Companion files
Practitioner guides to private markets and real estate.
Some of the books come with working files — models, templates, checklists. They are free, they need no account and no email address, and they are listed here.
The equity bridge with the buyer’s column beside the seller’s, locked box against completion accounts on identical terms, five defensible working capital pegs, the earn-out with its six accounting switches — and the eighty questions on the price, live.
Twenty-eight quarters of a unitranche built from one sheet of assumptions, the add-back that is the only thing passing the test, the equity cure priced two ways, the basket capacity nobody adds up — and the sixty-four-point review checklist as a working document.
Four Excel workbooks: the complete Meridian House model, an empty template built to the same architecture, the three timed practice cases, and the 53-point review checklist.
The complete Meridian Industrial Trust analysis — FFO and AFFO bridges, the net asset value build, leverage and dividend safety — an empty template for a REIT of your own, two timed practice cases, and the 54-point analysis checklist.
The complete measurement of one fund from cash flows to reporting line, a PME calculator for a fund of your own with all four methods, the three practice cases, and the 51-point measurement checklist.
The asset business plan with a live rent roll and projection, the hold-refurbish-sell comparison, the accept-or-hold-out arithmetic for a vacancy decision, and the monthly asset dashboard.
The residual appraisal live, the worked scheme rebuilt from first site visit to disposal, land diligence and the risk register.
The Harbourgate reporting pack, five worked cases, and a sheet that rebuilds your own capital account from the statements.
Net effective rent, valuing an ageing asset, and the development letting decision.
A borrowing base running off a live tape of 1,880 contracts, the Meridian transaction, and trigger design.
The Thalia transaction stage by stage, the bid funnel, and the fees behind the roll decision.
The money and the instruments, the 35 per cent content test, and proposal economics.
The JV waterfall and promote, the complete venture, and failure to fund, clawback and removal.
The margin set beside the return on capital, the rate floor priced, and the round trip that shows what prepayment protection is really for.
What a minimum payment actually does to a balance, what the smallest-first method costs against the highest-rate one, what a starter emergency fund costs while you are still in debt — and a calculator for one debt of your own.
The grey zone in expense allocation expressed as basis points on the fee, a broken deal allocated four defensible ways, and the sample size that turns “test periodically” into something an examiner can be told.
The two illustrations the book leaves without figures, computed: what ten early years are worth against thirty late ones, what a percentage point of fees costs over a working life, and a calculator for your own numbers.
The reserve arithmetic Chapter 12 says you should be able to run and never runs: what defending a position costs round by round, how many companies a reserve pool actually defends, and how much a fund should reserve.
The worked deal model with its return decomposition and stress case, the bottom-up NAV build, the 30-point underwriting checklist, and the interview and case preparation.
The distribution waterfall with the catch-up solved in the cell, the quarter-end close simulation cross-checked two ways, a capital call and capital account file, and all ten operational checklists.
The lender's model reproducing both published coverage ratios, the 33-point underwriting checklist, the interview and memo preparation — and the Ridgeline fund-economics bridge, where 11 per cent gross and 8 net turn out not to be measured on the same thing.
Four workbooks: the fund waterfall of Chapter 6 with the three readings of a “20% promote” side by side, underwriting and exit sensitivity, debt sizing under all four binding tests, and the budget, variance and close cycle.
Three workbooks: both worked waterfalls of Chapter 12 to the dollar, the value-add and development capital stacks, and the performance measures and LP diligence scorecard.
Three workbooks: a fulcrum finder that names the breaking tranche by arithmetic, the recovery model with fees, DIP priority, form and time, and the sum-of-the-parts valuation against an independent liquidation floor.
Three workbooks: the Calder Toll Road debt sculpt of Chapter 4, where leverage is a result rather than an input; the fifteen diligence workstreams with the risk register of Chapter 8; and one asset valued twice to the last decimal, which computes the seven claims Chapter 6 makes without arithmetic and corrects three of them.
Five workbooks: the market approach with its calibration confronted against the entry price, the terminal value computed two ways, the equity waterfall with the preference stack solved in the cell, one credit loan marked from performing through to recovery — and the hardest judgment calls computed, where a conventional thirty per cent marketability discount turns out to be unreachable by the method meant to evidence it.
One workbook: the three refinements Chapter 19 names and never performs — the over-market stream the headline capitalises in perpetuity four paragraphs before warning against it, the twenty-eight per cent that time-weighting removes, and the leasing cost that takes 5,788,000 down to 807,365 — plus the renewal rate below which the whole mark-to-market is worth nothing, which is 43.7 per cent.
One workbook: the escrow rule in one line — full collateralisation at one minus the assumed tax rate, which the chapter never names — the two negotiations it recommends that pull against each other so exactly that every dollar won on the tax clause lands in the uncollateralised column, and the “roughly doubles” that turns out to be five thirds.
One workbook: the 11.2-times cure ratio derived in closed form as 1/(k·c) — independent of the breach, entirely dependent on the two terms a borrower negotiates — the four-point window of income decline in which the cheap cure exists at all, a worked cure sized to a threshold the loan does not impose and 3.8 times what the covenant requires, and the extension paydown priced at a quarter of the figure the book floats.
One workbook: the Northbeck cash profile the chapter builds and stops one year short of — 2027 carrying two obligations at once for €54,063 against a lump sum of €27,063, the buffer sized at €90 of carry with a break-even that needs no penalty figure, the repurchase cap that turns out not to bind at any realistic size, and a surrender formula that counts one term twice in the direction Article 26(1) sanctions.
One workbook: Chapter 19's six findings reproduced to the unit, then asked the questions the book does not — the eighteen basis points that are really thirty-one once the retention, the indemnity and the premium are counted, the two workstreams out of ten that paid for everything, the ground investigation the chapter skips priced at a 2.4 per cent break-even, and the one piece of scoping arithmetic in the book, corrected.
One workbook: Chapter 17's funnel run on a calendar — the printed column against the rule printed beside it, the eight commitments arriving as two lumps on two committee dates and never reaching the hundred million, a third of the money moving on which week in June the raise begins, the first close that pays 62.5 per cent of its own management company's burn, and the target size below which Chapter 20's first-close rule and Appendix C's break-even cannot both hold.
One workbook: the back book sized on Chapter 16's own segmentation — 82,062 analyst-days for a quarter of a million customers — the remediation team costed two ways, the month beyond which no headcount recovers a one-year transition window, the fifth of the portfolio that never answers a letter, and the premium the chapter's base plan actually costs.
One workbook: Chapter 13's three business models reconciled to the figures it prints, the levered return it asserts and never builds — 18.6 per cent flat, 22.6 with a three per cent escalator — what a twelve-month energization slip really costs against an eight per cent rent miss, and the powered-land hit rate below which a fourfold multiple on the winners is still a losing business.
One workbook: the reverse stress test the book states and does not show, computed on its own Fund IV — the severe quarter as a waterfall, the credit facility's advance rate moving the answer from 34.2 per cent to 17.2, the undrawn borrower commitments that turn out to exceed the gate they are supposed to sit behind, and the book's approximately 19 per cent confirmed inside the error band rather than corrected.
One workbook: the exit carried through to equity, multiple and IRR — the units a committee actually uses; the 150,000 line the book prices at 11 million of protected revenue and then leaves out of its own total; and the half-turn of exit multiple that is worth more than the whole four-year programme beside it.
One workbook: what a co-investment programme really saves in basis points, the total-loss rate that erases the whole fee advantage, the deal count that turns out to be the binding constraint — and the collision between two policies from the same book that pushes a portfolio out of its band on its own.
One workbook: why dilution at a fairly priced round takes nothing from you, what a pay-to-play clause is worth in turns of multiple, and the reserve sized as an expectation rather than a habit — where the optimum is smaller than the instinct and twenty per cent loses money.
One workbook: what a pro rata cheque really costs and why the book's own sentence about it is wrong, the band of exit values in which defending your ownership loses money while the position still looks like a winner, and the reserve tested like-for-like against simply backing twice as many companies.
Three workbooks: the Northbridge borrowing base and why it is not availability, the arithmetic a subscription line does to the IRR — and to the multiple — and NAV facility underwriting with the one figure in the book that does not reconcile.
One fund read six ways: DPI against total value, the net return with the valuation and on realised cash alone, the harvest window, the public market equivalent, and what a subscription line does to each.
What a five-person single-family office costs to run, what a multi-family office charges on a tiered schedule, and the level of wealth at which the two curves cross — with the decision reframed as a hurdle rather than a bill.
What the liquidity sleeve costs, what separates two share classes of the same fund, and the headline internal rate of return at which a closed-ended fund stops beating an evergreen one and merely ties it.
The cohort forecast of Chapter 15, run for eighty-four months: gross flows that rise every single month while the organic growth rate peaks in month twenty-nine and falls by three quarters — and the signature that tells a forecast miss apart.
The value creation bridge of Chapter 13, built in full and reconciling to zero: operating work is 41 per cent of the value created, deleveraging almost 20 — and the add-on figure most decks print is wrong by 73.
The rating model of Chapter 3, built and then broken: three managers with a flaw the book calls terminal all receive the model’s top rating, and refining the rubric makes it easier to hide one, not harder.
The occupancy cost of every unit in a shopping centre, the sixteen per cent of the rent roll no tenant can sustain, the right-size-convert-or-hold decision — and an eighty-question retail acquisition checklist.
A six per cent office yield that returns 3.2 per cent once the re-letting cycle is paid for, the headline-to-net-effective rent arithmetic, the refurbish-reposition-convert-or-sell decision, and eighty questions to ask before bidding.
The model behind a 520-bed scheme, revenue per available bed against the rate card, the letting campaign week by week, and the cost base per bed.
The cohort engine behind a 590-unit store, the rate-increase decision on existing customers, a lease-up tracker and a bid sheet.
The resident engine behind a seventy-bed nursing home, the payor mix and the conversion of self-funders, a rota and employer-cost calculator, and a bid sheet.
The model behind a €179.5 million sale and leaseback, rent cover measured on the entity that actually signs the lease, a lease abstract, a covenant monitor and a bid sheet.
An eighteen per cent discount to net asset value that is really 9.18, the unsmoothing arithmetic, the portfolio bid, the sell-or-roll decision and a seventy-two-question checklist.
The conduit deal model, the loss severity bridge from appraisal to realised, the ARA and control mechanics, the B-piece return, the CRE CLO equity, and the refinancing gap.
The IRB capital calculation, the SEC-IRBA tranche formula for any attachment and detachment you enter, the bank's economics and both break-even margins, the investor's return, and the bargaining range on one sheet.
The standard formula charge on any asset, the four candidates charged and ranked, Annex IV and what diversification is worth, the marginal cost at two different insurers, and every lever a manager controls.
The bridge from market value to recovery value with every adjustment switchable, the three sizing tests with the binding one named automatically, the thirty-eight loan tape and its seven-tranche waterfall, and the unsmoothing of a reported net asset value.
One corporate hedging programme over five years: the number the treasury report calls a cost and which is not one, the collar nobody priced, and the margin call that arrives twelve months before the money it is protecting.
One acquisition quoted seven ways: the three prices that are the same number, the lease qualification worth four points of the headline, the four accretion columns a banker gets to choose between, and the margin ladder that admits no solution.
One estate over ten years: the six numbers people call the entry yield, the void that makes the rent line fall in five years out of ten, all twenty-four ways of attributing the return, and the waterfall that pins the investor at the hurdle.
One credit model with five defensible accuracies, the field that leaks and takes 0.1772 with it, the applicants the model never saw, and the cut-off priced in two different years — where it moves thirty points and turns a profit into a loss.
Four multiples for one venture fund, the power law measured with a tick box on every position, initial cheques set against follow-on reserves, and the carried interest under three conventions that differ by more than the entire management fee.
Five cash balances for one company on one morning, all correct; net debt three ways; the thirteen-week forecast error turned into a reserve and a derived minimum; the revenue fall at which the revolver is no longer there; three ways to hold 250,000,000, priced; the discount, the supply chain finance programme on both readings, the pool, the swap and the hedge that lost money and was right.
One quarter, one set of correct accounts: revenue 3.0190 per cent above budget, gross profit 3.7591 per cent below it and operating profit 16.3209 per cent below it. The discount and the mix separated and sized, five comparators that read the same quarter 10.6610 per cent apart, a corporate pool that moves 11,924,386.35 of segment profit between two defensible denominators, and the discount the pack convicted that repays in 4.4114 years.
One power station, one retail book and a position report that is correctly computed and describes a different company: 91.7031 per cent hedged against a true short of 2,542,000 MWh. The plant valued four ways off one price duration curve, the load shape that takes 86.0857 per cent of a booked margin, and a 40.00 rise that calls 198,400,000 the next morning against 158,000,000 of liquidity.
One borrower, four defensible EBITDAs and four leverage ratios from 3.19x to 6.47x, the add-back bridge with a switch on every line, both covenants and which one binds, the downside carried through to cash, and expected loss set against the margin.
Three correct returns for one private programme on one call schedule, the multiples and the over-commitment a target allocation really needs, the public market equivalent against any index, and what unsmoothing does to the beta the risk model was solved on.
The four returns of one mandate on one set of inputs, chain-linking against an internal rate of return with the cash flows editable, Brinson attribution that reconciles to the active return, and the fee schedule against a passive counterfactual.
The closed form and a thousand-step lattice as live formulas with convergence recomputed on your inputs, all five Greeks with a finite-difference check beside each, what eight points of volatility are worth, and the delta hedge run out step by step.
Three methods on one set of facts with the gap decomposed, the five-year build and the terminal value with its share of the answer on the page, the control premium converted into the annual synergies it demands, and the reverse solve for the growth a price assumes.
Value at risk built one visible step at a time with the correlation grid switchable to crisis, the stress book and the reverse stress, and a liquidation engine that runs six redemptions under two selling policies and reports what is left to sell after each.
The list-to-pocket waterfall on one account and on a whole company of 180, the nine leakages ranked, the break-even volume of a discount with the rebate cliff solved by bisection, the price-volume-mix bridge with the elasticity as an input cell, and a blank calculator for your own book of business.
Both bridges from the contract sum to the final account with all eleven lines and their workings, the appraisal run six ways with the break-even solved in closed form, six one-input sensitivities that print every line, and a calculator for a project that has not been signed yet.
One factory costed twice on the same 13,440,000 of overhead, with the identity that both methods distribute it in full; the five pools and their driver rates with capacity and the avoidable split; the decisions from the batch and the make-or-buy to the death spiral; and a calculator for costing a part of your own.
One export order priced down all six routes at the terms they are really quoted at, the documentary credit in every form with the risk it removes set against the risk it introduces, the cost of the money from a discount that is not a yield to a receivable sold outright, and a calculator for your own payment term.
The Larchmere case priced clause by clause with both bridges and the month-nineteen crossover, nine standalone clause calculators, a calculator that prices your own contract before you sign it, and the one-page negotiation ledger with the value of every concession.
The loan split that makes a portfolio weigh more than its average loan, the output floor with its binding year computed, the liquidity buffer before and after encumbrance, and the subordination rule written both permitted ways.
The interpolation that decides a concession, the cost of a bond issue under each convention, the placement discount and the leak computed at a constant discount, and the register of every choice with its value in euros.
The four capital readings of one balance sheet, the deposit beta measured two ways, the economic value of equity across the whole surface of its assumptions, and the liquidity that lasts 3.9 days.
Four ways out of one broken loan, each discounted to today so that they are comparable, and the 58.7 cents in the dollar at which a discounted payoff beats enforcement.
Die Liquiditätsquote, die eine Warteschlange wirklich verlangt: 241 Anfangsbestände, 13,2 Prozent, zwei Punkte davon nicht dem Manager zuzurechnen. Vier Arbeitsmappen auf Deutsch.
Two infrastructure deals modeled end to end and one of them does not work: a data center that costs 657.4 million to fund and returns 16.60 per cent with only 2.41 points of it coming from the exit, a solar and storage project that returns 5.59 per cent with the tax credit and 0.13 without it, and a funding loop unrolled one pass per column so there is no circular reference anywhere.
A business the bank would finance at 1.56 times covers its own debt 0.96 times in its first year: the add-back bridge from 917,000 of broker SDE to 676,000 of EBITDA, an SBA 7(a) loan fixed by the valuation cap, and a cleaning company priced for the day its largest customer leaves.
Four managers give four numbers for the same pile of stock: a cycle service level of 98 per cent that is a fill rate of 99.3846 per cent, 90.5155 per cent of the variance coming from the supplier’s lead time, and 24,154.77 EUR a year for making that supplier reliable, on one part number from the first page to the last.
A management fee invoiced at 462,834.23 under both designs while one investor pays less and the others pay more, a 6.8 million capital call the facility cannot bridge once a pending transfer leaves the borrowing base, and an operations team at 69 per cent of capacity over the year that is 18 per cent over it in April.
A nonpaying first lien offered at 42 cents on its balance, worth 69,754 at 20 per cent if the court is done by month 30 and 114,835 in a non-judicial state, a modification that fits the borrower’s budget yet sells for less than the keys, and an eight-loan pool that fails its own floor at the asking price.
A credit portfolio of $500,000 in seventeen lines that displays 7.46 per cent on the day of purchase and earns 6.02 per cent over three years, a BDC showing 14.67 per cent that earns 7.76, and an interval fund paying 9.00 per cent while its net asset value falls, because defaults, fees and liquidity take their share and a few lines decide the portfolio.
An analyst initiates on a fictional industrial company with a Buy and a $101 price target, twenty times FY2027 adjusted EPS of $5.06; one quarter later the company beats, the estimates move 0.9 per cent, the shares 5.6 per cent, and the firm's own rating grid turns the Buy into a Hold.
A tax sale that advertises 18 per cent, a certificate won at 0.25 per cent that earns 20.3 per cent a year if redeemed in month 2 and 4.4 per cent if held to month 12, and sixty certificates expected to return 7.54 per cent, below their own 8 per cent hurdle, because the few parcels nobody redeems decide the portfolio.
A data center bought for $900 million whose signed lease pays back 34.2 per cent of the price, a power clause that takes the equity return below target at a 4 per cent rise in market prices, a fiber network that needs 39.4 per cent penetration to earn 10 per cent, and 850 towers whose best ground buyout is multiple arbitrage.
A group with $145 million of liquidity on paper and $22.5 million it can actually use, a 13-week cash flow that dates the crisis, an uptier and a drop-down priced holder by holder, and a Chapter 11 plan whose value decides who is paid and who is not.
A fictional industrial park bought clause by clause: the deposit and the hard date, title and survey, reps that survive, prorations, the closing statement and the funds flow.
A first private equity fund launched by the numbers: structure across three vehicles, the launch budget, terms, a second close and its equalization, and ten years of management company accounts.
A platform owned after the deal: a regulated water utility, a solar portfolio and a schools concession, run through covenants, a refinancing, quarterly valuation and the exit.
A $600 million fund of funds across primaries, secondaries and co-investments: pacing, the J-curve, both layers of fees and the gap between gross and net.
A 150 MWac solar project taken from the first land option to commercial operation: energy yield, the utility PPA, the virtual PPA and the merchant case, tax credits, debt, tax equity and the value at each milestone.
A tank storage platform sold at auction: the seller's plan and the diligence bridge, the DCF, the locked box, acquisition debt, the walk-away price and why the lower bid won.
A regulated water utility bought at 1.60x its rate base and a seawater desalination plant bid as a build-operate-transfer project: rate cases, PFAS and lead, tariffs, sculpted debt and the price of water.
A $750 million real estate debt fund: sizing a bridge loan with the four tests, repo leverage, a loan that goes wrong and the workout, and the net IRR to investors.
A 200 MW / 800 MWh battery taken from interconnection to valuation: revenue stacks, tolls and floors, degradation and augmentation, fire codes, the investment credit and the foreign entity rules.
One two-unit station valued four ways: a licence renewal, a power contract with a data center buyer and an uprate, with the fuel, the decommissioning fund, the tax credits and the risks priced.
What happens after the keys are handed over: a 36 MW campus run for three years, its contracts, its power bill, its incidents and service credits, its staffing and the retrofit that liquid cooling forces.
A closed-end fund's calls, equalization, waterfall and capital accounts, and an evergreen fund's dealing, gates, performance fee and a NAV error corrected investor by investor.
Three infrastructure loans from the lender's chair: a schools PPP whose quiet stress is inflation, a wind farm whose contract ends before its debt, and a holdco loan that lives on another company's dividends.
Reading guides: which book to read first, field by field →
Articles — worked arithmetic from the books →
Free Excel templates — LBO, DCF, 13-week cash flow, project finance, REIT and more →
REPE interview questions, a worked case study and the modeling test →