Project finance advisory — Mastia Consulting

What this is for

Infrastructure and energy assets get financed against a cash flow that has not happened yet. The work is establishing how much debt that cash flow can carry, on what profile, with how much room for things to go wrong, and then defending those numbers in front of people paid to find the hole in them.

We do that for sponsors, developers and funds. Sometimes as the modelling arm of a larger team, sometimes as the only external party in the room.

Where we are most useful

Assets with a contracted period and a debt tail. The interesting question is rarely the base case. It is what the structure does in year eleven, when the offtake has ended and the price curve is doing the work a signed contract did before.

The work

What we actually do

Debt sizing

How much the asset carries against target coverage, and what that leaves for the equity contribution.

Amortisation and tenor

Sculpted, annuity or straight line, with what each does to the sponsor's return profile and to the tail.

Covenant package

Coverage tests, lock-up levels, cash sweep and cure mechanics, sized with headroom that survives a bad year.

Lender process support

Model releases, sensitivity runs on request, and answers to the technical adviser inside agreed response times.

In detail

The parts most people ask about

Coverage ratios

DSCR is the number a lender sizes against and the one a covenant is written on. It is also the number most often quoted without saying which definition is in use. We state ours in writing before the first model release, and we settle two questions with the lender before anyone runs a sensitivity: look-back or look-forward, and annual or semi-annual.

Definitions we publishshow the working

DSCR(t) = CFADS(t) ÷ debt service(t) debt service = scheduled principal + interest paid + net hedge settlements + commitment and agency fees CFADS = revenue − opex − tax − capex, including major maintenance ± movement in working capital ± movements in reserve accounts, if the term sheet includes them before debt service and before distributions LLCR = [ NPV(CFADS over the remaining loan life, at the debt cost) + DSRA and other available reserve balances ] ÷ debt outstanding PLCR extends the numerator to the end of the asset's life. The tail ratio, PLCR ÷ LLCR, is the number that actually measures the debt tail.

One convention worth arguing about

Placing reserve account movements inside CFADS is a choice, not a settled fact. In most credit agreements the DSRA sits below senior debt service in the waterfall, so funding it out of CFADS makes the ratio partly self-referential and flatters release years. Which one applies is set by the term sheet, not by us. We model the definition the term sheet gives, and the file states which one it uses.

The distinction between DSCR and LLCR carries more weight than it looks. DSCR is a flow test: it says whether this period works. LLCR is a stock test: it says whether the remaining debt is supportable at all. A structure can pass one and fail the other, and which one binds first tells you what kind of problem you have.

Sculpting, and where the circularity really is

Sculpted amortisation sets each period's principal so that DSCR lands on a target instead of falling out of a fixed schedule. It maximises debt capacity against a coverage constraint, at the cost of a repayment profile only as reliable as the cash flow forecast underneath it. On a contracted asset that is usually a good trade. On a merchant one it deserves an argument.

For the plain case there is no circularity at all: the debt quantum is the present value of CFADS divided by the target ratio, discounted at the debt cost, and the split between principal and interest then falls out in one pass. Anyone running an iteration block to sculpt a vanilla facility is iterating towards an answer that has a closed form.

Where iteration is genuinely requiredshow the working

Closed form, plain case: Debt = Σ_t [ CFADS(t) ÷ target DSCR ] ÷ (1 + r)^t Debt service(t) = CFADS(t) ÷ target DSCR Genuine circularities, each of which does need a solver: · interest during construction capitalised into the facility · arrangement fees charged as a percentage of the drawn amount · a DSRA sized on forward-looking debt service · tax, where the interest deduction depends on the debt, which depends on CFADS, which depends on tax The last one is the hard one and the one most often fudged.

Iterative calculation, and the honest argument against it

The usual objection is that two users see two different answers. For a contraction mapping like sculpting that is overstated: Excel converges to the same fixed point from any starting value, within tolerance. The real cost is that switching iteration on globally suppresses the circular-reference warning, so an unintended circularity, the accidental self-reference in a new row, stops announcing itself. We use an explicit solver block with a stated tolerance, and a staleness check that compares the inputs the block was solved against with the inputs now in the file. A solver block without that check trades a small numerical uncertainty for a silent stale answer, which is worse.

Send the term sheet

The structure as it stands and the date you are working to. You get a straight answer on whether it is sized right.

Emailinfo@mastiaconsulting.com LinkedInin/plarios