DSCR-Driven vs Gearing-Driven Debt Sizing: Which Constraint Binds, and What It Tells You

Every project finance term sheet contains two debt sizing constraints. A maximum gearing — say 75% debt to 25% equity — and a minimum debt service coverage ratio, typically 1.30× for a contracted renewable project.

Both are stated. Only one binds. And which one binds is not a detail of the arithmetic — it determines what the financing negotiation is actually about, which assumptions matter, and whether improving the project's revenue does anything at all to the amount of debt available.

The crossover between them is a single number: the project's cash yield on cost. On the market terms above, at roughly 9.3% of CFADS to capex, the two constraints give the same answer. Below it, DSCR binds. Above it, gearing binds.

This post opens Series E with the mechanics of both constraints, where the crossover sits, why it moves by market and technology, and what a modeller should do differently depending on which side of it a project falls.

ℹ️ Note: Market terms are indicative and move. Every figure below is a labelled assumption; the method is the point, not the numbers.

What Sizes Project Finance Debt?

The lower of two tests. Debt cannot exceed a stated percentage of project cost, and the cash flow available for debt service must cover debt service by at least a stated multiple in every period. The facility is sized at the minimum of the two answers.

As the standard formulation has it, the amount of debt raised is expressed by a maximum gearing ratio — for example a maximum of 75% debt and 25% equity — and a minimum DSCR, for example no less than 1.40×. Lenders structure so that debt does not exceed a given debt-to-capital ratio and the DSCR does not fall below a given level.

Everything else — sculpting, tenor, reserve accounts, the sweep — operates within whichever of those two the model hits first.

The DSCR Constraint

Coverage sizing works backwards from cash flow. Take the cash available for debt service in each period, divide by the target coverage ratio to get the maximum permissible debt service, and discount that stream at the debt interest rate over the tenor.

Max debt service in period t  =  CFADS(t) ÷ target DSCR
Max debt                      =  PV of that stream at the debt rate, over the tenor

For a flat cash flow this collapses to an annuity:

Max debt = (CFADS ÷ target DSCR) × annuity factor(rate, tenor)

Three inputs move it, and only one is about the project. The target DSCR and the debt rate and tenor are set by the lender and the market. CFADS is the project's.

Market levels for contracted renewables cluster tightly. Projects with long-term offtake agreements typically require 1.20× to 1.35×, with 1.30× the most common single figure and the level most often written into a term sheet to provide headroom against production variability.

What Is Actually in CFADS?

More contested than the formula suggests, and every item is worth a specific answer before the sizing is agreed.

Cash flow available for debt service is, broadly, revenue less operating costs less tax, adjusted for working capital and for movements in reserve accounts. The definition sits in the credit agreement and it is negotiated, because each inclusion or exclusion moves the debt directly through the sizing formula.

The recurring arguments are these.

Tax. A project with accelerated depreciation pays little or no cash tax in early years and a lot later. Sizing on pre-tax cash flow overstates capacity in the tail; sizing on post-tax reflects it. The depreciation post in Series B is directly relevant — a hundred percent bonus deduction in year one produces a cash tax profile that looks nothing like a straight line.

Maintenance capex and major overhauls. A blade replacement programme or a battery augmentation is not an operating cost in any accounting sense and is absolutely a cash outflow the lender must be repaid around. Whether it sits in CFADS or is funded from a reserve is a real choice with real consequences for both the sizing and the reserve account structure.

Reserve account movements. Funding a debt service reserve consumes cash; releasing it produces cash. A definition that nets reserve funding into CFADS reduces early-year sizing; one that treats it as a use below the line does not.

Curtailment compensation and deemed generation. Where a PPA pays for energy not delivered, that payment is revenue — but it is contingent, sometimes disputed, and lenders differ on whether to credit it in the base case.

The general instruction is that CFADS is a defined term, not an accounting fact. A sponsor comparing two term sheets on headline DSCR without comparing the CFADS definitions underneath them is comparing two different numbers that happen to share a name.

The Gearing Constraint

Far simpler, and that simplicity is the point.

Max debt = gearing % × total project cost

No cash flow, no discounting, no forecast. It is a statement about how much equity the lender requires the sponsor to have at risk, expressed as a share of what the project cost to build.

The gearing constraint exists because coverage sizing alone can produce uncomfortable outcomes. A project with strong contracted cash flows and low capital cost can support debt well above its cost — and a lender that funded it would hold an asset worth less than its loan from day one, with a sponsor holding almost no equity and correspondingly little reason to protect it.

Which One Binds?

It depends on a single ratio: cash flow available for debt service, divided by project cost.

Work it through. Assumptions, labelled as such:

Total project cost                          $250,000,000
Maximum gearing                                      75%
Target DSCR                                       1.30×
Debt interest rate                                  6.5%
Debt tenor                                      18 years
Annuity factor, 18 yrs @ 6.5%                    10.4325

Gearing gives a fixed answer:

Max debt = 75% × 250,000,000                = $187,500,000

DSCR gives an answer that moves with CFADS:

CFADS	Max debt service	DSCR-sized debt	Implied gearing	Binding constraint
$22.0m	$16.92m	$176.5m	70.6%	DSCR
$23.4m	$17.97m	$187.5m	75.0%	neither — crossover
$24.0m	$18.46m	$192.6m	77.0%	Gearing
$26.0m	$20.00m	$208.6m	83.5%	Gearing

The crossover is at CFADS of $23.365m, which against a $250m project cost is a cash yield on cost of 9.35%.

That number is worth internalising, because it converts a modelling question into a back-of-envelope one. A project generating less than about 9.3% of its capital cost as annual CFADS is DSCR-constrained. Above that, it is gearing-constrained. Change the target DSCR, the debt rate or the tenor and the threshold moves — but for any given term sheet there is one such number, and it can be computed before the model is built.

Why Does It Change by Market?

Because the crossover depends on the term sheet, and the project's position relative to it depends on local capital costs and revenue.

The term sheet moves the threshold. A longer tenor raises the annuity factor, which raises DSCR-sized debt and makes gearing more likely to bind. A higher interest rate does the opposite. And a higher target DSCR — 1.45× for a merchant project against 1.25× for a strongly contracted one — lowers DSCR-sized debt sharply, pushing merchant projects firmly into DSCR-constrained territory.

The project's economics move its position. A market with high PPA prices relative to installed cost produces high cash yields and gearing-constrained projects. A market with expensive land, expensive interconnection or weak offtake pricing produces the reverse. The Series C posts are directly relevant here: a project carrying six dollars a megawatt-hour of basis and an eight percent curtailment rate has a materially lower CFADS against the same capex, and can cross from gearing-constrained to DSCR-constrained on locational factors alone.

Contracted versus merchant is the largest single driver. It moves the target DSCR, it moves the tenor a lender will offer, and it moves the revenue assumption the CFADS is built on — all three in the same direction.

What Does Each One Binding Tell You?

This is the part that changes behaviour, and it is usually left implicit.

	DSCR-constrained	Gearing-constrained
What limits debt	The project's cash flow	The sponsor's required equity
More revenue →	More debt	No more debt
Lower capex →	Same debt, better returns	Less debt
Negotiate on	Target DSCR, tenor, revenue assumptions	The gearing cap
Sensitive to	Every revenue assumption in the model	Cost estimates only
Typical of	Merchant, weak offtake, high-cost markets	Strongly contracted, low-cost markets

Two consequences deserve stating plainly.

In a gearing-constrained deal, improving the revenue case does not raise the debt. A sponsor spending weeks arguing the basis assumption down or the capture rate up, in a deal where gearing binds, is improving the equity return and not the leverage. That may still be worth doing — but it is a different objective from the one usually being pursued in that conversation.

In a DSCR-constrained deal, reducing capex does not reduce the debt. Value engineering that cuts twenty million from the build cost, in a DSCR-constrained deal, cuts twenty million from the equity cheque and leaves the debt unchanged. That is a very high return on the effort, and it is systematically under-pursued relative to revenue optimisation.

Knowing which regime you are in tells you where to spend the next fortnight.

There is a third consequence that is easy to miss and matters at portfolio level. The two regimes respond differently to a market-wide shock, which means a portfolio is not uniformly exposed. A rise in interest rates lowers the annuity factor and therefore DSCR-sized debt, while leaving gearing-sized debt untouched — so a rate move re-sizes the DSCR-constrained projects in a portfolio and leaves the gearing-constrained ones alone until they cross over. A rise in construction costs does the reverse: it raises gearing-sized debt in absolute terms while making the project worse, and squeezes DSCR-constrained deals through the equity cheque instead.

A development company holding twenty projects should therefore know, for each of them, which side of the crossover it sits on. That single classification tells it which half of the portfolio is exposed to rates and which half is exposed to capex — and those are two different hedging problems with two different answers.

Where Does the Tenor Come From?

From the contract, less a tail — and the tail is a sizing input hiding as a documentation term.

A lender will not let debt mature at the same moment the revenue contract ends. It requires a tail: a period of contracted or credibly forecastable revenue extending beyond final maturity, so that if the repayment schedule slips there is still cash flow to repay from. A 20-year PPA therefore typically supports something closer to a 17- or 18-year facility.

That single decision moves the sizing more than most negotiated points. On the worked term sheet, extending the tenor from 18 years to 20 raises the annuity factor from 10.4325 to about 11.02, lifting DSCR-sized debt by roughly 5.6% — larger than the 4.0% gained by moving the target DSCR from 1.30× to 1.25×. Tenor is usually negotiated as a scheduling question and is in fact a leverage question.

It also interacts with everything Series C established. A facility maturing in year 18 against a PPA ending in year 20 has a two-year tail of contracted revenue. A facility maturing in year 18 against a PPA ending in year 15 has a three-year merchant tail, which is a completely different proposition and is sized quite differently — the subject of a later post in this series.

The practical instruction is to treat tenor, tail and target DSCR as one negotiation rather than three. They trade against each other, they all feed the same formula, and a sponsor optimising one in isolation will usually give back more than it gains on the other two.

Where Do LLCR and PLCR Fit?

Alongside DSCR, and it is worth being clear that they do different jobs.

DSCR is a period measure: in this year, does cash cover debt service? It is what sizes the facility under the coverage constraint and what the covenants test.

The loan life coverage ratio and project life coverage ratio are whole-of-life measures: the present value of cash flow over the remaining loan life, or the remaining project life, divided by outstanding debt. They answer a different question — not "can the project pay this year" but "is there enough value in the remaining asset to cover the balance."

Lenders examine annual or semi-annual DSCR profiles alongside LLCR, minimum cash balances and reserve account requirements, and each catches something the others miss. A project can pass every annual DSCR test and have an LLCR below one, if the cash flow is front-loaded and the balance is not amortising fast enough. It can also pass LLCR comfortably and breach a single-year DSCR because of a one-off maintenance event.

They are covered properly two posts from here. The point for sizing is narrower: the facility is sized on DSCR and gearing, and LLCR is usually a covenant rather than a sizing test — but where an LLCR floor is imposed at financial close it becomes a third constraint, and the MIN in the model needs a third argument.

How Do You Build This in Excel?

As two independent calculations and an explicit MIN, with the binding constraint reported as an output.

The core

Gearing_Debt   = Gearing_Pct × Total_Project_Cost

DSCR_Debt      = NPV(Debt_Rate, CFADS_range / Target_DSCR)

Sized_Debt     = MIN(Gearing_Debt, DSCR_Debt)

Binding        = IF(Gearing_Debt < DSCR_Debt, "GEARING", "DSCR")

That last line is the one almost no model publishes and every model should. A number that says only "$187.5m" hides the fact that it came from a cost estimate rather than a cash flow forecast, and a reader cannot tell which sensitivities matter without it.

The crossover, computed once

Crossover_CFADS = Gearing_Pct × Cost × Target_DSCR ÷ Annuity_Factor
                = 75% × 250,000,000 × 1.30 ÷ 10.4325
                = $23,365,000

Crossover_Yield = 23,365,000 ÷ 250,000,000               = 9.35%

The headroom to the crossover

Project CFADS                                   $24,000,000
Crossover CFADS                                 $23,365,000
Headroom                                           $635,000    (2.6%)

A 2.6% fall in CFADS flips this deal from gearing-constrained to DSCR-constrained. That is well inside the range of a single adverse assumption — the basis moving a dollar a megawatt-hour, the curtailment rate moving two points, an opex escalation. A deal sitting that close to the crossover should be modelled on both sides of it, because the sensitivity profile changes completely when it crosses.

For sculpted debt

The flat-annuity shortcut above is a sizing approximation. Real facilities sculpt principal so that DSCR is constant, which requires discounting the actual CFADS profile rather than an annuity:

DSCR_Debt = SUMPRODUCT(CFADS_range / Target_DSCR, Discount_Factors)

This is exact where the sculpt achieves the target in every period. Where it does not — because of a debt service floor, a minimum amortisation, or a tail constraint — the sculpt itself becomes circular, which is the subject of the next post.

ℹ️ Note: Size on the lender's CFADS, not the sponsor's. The two differ by the haircuts Series C described — basis, curtailment, accreditation — and it is normal for the same project to be gearing-constrained in the sponsor's model and DSCR-constrained in the bank case. That disagreement is worth surfacing early, because it is really a disagreement about revenue wearing the costume of a sizing dispute.

To build both constraints, the crossover and the headroom test, prompt Dezzmond with your capex, CFADS profile and term sheet.

What Do Sponsors and Lenders Actually Check?

• Which constraint binds, and is it reported as an output rather than buried in a MIN?
• What is the crossover CFADS, and how much headroom is there to it?
• Does the model use the lender's CFADS or the sponsor's?
• Is the target DSCR appropriate to the revenue contract, and does it step up for merchant tail years?
• Has a sensitivity been run on both sides of the crossover where headroom is thin?
• If gearing binds, has anyone tested whether the gearing cap is negotiable? It is a credit policy parameter, not physics.
• If DSCR binds, has capex reduction been pursued as hard as revenue optimisation?

Frequently Asked Questions

What is the difference between DSCR and gearing sizing?

DSCR sizing discounts the maximum permissible debt service — CFADS divided by the target coverage ratio — at the debt rate over the tenor. Gearing sizing simply applies a percentage to total project cost. The facility is the lower of the two.

What DSCR do lenders require?

For renewable and infrastructure projects with long-term offtake agreements, typically 1.20× to 1.35×, with 1.30× the most common figure. Merchant exposure pushes the requirement higher.

How do you know which constraint binds?

Compare the project's CFADS to the crossover value, which is gearing % × cost × target DSCR ÷ annuity factor. On the worked term sheet that is a cash yield on cost of about 9.35%: below it DSCR binds, above it gearing binds.

Does improving revenue always increase debt capacity?

No. In a gearing-constrained deal the debt is fixed by project cost, so a better revenue case improves equity returns and leaves leverage unchanged.

What is a debt tail?

A period of revenue extending beyond final maturity, so that a slipped repayment schedule still has cash flow behind it. A 20-year PPA typically supports a 17- or 18-year facility, and where the PPA ends before maturity the remainder is a merchant tail and is sized quite differently.

Is CFADS a standard definition?

No. It is a defined term in the credit agreement, and tax, maintenance capex, reserve movements and deemed generation payments are all negotiated. Two term sheets with the same headline DSCR can size very differently on different CFADS definitions.

Why would a sponsor and a lender disagree about which constraint binds?

Because they are using different CFADS. The lender applies haircuts to basis, curtailment, capacity accreditation and merchant pricing, and those can be enough to move a project across the crossover. The dispute is about revenue, not about sizing.

Closing: Two Numbers, One That Matters

Debt sizing is usually taught as a calculation and negotiated as a number. It is more useful to treat it as a diagnosis.

A gearing-constrained project is one the lender is comfortable with on cash flow and is limiting on principle — it wants the sponsor to have skin in the game. The conversation to have is about the gearing cap, which is a credit policy decision and is sometimes movable.

A DSCR-constrained project is one where the cash flow genuinely will not support more debt. No amount of arguing about equity contribution helps; the conversation is about the revenue assumptions, the tenor, and the coverage level — and about whether the capital cost can come down, which is the lever most sponsors reach for last and should reach for first.

The crossover between the two is computable in a single line before any model exists, and a project within a few percent of it is a project whose entire sensitivity profile is about to change. That is worth knowing at term sheet stage rather than at credit committee.

The next post takes the sizing calculation into its more difficult form: sculpting principal to hold a constant coverage ratio, and the circularity that every model runs into when it tries.

Sources: Wall Street Prep — Debt Sizing in Project Finance · Edward Bodmer — Debt Sizing · eFinancialModels — Understanding Debt Sizing and Project Finance Ratios · Renewables Valuation Institute — Debt Sizing with Target DSCR