Private Markets II: Private Debt, Real Assets/Infra & Private-Company Valuation

Ring: Private Wealth & Private Markets (Finance-Plus galaxy) · Node PW1.03 · ~18 hours · Mastery gate ≥ 85%

A quoted share price is a fact you can look up. Nothing in this module is. A direct-lending term sheet buries its real yield inside an original-issue discount and a payment-in-kind toggle that a spreadsheet's headline "coupon" cell will not show you. An infrastructure concession's cash flow does not go on forever the way a listed REIT's does, it runs out, by contract, on a specific date, and treating it like a perpetuity is not a simplification, it is an error. And a private company has no price at all: two entirely competent analysts can look at the same business and land 30% apart, not because either one made an arithmetic mistake, but because they were silently valuing two different things, a controlling interest versus a minority sliver, a stake with a buyer waiting versus one with none. The three subjects in this module (private/direct debt, real assets and infrastructure, and private-company valuation) are united by exactly that fact: value here flows through a negotiated structure, not a public clearing price, and your job is to read the structure with the same rigor you would bring to reading a quoted price. This Ring is additive institutional/CFA breadth, it presents the toolkit and the plumbing alongside, not in place of, the program's process-over-P&L, concentration-over-diversification philosophy.

You already own the machinery this module extends. M3.04 taught you to discount a cash flow; M3.06 taught you to build a comp; M3.10 taught you what breaks when a business is not a clean going concern. Nothing here re-derives any of that. What is genuinely new is the layer on top: how a privately negotiated loan prices differently from a publicly traded one; how a cash flow that is contractually guaranteed to end changes everything about the terminal value you learned to build for a business that keeps existing; and how the exact same guideline-public-company multiple has to be adjusted up, down, or not at all, depending on a question most learners never think to ask first, whose interest, exactly, am I valuing, and can they sell it whenever they want?

Learning objectives

By the end you can:

1. Explain why direct lending exists as a distinct asset class (bank retrenchment, BDC/AIF fund vehicles), and describe unitranche/first-out–last-out (FOLO) structuring, original issue discount (OID), payment-in-kind (PIK) toggles, and the maintenance-versus-incurrence covenant distinction.
2. Compute a direct loan's all-in yield (cash coupon plus OID pickup) to an expected life, read a leverage/interest-coverage covenant test for cushion or breach, and price a call-protection premium.
3. Distinguish core, core-plus, value-add, and opportunistic infrastructure/real assets by their risk drivers, and explain why "infrastructure" spans a far wider return band than the label suggests.
4. Build a concession-life discounted cash flow to equity (separating cash yield from total return) and explain why the terminal value is contractually zero, never a going-concern perpetuity.
5. State the levels-of-value hierarchy (marketable minority → controlling → nonmarketable) and identify, for any private-company valuation task, exactly which adjustment (control premium, discount for lack of control, discount for lack of marketability) applies, in which direction, and in what order.
6. Compute a control-premium/DLOC bridge and a discount for lack of marketability (DLOM) using both an empirical-study anchor and the Chaffe option-pricing proxy, and name the Mandelbaum factors that justify a chosen figure.
7. (Productivity objective: R10 duality.) Know which category of AI tool belongs in private-markets diligence work, and why this specific corner of the corpus, a judgment call with no lookup-able right answer, is the single highest AI-over-trust zone you will meet.
8. State this module's mastery gate (≥85%) and pass it.

The duality, stated once (R10). As in every module of this branch, objectives 1–6 and 8 are the understanding gate; objective 7 is the productivity payoff you keep. Nowhere in the corpus is the gap between the two more dangerous than here: a DLOM or control-premium figure is never a fact to retrieve, it is a defensible judgment built from named evidence, and an AI assistant will hand you a confident, specific percentage with no defensible basis whatsoever, the purest form of AI0.01's "plausible, not true" applied to a number that decides what a family actually inherits or what a lender actually recovers. Verification here is not a courtesy step; it is the entire discipline.

Prerequisites & connections

Builds on. M3.04's discounted-cash-flow mechanics and M3.06's comparable-company and precedent-transaction machinery are assumed fluent, this module never re-derives how to build a DCF or a comps table, only how to adjust their output for a private, illiquid, or contractually-finite setting. M3.10 supplies two things you will lean on repeatedly without re-reading them here: its distressed/IBC Section 53 waterfall (§4.7 there) is exactly where a defaulted direct loan ends up, and its general triangulation discipline, never average disagreeing methods, explain the gap, is the mindset this module's private-company bridge applies to a much narrower, more personal decision. E11.01's leverage/coverage ladders, covenant vocabulary, and rating logic are assumed; this module extends them into the privately negotiated, bilateral setting rather than the broadly syndicated or publicly rated one. E11.02 defined the control premium off a public target's unaffected trading price, a private company has no such price, so §4.10 shows you how the same concept travels into a world with nothing to anchor against except the value you just built yourself. M6.04's behavioral-finance material (loss aversion, the endowment effect, anchoring) is the quiet undercurrent of every family-business valuation conversation you will ever sit in on, an owner's attachment to "what my company is worth" is rarely neutral. PW1.02's fund-economics vocabulary (J-curve, called capital, carry on realized gains) is contrasted directly in §4.4 against direct lending's current-income engine, same GP–LP alignment grammar, a different cash-flow shape entirely.

Feeds into. PW1.01's concentrated-position unwind and estate-transfer planning for a business-owning client consumes this module's private-company bridge as a direct input, you cannot advise on "what will my children actually receive" without first knowing which interest, at which discount, is actually being transferred. AL1.02 will place infrastructure and real assets inside a portfolio-allocation lens (expected return, inflation linkage, diversification), this module supplies the individual-asset valuation machinery that allocation decision has to sit on top of. ES1.01 and ES1.02's fiduciary-duty and conflicts-of-interest material is the formal ethical frame around a technical fact this module states bluntly in §4.12: DLOM and control-premium judgment calls are a documented, litigated site of appraiser conflicts, and the discipline that protects you is technical (cited evidence, written-down reasoning) before it is aspirational.

PW1.04 owns the step that comes before this one. Everything above assumes a value has already been built and asks what to do to it: which level of value it sits at, and which discount or premium moves it to the level the question requires. PW1.04 owns the build itself. That means the income and market approaches to a private company, the normalisation adjustments that turn reported earnings into a figure a multiple can be applied to, and the cap table and preference stack that decide which shareholder actually receives what. The normalisation list is the one most often skipped: owner compensation, related-party rent, non-recurring items, and excess or deficient working capital. Read it before this node's bridge if you are starting from a set of financial statements rather than from a value somebody has already produced.

Core teaching content

4.5 The infra/real-asset risk spectrum

"Infrastructure" is not one risk-return band, and treating it as one is the fastest way to misprice an asset. Core assets are operating, with contracted or regulated cash flows and minimal volume risk, an operating toll road paid on an availability basis, a regulated utility, a wind or solar farm fully covered by a long-dated power purchase agreement (PPA). Core-plus carries modest lease-up or re-contracting risk. Value-add carries real construction or ramp-up risk and some merchant (uncontracted) price exposure. Opportunistic is greenfield development or distressed turnaround with full merchant exposure. Illustrative equity-return bands by tier (mid-2026 fundraising anchors, check the current data from a source such as Preqin or Cambridge Associates infrastructure benchmarks): core roughly 7–9%, core-plus roughly 9–12%, value-add roughly 12–16%, opportunistic 16–20%+. The word "infrastructure" spans that entire range, which is exactly why AL1.02's allocation-level treatment of infra as an asset class only becomes useful once you have first located a specific asset on this spectrum, which is this module's job.

Infrastructure's odd character (bond-like discipline in a good year, real equity risk in a bad one) comes from stacking a residual equity claim on top of long-duration, often-leveraged, contracted cash flows: the contracted revenue discounts like a long bond, but a genuine shock (a traffic collapse, a PPA counterparty default) hits the leveraged equity claim first and hard, exactly the way M3.10's equity-as-option framing describes a distressed firm's stub equity. Keep that connection in your pocket; §4.7's worked examples make it concrete.

4.6 Cash yield versus total return: the return-of versus return-on capital distinction

Two numbers get quoted about the same asset and they answer different questions. Cash (distribution) yield is a single-period snapshot: this year's distributable cash flow, divided by the equity value today. Total return (IRR) is the return over the entire remaining life of the asset, and for a depleting or amortizing asset it necessarily blends return on capital (a genuine profit margin) with return of capital (the concession or contract paying down toward its own contractual end). A high current cash yield on an asset that is giving your principal back is not the same thing as a high sustainable return, exactly the way a bond trading near par yields differently from one trading at a steep discount that is about to be redeemed.

This is precisely where a REIT's cap rate (M5.09) and an infrastructure concession's distribution yield must not be read the same way. A REIT is underwritten, for valuation purposes, as a perpetual cash machine, buildings get re-let, refurbished, and held indefinitely, which is exactly why M5.09's NAV and cap-rate math assumes an ongoing entity. A concession has no such perpetuity: its cash flow runs to a contractually fixed date and then, for your purposes as the equity holder, stops. Quoting a concession's current yield the way you would quote a REIT's stabilized cap rate, as if it signals a durable, indefinite return, is a common and expensive category error, and §4.7 exists to fix it. Practice P8 gives you the cleanest possible numeric illustration of the gap this section is naming: a flat cash flow bought at a price, with a hard stop and no terminal value, yields a very respectable current cash number and a materially lower true IRR, because part of every payment you receive is simply your own money coming back to you.

4.8 India and US anchors

In India, the National Highways Authority of India (NHAI) monetizes completed public roads to private or InvIT operators through TOT (Toll-Operate-Transfer) bundles, a fixed-term concession over an already-built road, which removes construction risk but leaves real traffic risk with the operator. The Hybrid Annuity Model (HAM), by contrast, has NHAI itself fund roughly 40% of project cost during construction and pay the balance as a fixed annuity through the operating period, sharply reducing the operator's revenue risk relative to a pure toll-based Build-Operate-Transfer (BOT) concession. That single structural difference is exactly why HAM assets sit near the core end of §4.5's spectrum while pure toll/BOT assets sit further out on real traffic risk, precisely the gap Worked Example 3 surfaces between an assumed "core" 11% discount rate and the roughly 17% the market may genuinely be demanding for a traffic-exposed toll concession. InvITs (Infrastructure Investment Trusts regulated under the SEBI (Infrastructure Investment Trusts) Regulations, 2014, as amended) are required to distribute at least around 90% of Net Distributable Cash Flows to unitholders on a semi-annual basis (check the current requirement against the live regulation); that mandatory-distribution design is exactly why an InvIT's quoted "distribution yield" is a §4.6 cash yield, not a promise about total return.

In the US, long-dated toll-road and airport concessions (the Chicago Skyway and Indiana Toll Road privatizations of the mid-2000s are the standard reference transactions for how a US toll concession gets structured and financed, pull the actual transaction documents for exact terms and figures rather than citing a remembered number) anchor the core end of the spectrum. Contracted renewable power and midstream energy infrastructure occupy a similar core/core-plus band when the offtake is genuinely contracted (Worked Example 4's shape), and slide toward value-add or opportunistic as merchant price exposure rises, the same risk-tiering logic, a different regulatory alphabet.

4.12 Putting it together: the full valuation bridge, and the fiduciary warning

State the whole bridge as a checklist, because in practice it is executed as one: (1) build a base value on a marketable-minority basis, using guideline public comparables per M3.06; (2) identify exactly which interest you are valuing (controlling or minority) and for what purpose (a sale, a gift, a divorce, financing, an ESOP); (3) if controlling, multiply by (1 + control premium), and apply a DLOM to that controlling value only if it also genuinely lacks any near-term liquidity path (contrast Worked Example 5's two scenarios: a real deal in motion needs no DLOM, an illiquid family patriarch's controlling stake with no buyer in sight still does, though typically at a smaller discount than a minority stake would carry); (4) if minority, apply DLOC first only if your base value was built on a controlling assumption, then apply DLOM; (5) write down every Mandelbaum factor that moved your discount within the empirical range before you know what number the client is hoping to hear.

State the fiduciary warning just as bluntly. Gift- and estate-tax valuations are exactly where an appraiser's incentives can quietly pull a DLOM or control-premium figure toward the convenient answer, a family that benefits from a lower value for gift-tax purposes, a divorcing spouse who benefits from a higher value on the other side of the table, an ESOP trustee bound by fiduciary duty to the plan rather than to the selling founder. This is a real, litigated, IRS- and Department of Labor-scrutinized failure pattern, not a hypothetical one, and it is precisely why Revenue Ruling 59-60's demand for a reasoned, written, factor-based approach still matters seventy years later. The discipline that holds is the same one AI0.01 named for an entirely different reason: document your reasoning before you know what answer is wanted, cite your empirical anchors by name, and be able to defend every factor to a skeptical reviewer, or a Tax Court. Cross-reference ES1.01 for the formal ethics and conflicts-of-interest frame; this module supplies the technical judgment that frame exists to discipline. And cross-reference PW1.01 directly: this entire bridge is the exact input a private-wealth advisor needs before they can competently answer a business-owning client who asks, in plain language, what their company is actually worth and what their children will actually receive.

{
  "skill": "Screening current direct-lending and infrastructure pricing benchmarks before a model is built",
  "use": "Use a research-enabled AI tool for a first-pass read on current SOFR/benchmark rates, direct-lending spread indices, and listed-InvIT distribution yields, so a model's rate and spread assumptions start from something current rather than a stale training-data guess.",
  "tools": ["Research/agentic browsing tools", "A market-data terminal if available (Bloomberg, Refinitiv)"],
  "prompt": "Find the current SOFR fixing, the current spread of a direct-lending benchmark (e.g., the Cliffwater Direct Lending Index) over SOFR, and the current distribution yield of two listed Indian road or power InvITs. For each figure, name the exact primary source and its as-of date.",
  "verify": "A bare chat assistant's memory of every one of these figures is frozen at its training cutoff (AI0.01), and each one moves with markets and semi-annual InvIT distribution announcements — date-check every figure against its named primary source (the relevant central bank/reference-rate administrator for SOFR, the index provider's own published page, the InvIT's own regulatory filing) before it enters a model, and re-pull before every live engagement rather than trusting a number from last month's chat.",
  "diy": "You must already know what each benchmark means (why SOFR replaced LIBOR, what an InvIT's Net Distributable Cash Flow actually contains) to sanity-check a pulled figure against the mechanism this module taught, rather than pasting in a number you cannot interpret.",
  "market": "IN"
}

Common mistakes & how experts think differently

1. Treating a unitranche's single blended rate as the whole story. The borrower sees one number; the First-Out and Last-Out tranches behind the Agreement Among Lenders carry genuinely different spreads, priorities, and recovery profiles. An expert asks "what's the FOLO split?" before assuming a unitranche behaves like a single, uniform loan.
2. Confusing OID with PIK. OID is a one-time funding discount, pure yield pickup, no compounding effect on the coupon. PIK compounds the principal balance itself, growing the lender's real exposure every year it is elected. Mixing them up mis-states either the lender's yield or its exposure at exit.
3. Reading a covenant breach as an automatic default. Most breaches trigger a negotiated waiver or amendment (often for a fee) not acceleration. The lender's real leverage is the ability to accelerate, which is what brings the borrower to the table; experts read a breach as the start of a negotiation, not the end of the loan.
4. Valuing an infrastructure concession's current yield like a REIT's cap rate. A REIT (M5.09) is underwritten as a perpetual cash machine; a concession's cash flow runs out on a contractually fixed date, and its yield mixes return of and return on capital (§4.6, Practice P8). Quoting one number without asking which kind of asset it describes is a category error.
5. Applying a Gordon-growth terminal value to a concession DCF. The terminal value is contractually zero (or a stated handback figure), never a perpetuity, no matter how attractive the early-year growth looks (§4.7, cross-ref M3.05). This is the single most expensive mechanical mistake this module guards against.
6. Stacking a DLOM on top of an already-illiquid private-transaction-comp base, or applying both a full control premium and a full minority discount to the same interest. Double-counting the same adjustment under two different names (§4.9, Practice P16), always ask which single adjustment your starting value already implies before adding another.
7. Subtracting the control-premium percentage directly to get a "minority discount." DLOC = 1 − 1/(1 + premium) is a reciprocal relationship; a 30% premium implies roughly a 23% discount, not 30% (§4.10, Practice P12). Experts run the reciprocal by hand every time rather than trusting the intuition that the two numbers should match.
8. Quoting a single "rule of thumb" DLOM ("private companies just get a flat 30% off") without running the Mandelbaum factors or an option-pricing cross-check. The empirical literature spans roughly 10% to 50%+ depending on holding period, dividend policy, and transferability restrictions; a defensible appraisal shows its work rather than asserting a number (§4.11).
9. Accepting an AI-drafted valuation memo's stated DLOM or control premium as if it were sourced. A fluent, confident percentage with no cited study and no Mandelbaum reasoning attached is a hallucination wearing a business suit: treat it exactly as AI0.01 taught you to treat any other confident, unverified figure, and trace it before it appears in anything you sign.

Worked examples

Worked example 1: Cascade Industrial Holdings (synthetic, US$): unitranche structuring, all-in yield, and a covenant test

A sponsor is financing a buyout of Cascade Industrial Holdings (synthetic), EBITDA $50.0mm, with a $250.0mm unitranche facility, exactly 5.00x EBITDA. Behind the single facility the lenders split it via an AAL into a First-Out tranche of $100mm (2.00x EBITDA) priced at SOFR+350bps and a Last-Out tranche of $150mm (3.00x EBITDA) priced to make the blended rate the borrower is quoted come out to SOFR+575bps.

Solution. Step 1: solve the Last-Out spread. Set (100×350 + 150×x)/250 = 575 and solve: x = (575×250 − 100×350)/150 = (143,750 − 35,000)/150 = SOFR+725bps. Check: (100×350 + 150×725)/250 = (35,000 + 108,750)/250 = 143,750/250 = 575bps: confirmed.

Step 2, the blended cash coupon. Using an illustrative SOFR of 4.30% (as of mid-2026, check the current fixing), the blended cash coupon is 4.30% + 5.75% = 10.05%, so cash interest is 10.05% × $250mm = $25.125mm.

Step 3: covenant test. Interest coverage = EBITDA/cash interest = 50/25.125 = 1.99x against a stated covenant minimum of 1.75x, a cushion of (1.99/1.75 − 1) = 13.7%. Gross leverage = 250/50 = 5.00x against a covenant maximum of 5.75x, a cushion of (5.75/5.00 − 1) = 15.0%. With $8.0mm of cash on the balance sheet, net debt is $242.0mm and net leverage is 242/50 = 4.84x. Both covenants pass with real, but not enormous, headroom, exactly the profile a direct lender underwrites, deliberately, so that a modest EBITDA shock (Practice P4 tests a 15% decline) can flip the interest-coverage test to a breach.

Step 4, all-in yield including OID. The facility is funded at 99 (1 point of OID). Per 100 of face value, the lender's cash flows over a 4-year expected life (before any refinancing) are [−99.00, 10.05, 10.05, 10.05, 110.05]. Solving for the internal rate of return of that stream gives an all-in yield of 10.37%, a 0.32 percentage-point pickup over the 10.05% cash coupon alone, purely from the OID. (Practice P2 recomputes this at a deeper 97 issue price.)

Step 5: call protection. Say the facility carries a non-call period in year one, then a 102 call price in year two, 101 in year three, and par thereafter. If the sponsor refinances the loan away at the start of year three (after two full years), the applicable premium is the year-three price of 101, costing 1% × $250mm = $2.50mm.

Read it. Five mechanically distinct numbers (a blended spread, a cash coupon, two covenant cushions, an OID-inclusive yield, and a call premium) all live inside what the sponsor experiences as "one loan at SOFR+575." An analyst who reads only the headline spread has read none of the actual economics.

Worked example 1b: the mezzanine layer: PIK toggle and a warrant kicker

Below Cascade's unitranche sits a $30.0mm mezzanine tranche paying 8% cash plus a 4% PIK toggle (12% total coupon), plus warrants for 5% of fully diluted equity.

Solution. Step 1, the PIK-only balance. If the borrower elects full PIK on the 4% strip every year, the principal compounds: after 5 years, balance = 30.0 × 1.04⁵ = $36.500mm.

Step 2, the full cash-flow build. Computing cash-pay interest on the then-current PIK-accreted balance each year: year 1 begins at 30.000 (cash-pay 2.400, PIK accretion 1.200, ending balance 31.200); year 2 begins at 31.200 (cash-pay 2.496, PIK 1.248, ending 32.448); year 3 begins at 32.448 (cash-pay 2.596, PIK 1.298, ending 33.746); year 4 begins at 33.746 (cash-pay 2.700, PIK 1.350, ending 35.096); year 5 begins at 35.096 (cash-pay 2.808, PIK 1.404, ending balance 36.500), at which point the note is repaid in full and the warrants, assumed, illustratively, to be worth 5% of an exit equity value of $80.0mm, or $4.00mm: are exercised. The investor's full cash-flow stream is [−30.0, 2.400, 2.496, 2.596, 2.700, 43.31] (year 5 = 2.808 cash-pay + 36.500 principal + 4.00 warrant value ≈ 43.31).

Step 3, the IRR, warrant included. Solving that stream gives an IRR of 13.89%, noticeably above the 12% cash-plus-PIK coupon alone, entirely because of the warrant kicker. Strip the warrant out and the IRR would fall back toward the coupon; the kicker is what makes mezzanine capital tolerate its subordinated position in the capital stack.

Read it. PIK toggles and warrant kickers are two separate levers pulling in different directions on the same instrument (one defers cash to the borrower's benefit, the other transfers upside to the lender's benefit) and a lender's real expected return depends on modeling both, not on quoting the stated coupon.

Worked example 2: Saraswati Precision Components Pvt Ltd (synthetic, India ₹): a structured NCD, security cover, and a hand-checked XIRR

A Category II AIF extends ₹150 crore of secured, privately-placed non-convertible debentures (NCDs) to Saraswati Precision Components (synthetic), issued 15 March 2023, cash coupon 11.0% p.a. paid annually, maturing 15 March 2026 (3-year tenor) with a ₹15 crore redemption premium at maturity.

Solution. Step 1, the cash-flow stream. Annual coupon = 11.0% × ₹150 crore = ₹16.5 crore. The investor's cash flows are [−150.0, 16.5, 16.5, 181.5] crore, where the final payment bundles the last coupon, the ₹150 crore principal, and the ₹15 crore redemption premium (16.5 + 150 + 15 = 181.5).

Step 2, simple IRR versus a true XIRR. Ignoring the exact calendar (treating each period as a clean year) gives an IRR of 13.910%. But 15 March 2023 to 15 March 2024 spans a leap day, so the actual day-count is not a uniform 365 days throughout, computing the true, date-based XIRR (actual/365) gives 13.895%, a small but real difference that would matter on a larger book or a less regular schedule (Practice P6 gives you a genuinely irregular-date case where the gap is much larger).

Step 3: security cover. The debentures are secured against collateral (property and receivables) valued at ₹200 crore, giving a security cover ratio (SCR) of 200/150 = 1.33x, comfortably above an illustrative 1.25x regulatory floor (check the current minimum under the applicable SEBI framework, the required cover has been recalibrated across several circulars), a cushion of (1.33/1.25 − 1) = 6.7%.

Step 4: borrower coverage. With borrower EBITDA of ₹60 crore against ₹16.5 crore of interest, interest coverage is 60/16.5 = 3.64x, a healthy cushion consistent with a bullet (non-amortizing) structure, where the covenant discipline leans more heavily on interest coverage and security cover than on debt-service coverage.

Read it. A "coupon + redemption premium" structure (common in Indian private placements partly for withholding-tax efficiency) hides its true yield exactly the way a US OID does; you must build the actual cash-flow stream and solve for the rate, never read the stated coupon as the answer.

Worked example 3: Deccan Expressway Trust (synthetic, India ₹): a toll-road concession-life DCF

Deccan Expressway Trust (synthetic) holds a toll concession with 15 years remaining. Year-1 toll revenue is ₹120 crore, growing 5% annually; opex runs 25% of revenue; a major-maintenance reserve of ₹15 crore is set aside every year; and outstanding debt of ₹300 crore, at 9% interest, amortizes in equal ₹20 crore annual principal installments, exactly extinguishing the debt in year 15, the same year the concession itself expires.

Solution. Step 1: build the schedule. The full 15-year build (₹ crore):

(Years 4, 6, 7, 9, 11, 13 follow the same mechanical pattern between the rows shown.) Distributable cash flow to equity climbs from ₹28.00 crore in year 1 to ₹141.39 crore in year 15, driven by both revenue growth and the shrinking debt-service burden as principal amortizes, then, at year 15, the concession reverts to NHAI and the cash flow simply stops.

Step 2: cash yield. Against a quoted (illustrative) equity value of ₹310 crore, the year-1 cash/distribution yield is 28.00/310 = 9.03%, a plausible InvIT-style current yield.

Step 3, the full DCF, and the gap it exposes. Discount all 15 years of distributable cash flow at an 11% "core" equity rate, the kind of rate an availability-payment (HAM) asset might command. That gives a DCF-intrinsic equity value of ₹468.78 crore, some 51% above the ₹310 crore quoted price. Equivalently, the quote sits 34% below the DCF: 310/468.78 − 1 = −33.87%. That gap is not evidence of an error; it is a prompt to check the discount rate against the asset's actual risk. Solving instead for the rate that makes the 15-year cash-flow stream exactly equal the ₹310 crore quoted price gives an implied IRR of 17.17%, and a hand check confirms it: discounting the identical 15 cash flows at 17.17% reproduces ₹310.00 crore almost exactly. Because this is a toll-based BOT concession with real, uninsured traffic risk (not a HAM availability-payment asset) a market-implied required return near 17% is a defensible reading, not a mispricing: §4.8 named exactly this distinction, and this is what it looks like in the arithmetic.

Step 4, total return over the concession's life. Summing all 15 years of distributable cash flow gives ₹1,201.07 crore against a ₹310 crore initial equity outlay: 3.87x the money over the life, blending real profit with the simple return of your own capital as the concession runs down to its contractual zero.

Read it. Never trust a single "yield" number on a concession asset without asking three questions in order: what is the cash yield today, what does a full-life DCF say at a stated discount rate, and does that discount rate actually match this specific asset's risk. HAM-like or genuinely traffic-exposed? Worked Example 4 shows the same discipline applied to a US asset with a different shape of risk entirely.

Worked example 4: Prairie Wind Partners (synthetic, US$): a contracted-plus-merchant-tail infrastructure DCF

Prairie Wind Partners (synthetic) operates a wind farm generating 400,000 MWh annually under a 15-year power purchase agreement (PPA) at $45.00/MWh in year 1, escalating 1.5% annually; opex starts at $6.0mm, growing 2% annually; a $1.0mm annual major-maintenance reserve is set aside; and $75.0mm of construction-era debt at 6.0% amortizes in equal $5.0mm annual principal payments, extinguished exactly when the PPA expires.

Solution. Step 1, the contracted period. Year 1: revenue = 400,000 × $45.00 = $18.000mm; opex $6.000mm; EBITDA $12.000mm; interest 6% × $75mm = $4.500mm; principal $5.00mm; distributable cash flow to equity = 12.000 − 1.000 − 4.500 − 5.00 = $1.500mm. As the contracted price escalates and the debt amortizes, the same build produces $3.310mm by year 5, $5.610mm by year 10, and $7.955mm by year 15, a strongly growing distribution profile, driven by both the PPA escalator and the shrinking debt-service burden.

Step 2: discount the contracted stream. At a 7.5% discount rate appropriate to genuinely contracted, low-volume-risk cash flow, the present value of the full 15-year contracted stream is $36.115mm.

Step 3, the merchant tail. After year 15 the PPA ends; assume 10 further years of merchant (uncontracted) operation at a conservative $30.00/MWh merchant price, 90% availability (aging-turbine derate), giving revenue of 400,000 × 0.90 × $30.00 = $10.800mm and, against $5.0mm of opex, a flat $5.800mm/year EBITDA (debt is fully repaid, so this flows straight through to equity). Because this cash flow is genuinely riskier, uncontracted price, aging equipment, discount it at a higher 12% rate for its own 10-year horizon: present value at year 15 = $32.771mm. Bringing that lump sum back to today at the contracted 7.5% rate for 15 years gives $11.076mm.

Step 4: total value and yield. Total equity value = $36.115mm + $11.076mm = $47.190mm. Year-1 cash yield = 1.500/47.190 = 3.18%, low today, by design, because early-year debt service is heavy relative to EBITDA; the yield grows every year thereafter as the asset delevers, a different growth mechanism than Deccan Expressway Trust's traffic-growth story but the same underlying shape.

Read it. Valuing the merchant tail at zero (the correct treatment for Deccan Expressway Trust's contractual reversion in Worked Example 3) would be wrong here: it would understate this asset's equity value by 23.5% (36.115/47.190 − 1), because a merchant tail is genuinely uncertain, not contractually void. The skill this pair of examples drills is not "always zero the tail" or "never zero the tail", it is reading which contract you actually hold before you touch the discount rate.

Worked example 5: Chandler Valve & Fittings, Inc. (synthetic, US$): DLOM via the Chaffe option-pricing proxy

Chandler Valve & Fittings (synthetic), EBITDA $8.0mm, is benchmarked against guideline public comparables trading at 9.0x EBITDA.

Solution. Step 1, the marketable-minority base value. EV = $8.0mm × 9.0 = $72.0mm; less $12.0mm net debt gives a marketable-minority equity value of $60.0mm.

Step 2, a real control sale, no DLOM. If the company is being sold outright to a strategic acquirer, apply a 25% control premium (within E11.02's empirical range): $60.0mm × 1.25 = $75.0mm. Because a real, near-term transaction is in motion, the acquirer has its own natural path to liquidity through the deal itself, no DLOM applies to this scenario.

Step 3, an illiquid minority gift, DLOM required. Separately, suppose the family instead gifts a 15% non-controlling interest for estate-planning purposes, with no pending sale. The pro-rata marketable-minority value is 15% × $60.0mm = $9.000mm, but this interest cannot be sold on demand, so a DLOM must be subtracted before it is a defensible gift value.

Step 4, compute the DLOM by hand, via Chaffe. Assume equity volatility σ = 45%, a risk-free rate r = 4.5%, and an expected holding period T = 4 years until a realistic liquidity event. Because the put is at-the-money (strike = today's value), d1 = (r + ½σ²)T / (σ√T) = (0.045 + 0.5×0.45²) × 4 / (0.45×2) = 0.585/0.9 = 0.6500; d2 = d1 − σ√T = 0.650 − 0.900 = −0.2500. Then N(−d1) = N(−0.650) = 0.2578 and N(−d2) = N(0.250) = 0.5987. The DLOM (put value as a fraction of the underlying) is e^(−rT)×N(−d2) − N(−d1) = e^(−0.18)×0.5987 − 0.2578 = 0.8353×0.5987 − 0.2578 = 0.5001 − 0.2578 = 24.22%.

Step 5, the final value, and a cross-check. $9.000mm × (1 − 24.22%) = $6.820mm. Check this against the empirical evidence families from §4.11: 24.22% sits squarely inside the restricted-stock-study range (roughly 20–35%) and well below the wider pre-IPO-study range (40–60%+), a sensible landing point given the company has a demonstrated dividend history and no unusual transferability restrictions (Mandelbaum factors 2 and 6 both push toward the lower half of the range).

Read it. The same $60.0mm marketable-minority base value produced a $75.0mm control-sale figure and a $6.820mm illiquid-minority-gift figure for a 15% slice, a nearly 5x difference in per-dollar-of-equity terms, entirely from asking whose interest and how liquid before touching a single adjustment.

DLOM sensitivity, computed the same way, across holding period and volatility:

Both drivers point the same direction (a longer expected holding period and a more volatile underlying business both raise the DLOM) which is exactly the intuition an unpriced put option should give you: more time and more uncertainty both make the forfeited right to sell more valuable.

Worked example 6: Konkan Agro Processors Pvt Ltd (synthetic, India ₹): DLOM via the restricted-stock method, plus a control/DLOC bridge

Konkan Agro Processors (synthetic) is benchmarked against listed small-cap agro-processing peers trading at 12.0x P/E, against normalized net profit of ₹18 crore.

Solution. Step 1, the marketable-minority base. ₹18 crore × 12.0 = ₹216 crore.

Step 2, a family gift of a 10% minority stake. Pro-rata value = 10% × ₹216 crore = ₹21.60 crore. Applying a restricted-stock-study DLOM of 30% (a reasoned choice within the empirical range, defended by the Mandelbaum factors, a closely-controlled family business with no dividend track record and real transfer restrictions in its articles points toward the higher end of the range): ₹21.60 crore × (1 − 30%) = ₹15.12 crore, the defensible gift value.

Step 3, a control acquisition of a 60% stake. Applying a 30% control premium: ₹216 crore × 1.30 = ₹280.80 crore total controlling equity value; a 60% stake is 60% × ₹280.80 crore = ₹168.48 crore. Because this is a real, negotiated acquisition, no DLOM applies.

Step 4, the DLOC check. What minority discount is implied by that same 30% control premium, if you needed to move the other direction? DLOC = 1 − 1/1.30 = 23.08%, not 30%. This is the reciprocal-not-subtraction trap from §4.10, made concrete: a 30% premium and a 23.08% discount are the same fact stated from two directions, and treating them as equal in magnitude is simply wrong arithmetic.

Step 5 (extension), a controlling stake with no buyer in sight. If instead of a real acquisition this were an estate valuation of a family patriarch's controlling-but-illiquid 60%-equivalent position, with no pending transaction, a smaller control-DLOM might still apply: illustratively 12%: ₹280.80 crore × (1 − 12%) = ₹247.10 crore, versus the full ₹280.80 crore a real, in-motion deal would command. Control reduces (it does not eliminate) the marketability question.

Read it. Four adjustments, one base value, and every one of them depends on a fact pattern question you must ask before you touch a formula: is this interest controlling or minority, and is there a real, near-term path to cash or not?

Practice set

Work every problem before reading its solution. Figures are to about ±2% unless an exact formula makes a tighter match natural.

P1. Blended FOLO rate. A unitranche facility splits into First-Out $60mm at SOFR+300bps and Last-Out $90mm at SOFR+700bps. What is the blended spread quoted to the borrower?

Solution. Blended = (60×300 + 90×700)/(60+90) = (18,000 + 63,000)/150 = 81,000/150 = SOFR+540.0bps.

P2. All-in yield with deeper OID. Using Worked Example 1's cash coupon (10.05%) and 4-year expected life, recompute the all-in yield if the facility is instead issued at 97 (not 99).

Solution. Cash flows per 100 face: [−97.00, 10.05, 10.05, 10.05, 110.05]. Solving for the IRR gives an all-in yield of 11.02%, noticeably higher than the 99-price case's 10.37%, since a deeper discount funds the same eventual par repayment with less cash out today.

P3. Mezzanine PIK-only balance, no cash pay. A $20mm note pays its entire 11% coupon in kind. What is the principal balance after 7 years?

Solution. 20 × 1.11⁷ = $41.523mm.

P4. Covenant cushion under an EBITDA shock. Using Worked Example 1's facility (cash interest $25.125mm, covenant minimum ICR 1.75x), what happens to the interest-coverage test if EBITDA falls 15% from its $50.0mm base?

Solution. EBITDA falls to 50.0 × 0.85 = $42.50mm. ICR = 42.50/25.125 = 1.69x, below the 1.75x minimum, a cushion of −3.3%, i.e., a breach. A modest, plausible EBITDA decline is enough to flip a comfortably-passing covenant into a technical default, which is precisely why direct lenders retain maintenance testing rather than waiting for an incurrence trigger.

P5. Call premium at an earlier refinancing. Using Worked Example 1's $250mm facility and its 102/101/par call schedule, what premium applies if the sponsor refinances at the start of year two instead of year three?

Solution. Year-two call price is 102, so premium = $250mm × 2% = $5.00mm: double the $2.50mm premium computed in Worked Example 1 for a year-three refinancing, illustrating why call protection is priced to decay over time.

P6. A genuinely irregular XIRR. An investor pays $100 on 10 Jan 2024, receives $8 on 10 Jul 2024, $8 on 10 Jan 2025, and $116 on 30 Sep 2025 (note the final date is not exactly six months after the prior payment). Compute the XIRR.

Solution. Solving the date-weighted (actual/365) cash-flow equation for these four genuinely irregular dates gives XIRR ≈ 19.01%, a case where the schedule's irregularity (the final stub period runs about 8.6 months, not 6) makes a naive annualized-IRR shortcut materially wrong, and only a true date-based solve is defensible.

P7. Security cover ratio, a failing case. Collateral is valued at ₹180 crore against ₹150 crore of NCDs, with a 1.25x minimum requirement. Does this pass?

Solution. SCR = 180/150 = 1.20x, below the 1.25x minimum, a cushion of −4.0%, i.e., a fail. Contrast this directly with Worked Example 2's 1.33x pass: the same structure, a smaller collateral pool, and the covenant flips.

P8. Cash yield versus IRR, the canonical gap. A concession pays a flat ₹10 crore annually for 5 remaining years, then stops (terminal value zero), and can be bought today for ₹40 crore. Compute both the current cash yield and the full-life IRR.

Solution. Cash yield = 10/40 = 25.0%: a striking headline number. But solving [−40, 10, 10, 10, 10, 10] for its IRR gives only 7.93%, because 25% of the purchase price is returned to you every single year, most of each payment is simply your own capital coming back, not a sustainable 25% return. This is §4.6's return-of-versus-return-on-capital distinction in its purest numeric form.

P9. A smaller concession DCF. A different concession asset has 5 remaining years of distributable cash flow: ₹20, ₹22, ₹24, ₹26, ₹28 crore, with terminal value zero. At an 10% discount rate, what is the NPV?

Solution. Discounting each year at 10% and summing gives NPV = ₹89.54 crore.

P10. Why is the terminal value zero here, but not in a going-concern DCF? Explain, referencing M3.05.

Solution. M3.05's terminal-value discipline (growth capped at the economy's growth rate, a sustainable steady-state ROIC) exists for a business that keeps operating after your explicit forecast window, there is a real, ongoing entity to value beyond year N. A concession's cash flow to you as equity holder contractually ends on a fixed date, written into the concession agreement itself; there is no "beyond year N" left to value. Zero is not conservatism here, it is what the contract says.

P11. Order the levels of value. From most to least valuable per dollar of underlying earnings, order: nonmarketable minority, marketable minority, controlling & marketable, nonmarketable controlling.

Solution. Controlling & marketable (rare, a control stake with a ready buyer) > marketable minority (the public/guideline-comp basis) > nonmarketable controlling (private, no ready market, but can direct cash flows) > nonmarketable minority (private, no ready market, and cannot direct cash flows), the worst-off position, correctly carrying the deepest combined discount.

P12. DLOC from a 40% control premium. What discount for lack of control is implied?

Solution. DLOC = 1 − 1/1.40 = 28.57%, again, meaningfully less than 40%, and again a reciprocal, not a subtraction.

P13. A fresh Chaffe DLOM. Compute the DLOM for σ = 55%, r = 4.0%, T = 6 years.

Solution. d1 = (0.04 + 0.5×0.55²) × 6 / (0.55×√6) = 0.8518; d2 = 0.8518 − 0.55×√6 = −0.4955. DLOM = e^(−0.24)×N(0.4955) − N(−0.8518) = 34.55%, noticeably higher than Worked Example 5's 24.22%, consistent with both a longer holding period and higher volatility.

P14. Applying a restricted-stock DLOM. A marketable-minority pro-rata value of $4.5mm is subject to a 28% restricted-stock-study DLOM. What is the final value?

Solution. $4.5mm × (1 − 28%) = $3.240mm.

P15. The full bridge, one pass. Comparable EV/EBIT is 10.0x; the target's EBIT is $6.0mm and net debt is $5.0mm. A 20% non-controlling stake is being gifted, subject to a 32% DLOM. Compute the final value.

Solution. EV = $6.0mm × 10.0 = $60.0mm; equity (marketable-minority) = $60.0mm − $5.0mm = $55.0mm; a 20% stake = $11.0mm; after a 32% DLOM: $11.0mm × (1 − 32%) = $7.480mm.

P16. Spot the error. An AI-drafted valuation memo starts from a private-transaction-comparables base value (already reflecting the illiquidity buyers priced into those deals), then separately applies a 30% DLOM and a 25% minority discount on top, before arriving at a final figure. What is wrong?

Solution. This double-counts illiquidity under (at least) two, and arguably three, different names. Private-transaction comparables already embed the market's own illiquidity discount in the price actually paid, layering a second, separately-sourced DLOM on top counts the same effect twice. Stacking a further, undifferentiated "minority discount" without checking whether it duplicates either the comp basis or the DLOM compounds the error a third time. The fix is qualitative, not arithmetic: identify which single lens your starting value already reflects (a public marketable-minority basis needs its own DLOM; a private-transaction basis generally does not need a second one layered on top) and apply only the adjustments that specific starting point actually requires, exactly the §4.9 discipline, and exactly the failure mode the ai-augment-json callout in §4.11 warns you to check for by hand.

Applied mini-project: the private-markets triage memo

Build a three-part memo, one part per domain this module covers, using entities you invent and clearly label (synthetic).

Part A: direct-lending credit memo. Invent a sponsor-backed borrower with an EBITDA figure of your choice. Structure a unitranche facility with a First-Out/Last-Out split, compute the blended spread and an all-in yield including OID over a stated expected life, and test the resulting leverage and interest-coverage covenants against thresholds you specify, state clearly whether each passes or fails, and by how much cushion.

Part B, a concession-life or PPA-life DCF. Invent an infrastructure or real asset with a stated remaining contract life, a revenue/opex/debt schedule, and a major-maintenance reserve. Compute both the year-1 cash yield and the full-life discounted equity value, with the terminal value stated explicitly (zero, or a named contractual handback figure), and justify, in one paragraph, which risk tier (§4.5) you placed the asset in and why that discount rate is the right one.

Part C, a private-company valuation bridge. Invent a family-owned business with a guideline-public-comparable multiple. State clearly which interest you are valuing (controlling or minority, for what purpose), apply the correct adjustment(s) in the correct order, and, before you compute a final number, write down at least three Mandelbaum factors that justify where your DLOM sits inside the empirical range.

Self-check rubric (10 points; pass at 8/10 with no zero on any item):

Case Lab note: this module does not currently map to a live guided case in the app's Cases tab; use the self-check rubric above as your grading standard until one lands.

Reading & resources

Everything below is an anchor to verify against and learn from, never a substitute for reading the actual primary document when a real figure matters. Free unless marked; tags run beginner-to-advanced.

• Shannon Pratt & Alina Niculita, Valuing a Business: The Analysis and Appraisal of Closely Held Companies, the standard reference text for the levels-of-value hierarchy, DLOM, and control premium. [Paid] [Intermediate–Advanced]
• David B.H. Chaffe III, "Option Pricing as a Proxy for Discount for Lack of Marketability in Private Company Valuations," Business Valuation Review, December 1993, the option-pricing DLOM method worked in full in Worked Example 5. [Paid/library access] [Advanced]
• Mandelbaum v. Commissioner, T.C. Memo 1995-255 (US Tax Court), the nine-factor test named in §4.11, in the court's own words. [Free, full text via public Tax Court/legal databases] [Intermediate]
• IRS Revenue Ruling 59-60 (1959): the foundational demand for a reasoned, factor-based valuation of closely-held stock. [Free: irs.gov] [Intermediate]
• American Society of Appraisers Business Valuation Standards; AICPA Statement on Standards for Valuation Services No. 1, the professional standards a real appraisal must follow. [Free summaries / paid full standards] [Advanced]
• Cliffwater Direct Lending Index (research and methodology notes) a live, publicly documented private-credit benchmark referenced in §4.1 and §4.4. [Free] [Intermediate]
• LSTA (Loan Syndications and Trading Association): market conventions, standard documentation, and recovery data for the leveraged/direct loan market. [Free, some content paid] [Intermediate–Advanced]
• ILPA (Institutional Limited Partners Association): private-credit fund reporting templates and fee/term guidance; read alongside PW1.02's fund-economics vocabulary. [Free] [Intermediate]
• SEBI (Alternative Investment Funds) Regulations, 2012 and SEBI (Infrastructure Investment Trusts) Regulations, 2014, the India fund/vehicle framework underneath §4.1 and §4.8. [Free: sebi.gov.in] [Intermediate]
• NHAI/MoRTH's own TOT and HAM framework documents, the primary source for India's road-concession structuring. [Free: nhai.gov.in, morth.gov.in] [Intermediate]
• US Investment Company Act of 1940, as amended by the Small Business Credit Availability Act (2018), the legal framework behind BDC structuring in §4.1 and §4.4. [Free: sec.gov] [Advanced]
• Aswath Damodaran's NYU Stern illiquidity-discount notes and lecture materials (pages.stern.nyu.edu), a valuation-theory bridge from the public-market DCF/comps machinery this module assumes into the private-market adjustment layer it adds. [Free lecture notes / paid textbook chapters] [Intermediate–Advanced]
• Moody's and PitchBook/LCD leveraged-loan recovery studies: the primary data behind §4.4's recovery-rate ranges; pull the current cycle's figures rather than the illustrative ones quoted here. [Paid/subscription] [Advanced]

Teach it back & journal

Feynman prompt. Explain to a colleague who has only ever valued public companies why the same guideline-public-company multiple can legitimately produce three different numbers for the same private business (a control-sale figure, an illiquid-minority-gift figure, and a controlling-but-illiquid figure) using Worked Example 5's $75.0mm, $6.820mm, and (if you extend it) a control-DLOM figure as your concrete anchors. If your colleague can, unprompted, ask "wait, whose interest, exactly, are we valuing, and can they sell it whenever they want?" before you get to the punchline, you have installed the one question this entire module is built around.

Journal prompt. Pick one of this module's three domains (direct lending, infrastructure valuation, or private-company adjustments) and find one real (not synthetic) example in the wild this week: a direct-lending fund's fact sheet, an InvIT's distribution announcement, or a news item about a family business sale or an estate dispute. Write three sentences: what structural feature from this module can you now name in it that you could not have named a week ago; what number would you want to independently verify before trusting anything written about it, and against which named primary source; and (honestly) would you have been more likely to accept an AI-generated summary of that same example uncritically before this module, and why has that changed?

This module's flashcards and mastery quiz are wired into the app: see the node's Quiz and Reviews.