that on the worked case three merchant years added just 5.9% to the facility. That is a lender declining to lend against a revenue stream it cannot forecast.
A sweep offers a different arrangement: lend against it, and take the cash if it materialises. Where merchant revenue turns out strong, the sweep retires the debt early and the exposure never bites. Where it is weak, the balance stays higher but so does the evidence that the conservative sizing was correct.
That can be an efficient risk-sharing structure where the lender is willing to leave upside with equity until the specified trigger or tail period. Whether the sweep increases, decreases or disappears over time is entirely document-specific.
The sponsor's counter-argument is the one this post has been making. A sweep during the merchant tail operates on exactly the years with the longest remaining tenor at the point cash arrives, which — per the per-dollar table — is when a swept dollar destroys the most value. A 100% sweep confined to the merchant years is therefore not the mild concession it appears to be, and should be priced with the same arithmetic as a continuous one.
The structure worth proposing instead is a sweep tied to realised merchant prices rather than to the period: full sweep above a price threshold, reduced below it. That gives the lender the deleveraging in exactly the scenario it is worried about, and gives equity the distributions in the scenario where it needs them most.
How Do You Model a Sweep in Excel?
As a balance-reducing cash flow with a revolving effect on the schedule, and with the equity cost reported explicitly.
The cascade
CFADS
less Senior debt service (scheduled)
less DSRA and reserve funding
= Excess cash flow
Sweep_Amount = MIN( ECF × Sweep_Pct , Outstanding_Balance )
Distribution = ECF − Sweep_Amount
Closing balance = Opening − Scheduled principal − Sweep_Amount
The MIN against the outstanding balance matters. Without it, a model with a high sweep percentage will drive the balance negative in the year the loan is repaid and generate interest income on a phantom negative debt.
The two outputs that matter
PF_EquityNPV_NoSweep = $115,430,150
PF_EquityNPV_WithSweep = $103,523,661
PF_SweepCost = $11,906,489 (10.3%)
PF_DebtRepaidYear_NoSweep = year 15
PF_DebtRepaidYear_WithSweep = year 9
PF_CostPerYearEarly = $1,984,415
Report the last of those. "A 50% sweep" is a term. "Two million dollars of equity value per year of early repayment" is a decision.
The per-dollar check
Value_Per_Dollar_Swept
= Debt_Rate × Annuity(Equity_Rate, Years_Remaining)
+ (1 + Equity_Rate) ^ −Years_Remaining
At 6.5% debt, 12% equity and ten years remaining that is $0.689. It is a useful sanity check on any sweep model, and a useful thing to be able to state in a negotiation.
The step-down schedule
For each ratchet step: report the projected year it triggers
If a step never triggers on the base case → flag it
ℹ️ Note: Model the sweep against the contractual excess cash flow definition, with its adjustments, not against a simple residual. On a deal with meaningful capex or working capital deductions the two differ substantially, and the difference is the part of the clause that was actually negotiated.
To build the sweep cascade, the equity cost decomposition and the ratchet schedule, prompt Dezzmond with your facility terms and cash flow profile.
What Do Sponsors and Lenders Actually Check?
- What does the sweep cost in equity NPV, expressed as a percentage and per year of early repayment?
- Is the cost front-loaded in the percentage? The first 25% is the expensive quarter.
- Do the step-downs trigger on the base case, and in which years?
- Is the excess cash flow definition doing work, or is it a simple residual?
- Is the retained portion genuinely distributable, or blocked by the same lock-up tests?
- Would a shorter fixed tenor achieve the same balance profile at a lower cost to equity?
- Does the model cap the sweep at the outstanding balance?
Frequently Asked Questions
What is a cash sweep?
A contractual mechanism applying defined excess cash or other specified proceeds to prepay debt. Its effect on distributions, covenant ratios and refinancing risk depends on where it sits in the waterfall and how the finance documents define the sweep base.
What percentage is typical?
There is no universal percentage. Filed agreements show full excess-cash trapping during a sweep period, scheduled sweep amounts tied to PPA expiry, and other structures. Use the executed sweep percentage and release tests.
What does a sweep cost equity?
The spread between the equity discount rate and the debt rate, applied for the remaining tenor. At 12% equity and 6.5% debt with ten years remaining, a dollar swept returns about 69 cents of present value.
Why does the cost rise with remaining tenor?
Because the benefit of prepayment — saved interest and avoided principal — arrives further in the future, and equity discounts it more heavily. A dollar swept with fifteen years remaining destroys 37.5 cents; with five years remaining, 19.8 cents.
Do tax credit transfer proceeds get swept?
It depends entirely on whether the mandatory prepayment provisions capture them, and many precedents predate transferability. A §6418 sale can produce more cash in one receipt than a year of distributions, so the treatment should be named explicitly in both the waterfall and the prepayment clauses.
How does a sweep interact with a lock-up?
They become the same mechanism where the lock-up provides that trapped cash is swept rather than released. Total sweep exposure has to be read across both clauses — a 25% ongoing sweep with a 100% lock-up sweep is a full sweep facility in a bad year.
Is a sweep better or worse than a shorter tenor?
For the lender, a sweep can deleverage opportunistically without increasing scheduled debt service in weak periods. For equity, a shorter fixed schedule may be easier to value and plan around. Compare the two structures on the same projected debt-balance path rather than assuming one is inherently better.
Closing: A Clause Priced in Percentages, Paid in Present Value
The cash sweep is negotiated as a number between zero and one hundred, and that framing obscures almost everything about it.
It is not a proportional term. The first quarter of a sweep costs nearly three times what the last quarter does, so the instinct to concede "just 25%" as a gesture gives away a disproportionate share of the clause's value. It is not a neutral transfer either — equity does get the money back, but at a rate that destroys thirty-one cents in the dollar on a ten-year remaining tenor, and more on a longer one.
And it is not the only way to achieve what the lender wants. A lender's objective is a balance profile; a sweep is one mechanism for producing it, and a shorter fixed amortisation is another that equity generally prefers because it is predictable.
The useful discipline is to convert the percentage into the two numbers that describe the trade: what it costs in equity NPV, and what it buys in years of early repayment. On the worked case that is $11.9m for six years — roughly two million dollars a year — and once those numbers are on the page, the conversation is about whether six years of earlier deleveraging is worth two million a year, which is a question a credit committee and a sponsor can actually answer.
the next post moves to the other end of the facility's life: the construction loan, interest during construction, and the drawdown mechanics that determine how much of the project cost is financed before a single megawatt-hour is sold.
Sources: SEC filing — Talasol project finance cash sweeps sized to repay by PPA termination, with reserve alternative · 2026 SEC-filed credit agreement — DSCR-triggered Excess Cash Flow sweep and release after consecutive test dates · 2026 Project Swift loan-guarantee agreement — Excess Cash Account and Cash Sweep Prepayment · Dezzmond analysis; worked-example sweep percentages and returns are illustrative.