Structuring, Sizing and Underwriting Infrastructure Deals — Two Deals Modeled End to End, With the Models Supplied
Julian R. Sterling
These are the companion files of The Project Finance Handbook by Julian R. Sterling.The book on Amazon →
These are the two Excel workbooks that go with the book, the files every figure in it came from, now corrected; around them, a blank set for your own deal, thirteen working documents to print, forty questions that mark themselves and three cases the book names and never takes to a number. Northgate is a 96 MW data center, built to suit and let for fifteen
years on a triple-net lease: it costs 614.4 million to build and
657.4 million to fund, carries 427.3 million of senior debt,
and earns a year-one net operating income of 61.61 million, which is a yield on
cost of 9.37 per cent on the funding requirement and
10.03 per cent on the building alone. Its equity returns
16.60 per cent, of which only 2.41 points come from exit yield
compression, and its rent can fall 43.2 per cent before cover reaches
1.000 times. Aurora is 220 MW of solar with 80 MW of storage, contracted for
fifteen years and merchant for ten more: it costs 468.5 million to build,
against 242.9 million of debt and a 249.5 million equity check,
and it is in the book because it does not work. With a thirty per cent investment tax credit it
returns 5.59 per cent to equity. Without one it returns
2.68 per cent at the project level and 0.13 per cent to
equity, and needs a power price of 105.14 a megawatt-hour to reach ten per cent
where the subsidized deal needs 67.31. Change any input and every dependent
figure moves, because nothing in either file is pasted.
Free to download. No sign-up, no email address, nothing to fill in.
The two workbooks, the blank set, the three cases, the question book, the printable documents and the read-me. Each group below can also be
downloaded on its own. Last revised 24 September 2026.
Get an email if these files are corrected
Optional. When a workbook on this page is corrected or a new worked case goes up, readers on the list get one email, plus a short note when a new book comes out. The downloads do not need it.
These carry the book’s own figures. Both were corrected on 24 September 2026 and now end with a checks sheet; what changed is at the end of the read-me inside.
Chapters 5, 7, 8, 14, 15 and 16, and Appendix A
Northgate — the data center
Ninety-six megawatts of IT capacity at 6.4 million a megawatt, let at
56.00 a kilowatt a month on a fifteen-year triple-net lease indexed at
2.5 per cent, against 2.9 million a year of non-recoverable cost. The loan
is the lower of two tests, both performed on the face of the sheet: a
1.45 times cover test on year-one net operating income, and a
65 per cent loan-to-cost test that exists precisely because it distrusts
the forecast the first one believes. The sheet names which of the two binds rather than
leaving the reader to infer it, and it amortizes on a twenty-five year profile over a
fifteen-year term, so the balloon that falls due in the month the only lease expires is
printed rather than buried.
The funding table is circular by construction, since the loan sizes off a cost that
includes the cost of the loan. It is not solved by iteration. The Solve
sheet unrolls the loop one pass per column with a cell that reads CONVERGED, which is why
the file behaves identically wherever it is opened. Interest during construction is charged
on the whole drawn balance, arrangement fee and debt service reserve included, because both
are drawn on the first day and carried for the whole build: that correction alone moved
2.8 million through the loan, the installment, the cover ratio, the balloon
and the return. The equity return is computed with the construction period inside it. Set
the exit cap rate equal to the yield on cost and the answer becomes what carry and
indexation alone produce. Set the stress switch to YES and the debt stops re-sizing itself, which is the whole difference between a sensitivity and a stress case.
Revised 24 September 2026. The IDC sheet showed 16.9 million of interest during construction,
the construction draws only, while the funding table carried 19.7: the seventh correction had reached
the Solve sheet and not the sheet that displays it. It now ties. The equity cash flow was pinned to a year-ten exit whatever
the exit year said; it now follows the input. The pre-tax project return the book quotes as 13.14 per cent was in no cell:
built the way Chapter 11 defines it, it is 12.34 per cent. The third test of Chapter 5 is on the
sheet, a loan life cover ratio of 1.619 and the rent at which the balloon refinances,
27.58. Thirty-one checks.
Two hundred and twenty megawatts alternating current of solar with eighty megawatts and
three hundred and twenty megawatt-hours of storage, sold under a fifteen-year power purchase
agreement and then into the market for ten more. Debt is sculpted to the cash flow rather
than leveled, which is what a resource that varies year to year requires and what a lease
does not. The base case is a P50 resource at a
28.5 per cent capacity factor and 549,252 megawatt-hours.
Holding the loan fixed at the 242.9 million actually signed and running the
resource down its own distribution, a P99 year at 483,727 megawatt-hours takes minimum cover from 1.350 to 1.197 times and the equity return from 5.59 to 3.06 per cent; a P90 year takes it to 1.247, just through a 1.25 lock-up.
The input that decides the answer is the investment tax credit, and it is a cell rather than
a footnote. Set it to zero and the deal returns 0.13 per cent to equity,
which is the most useful thing in the file and the reason the deal is in the book at all.
The terminal value is a cell too: the base case gives the plant no residual value in year
twenty-five, which is conservative and is the first thing a reader ought to argue with. Put
a multiple in and watch what it is worth. At eight times final-year cash flow available for
debt service the return is still nowhere near ten per cent, which is why the finding
survives the objection, and why the objection is printed instead of avoided.
Revised 24 September 2026. The file was delivered with the credit at 0 per cent, although the book, the read-me
and the sheet’s own instructions start from thirty: it is back at 30, and the no-credit series now sits beside the other,
5.59 against 0.13 per cent. A stress switch holds the schedule signed at close, which
the book’s resource table describes and no file could reproduce: on it a P90 year covers 1.247, not the
1.250 printed, which re-sculpts the signed loan to the weaker sun. Two simplifications are priced on their own sheet: the
reserve comes back in year 25 although the loan is repaid in year 18 (+0.10 points),
and tax losses are forfeited rather than carried forward (+0.32 on the
unsubsidized return). Neither rescues the deal. Twenty-eight checks.
Both workbooks with every input emptied. The checks sheets count what is still missing and test what holds for any inputs.
Chapters 2 to 16 · Appendices A and B
A blank Northgate and a blank Aurora
The corrected models with every blue input emptied, the construction drawdown profile included: 29 inputs for a levered, cost-bound or cover-bound asset on a level installment, 46 for a sculpted, contracted-then-merchant one. Nothing computes until every input is in. Then the checks test what must hold for any deal: the loop converged, sources equal uses, the reserve is the stated months of the service the loan actually carries, the loan is the lower of the two tests, the balloon is the value of the installments the profile schedules after the term, and the equity series changes sign once.
The book’s inputs are listed in a text file inside the archive. Type them back in and the files return 427.3 million, 1.826× and 16.60 per cent; 242.9 million, 5.59 per cent with the credit and 0.13 without.
Two workbooks, the list of the book’s inputs and the read-me
The documents the chapters describe
Thirteen pages to print and use. A4, with margins wide enough for US Letter.
Chapters 2 to 15 · Appendices C to E
Thirteen working documents
The two questions on one page. Sizing on two tests and a third for what is left at maturity. The CFADS definition reconciled against the executed one. The waterfall to the cent, with the five accounts and the gate. Sources and uses with the circularity written out. The covenant package in dollars of revenue. The construction package tested against the owner’s loss. The downside pack labeled sensitivity, stress or scenario. The counterparty map. Rates and policy, and the contract that absorbs a change of regime. The term sheet, the diligence questions with their bad answers, and the model review: thirteen errors, thirteen tests.
Northgate is filled in wherever the book computes a figure, as the worked example beside an empty column for your deal.
A workbook that marks itself. Answer in the yellow cells; the marking fills in as you go.
All parts of the book
Forty questions on the book
Seven on the foundations, seven on sizing, sculpting and funding, seven on covenants and construction, seven on revenue, the offtaker, rates and policy, six on model review, stress and the deal with its corrections, and six on the three cases below. Numbers are marked to a tolerance, multiple choice to the letter, and the score sheet names the part to revise first.
New material, not a restatement. Three situations the chapters name and never take to a number. Each case is one workbook: the note is the first sheet, the model is the rest, and a checks sheet closes it. Type your own figures in and every finding recomputes.
Chapter 9 · Case one
Six months late
The chapter costs a six-month delay at 13.0 million of interest and says the whole stream moves right, then stops. Together they take the equity return from 16.60 to 15.10 per cent. Damages struck against the carry, the way the chapter says contracts usually are, repair about a third of it; the rate that makes the equity whole is 208 thousand dollars a day, 1.18 times a day’s rent, because the deferred exit is part of the injury and no daily rate mentions it.
Chapter 13 · Case two
The break nobody modeled
The swap broken at the year-ten sale, which the book declines to compute. One point lower, the borrower pays 13.6 million; 375 points lower, 55.0. Move the exit cap with the rate, as the chapter says the world does, and the same fall adds 63.0 million to the proceeds: the return rises to 17.14 per cent. A rise pays out on the swap and still costs 0.48 points, because the cap rate is the risk and the swap only a partial hedge of it.
Chapters 6, 8 and 11 · Case three
How often a sculpted loan locks up
Aurora at exactly 1.35× over a 1.25× lock-up, against a normal spread of the sun calibrated on the book’s P90. The sponsor should expect 1.47 lock-up years in fifteen, with a 79 per cent chance of at least one. Sizing at 1.40× cuts that to 0.46 and costs 8.3 million of debt: the cushion is bought, a hundredth at a time.
Three workbooks, each with its note and its checks
Conventions used throughout
Blue text
an input: change it and everything recomputes
Black text
a formula: do not overtype these
Grey text
a note
Solve sheet
the funding loop, unrolled one pass per column, with a cell that reads CONVERGED once the passes stop moving
Stress switch
on both assumptions sheets: NO re-sizes the loan to the new case, YES holds it as signed, 427.3 million on Northgate and Aurora’s eighteen sculpted installments
Checks sheet
the last sheet of each workbook: the book’s figures recomputed, then the identities that must hold for any inputs
ADDED
a figure the book does not give, used in the cases and marked wherever it appears
There are no macros, no external links, no protection and no circular references anywhere, and
nothing is locked or watermarked. The files behave identically in Excel, LibreOffice and Google
Sheets, whether or not iterative calculation is switched on.
Why there is no circular reference
Project finance models are circular by nature: the loan is sized off a cost that includes the
interest and fees on the loan. The usual remedy is to switch on iterative calculation and let
the spreadsheet settle. These files do not do that, for a reason worth knowing.
With iteration enabled, a spreadsheet can display cells from different passes at the
same time. Measured on an early version of these very models: a debt service of 37.18
shown on one sheet where the model had computed 39.77 on another. A model built to demonstrate
that arithmetic should be visible cannot itself hide a number that depends on which order the
cells happened to settle in.
The loop is therefore unrolled. One pass per column on the Solve sheet, each
column taking the previous column’s loan and recomputing the cost, the interest during
construction, the fee and the loan again. Four passes are enough on Northgate; Aurora settles in the first, because its cover test binds and the cost of the loan never enters its size. Each file carries eight, and a check cell reads CONVERGED when the last two agree. The reader watches the loop close instead of
taking it on trust.
Three reviews, eight corrections
These models were reviewed three times before publication and failed all three. Eight errors in
all, and not one of them was a typing mistake: every one produced a plausible number. They are
set out in Chapter 14 because they are the ordinary errors rather than exotic ones, and because
there is no useful version of a book about model review that pretends its own arithmetic was
immaculate.
The largest moved the headline. An equity return computed with no construction period in it
reported 19.09 per cent where the corrected model reports
16.60, an overstatement of 2.49 points produced entirely by
putting the equity in at time zero and the first rent one year later on a twenty-four month
build. The second review found that this same correction had been applied to one model and not
to the other, which is the commonest way a correction fails and the reason the workbooks are
worth more than the text: a reader who does not believe a number can put their own assumption in
and watch what happens to the rest.
Go beyond the companion files
The Project Finance Handbook develops the complete framework behind these models, with the assumptions, trade-offs and worked examples tied together.
The workbooks open in Microsoft Excel, LibreOffice Calc, Google Sheets and Numbers; the PDF prints on A4 or US Letter. They use no
macros and no add-ins, so nothing needs to be enabled or trusted. If your spreadsheet asks to
update links on opening, decline, because there are none. Northgate and Aurora are illustrative
models built for a book: they are not investment advice, and not a representation about any real
asset or transaction.
What changed, and when
24 Sep 2026
Both workbooks corrected. Northgate: the IDC sheet totalled 16.9 while the funding table carried 19.7, and now ties; the equity cash flow follows the exit year instead of a fixed year ten; the pre-tax project return is added, 12.34 per cent as Chapter 11 defines it, where the book prints 13.14; the loan life cover ratio (1.619, on 691.7 of discounted income, not 691.8) and the balloon’s refinancing rent (27.58) are on the sheet; three wrong lines in the read-me. Aurora: delivered with the credit at 0 per cent, now at the book’s 30 with the no-credit series beside it; the read-me’s 98.29 is 105.14; a stress switch holds the signed schedule, on which a P90 year covers 1.247 and a P99 year 1.197, not the 1.250 and 1.202 printed. Elsewhere in the book: no exit cap rate takes Northgate’s return to zero (the 28.66 per cent printed returns 1.55; ten years of operating cash already exceed the equity), the merchant-price table is a sizing case, not labeled as one, and the closed-form coefficient is 0.0462, not 0.0461. Added: 31 and 28 checks, the blank set, thirteen working documents, forty questions and three cases.
How do you calculate the LCOE of a solar project?Present value of lifetime costs over present value of lifetime MWh: a $100M, 100 MW solar plant at 24 per cent capacity factor and 7 per cent has an LCOE of $52.48.
Closing the DealTwo defensible bridges 20.70 million apart, a peg worth 7.00 million, and the six choices that remove 6.60 of a 12.00 earn-out.
CMBS and CRE CLOsWhere the loss actually lands, from appraisal reduction to realised severity, and what the B-piece is really being paid for.
How to Read a Commercial LeaseThe three refinements chapter 19 names and never performs, and the renewal rate below which the mark-to-market is worth nothing.
How to Read a Credit AgreementWhere the default actually comes from, the cure that costs 5.5 times the other, and the capacity nobody adds up.
How to Read a Real Estate Loan AgreementThe cure ratio in closed form, the four-point window in which the cheap cure works, and the cure sized to the wrong threshold.
Office Real EstateA six per cent yield that returns 3.2 per cent once the re-letting cycle is paid for, and the headline-to-net-effective rent arithmetic.
Private Equity Real EstateBoth worked waterfalls to the dollar, the two capital stacks, and the arithmetic of the promote made changeable.
Private Markets PerformanceThirty-one of the thirty-three figures chapter 19 publishes reproduce exactly — and the two that do not are named rather than quietly adopted.
Raising a Real Estate FundThe chapter 17 funnel run on a calendar — when the first close actually lands, and why more travel does not help.
Real Estate FinanceFour people look at one building and reach four numbers; the lender is whole only above 105,109,489, twelve per cent below today’s value rather than forty.
Real Estate Financial ModelingProperty, development and fund models built line by line, and the modelling test worked end to end.
Real Estate Fund ManagementThe waterfall of 6.11, the build-to-core of 8.7 and the proceeds gap, reproduced as live formulas rather than asserted.
REIT Analysis and ValuationFFO of 532.0, AFFO of 381.0, net asset value and dividend safety — every figure a formula you can change.
Retail Real EstateThe occupancy cost of every unit in a centre, the sixteen per cent of the rent roll no tenant can sustain, and the right-size-convert-or-hold decision priced.
Sale and LeasebackA €179.5 million transaction end to end, with rent cover measured on the entity that actually signs the lease.
Self-Storage Real EstateThe cohort engine behind a 590-unit store, and the rate increase on existing customers priced against the move-outs it causes.
The Fund Finance ProfessionalChapter 8 builds the reported-to-eligible NAV bridge; chapter 9 computes every ratio without it. Two points at every state — and what a subscription line does to the IRR.
The Growth Equity InvestorWhat a pro rata cheque really costs, and the band where defending your ownership loses money.
The Private Credit InvestorThe two coverage ratios are not measured on the same thing: the erosion is 47.7 per cent, not the 28.7 the headline implies.
The Private Equity Fund Controller PlaybookThe book defines IRR, DPI, RVPI and TVPI, tells you to update them at the exit, and prints not one value. Computed: a 1.833× deal inside a fund at 0.892 TVPI.
The Venture Capital AssociateWhat defending a position costs, and how many companies a reserve pool actually defends.
Financial Risk ManagementA fund inside every limit that cannot meet a redemption — and the number that decides it is the one with no currency attached.
Business ValuationThree advisers land 26.8 per cent apart on one company, and the whole gap turns out to be 1.96 points of perpetual growth.
Quantitative FinanceThree models agree to a quarter of one per cent about a number that one unobservable input moves a hundred and three times as much.
Asset ManagementFour people quote four returns for one mandate, all correct and 2.7017 points apart — forty-eight times the manager’s net skill.
Alternative InvestmentsA manager reports 13.29 per cent and the endowment earns 6.26 — both correct, and only a third of the advertised advantage arrives.
Venture CapitalOne company out of twenty-eight returns 56.7 per cent of the fund, and half the capital goes in after the decision — at half the return.
Machine Learning for FinanceFive people quote the accuracy of one credit model, all five are right, and the number that decides how much money it makes is none of them.
Commercial Real Estate InvestingThe equity earned 8.6647 per cent and the investor received exactly 8.0000 — the preferred return, and nothing above it.
Mergers and AcquisitionsThe board paper says the deal creates 13,436,667 of value. The arithmetic says it destroys 17,530,855. Nobody is lying.
DerivativesThe treasury report says the hedge cost 1,233,698. That is the interest differential, not a cost.
Treasury ManagementFive cash balances for one company, all correct and 145,600,000 apart — and the revolver that is two-thirds of the liquidity leaves at a revenue fall of 8.4127 per cent.
Financial Planning and AnalysisRevenue 3.0190 per cent above budget and operating profit 16.3209 per cent below it, in the same quarter, with every figure correctly stated.
Energy TradingA position report that is 91.7031 per cent hedged and correctly computed, on a book that is short 2,542,000 MWh — and a margin call of 198,400,000 the next morning.
Construction Cost ControlA contract sum of 26,301,102 became a final account of 29,153,363 on the building that was drawn — and 85.8 per cent of what was lost was knowable on the day it was signed.
Trade FinanceSix routes to payment on one 4,200,000 export order cost between 178,040 and 223,268, a spread worth 14.8 per cent of the margin, and a day of buyer credit costs 1,031.76.
Cost AccountingOne factory costed twice on the same 13,440,000 of overhead, and 4,053,091 moves between four product families.
Pricing StrategyA list price of 148.00, a pocket price of 112.51, and the nine deductions in between — with what one point of price is actually worth.
Contract ManagementA bid 5.50 a unit cheaper, signed on paper worth 9.81 a unit more — and 94.68 per cent of it was knowable on the day of signature.
Capital MarketsFour defensible readings of one bond, the cost of an issue under each convention, and the placement discount computed at a constant discount.
Financial RegulationThree ratios pass and the fourth is in breach by 1,956.0 million: the output floor, the liquidity buffer after encumbrance, and the loan split.
Bank ManagementFour defensible readings of one balance sheet, the deposit beta measured two ways, and the liquidity that lasts 3.9 days.
The Real Estate WorkoutFour ways out of one broken loan, discounted to today, and the 58.7 cents at which a payoff beats enforcement.
If this book helped — or didn’t — a few lines on Amazon are worth more than they look: they are what the next reader goes on. Write a review. The workbook stays free either way.