Regulatory and filing pipelines as forward revenue

For most companies, next year's revenue is a forecast — a line you extrapolate from this year's and then argue about. For a whole set of Indian sectors it is something firmer: an entry already sitting in a public regulator's database, put there years ago, waiting to become revenue on a schedule you can partly read. A generic drug maker's US sales three years out are latent in the applications it has already filed with the American regulator. A power generator's revenue for the next two decades is written into power contracts already signed. A road developer's collections are fixed in a concession it has already been awarded. A telecom operator's capacity to earn is bounded by spectrum it has already bought at auction.

This module is about reading those pipelines as what they are: forward revenue that lives outside the accounts, in a public register, leading the reported number by years. The question is not whether the pipeline exists — it is documented and dated — but how to turn a list of pending filings and signed contracts into a sober picture of future revenue without mistaking the list for the money.

Part Eight has been building one idea: the observable commitment upstream of revenue moves before the financial statements do. Module 093 sorted those commitments into five classes; this module is about two of them — the regulatory pipeline and the contract book — in the sectors where a licence, an approval or an award is a legal precondition for earning at all. Where the state stands between a company and its customer, the state keeps a register, and that register is the forward book.

Two facts make the pipeline readable. First, it is public by law. Regulators publish what they are processing and what they have granted: pending and approved drug applications, auctioned spectrum and its holders, awarded concessions and their terms, licences issued. The forward book is not an investor-relations claim you must take on trust — it is a government record you can audit against the company's own telling.

Second, it leads by a knowable distance. Between a filing and its revenue sits a gate the company does not control and a lag it cannot compress: the regulator's review, the plant inspection, the construction period, the launch. A drug application takes years to clear; a road concession takes years to build before it tolls. That lag is the reason the pipeline is legible ahead of time — and the reason a company can look becalmed in its accounts while its future is filling up in a database, or look busy while its register has quietly emptied.

The pipeline becomes an estimate through four multiplications, each of which strips optimism out of the raw count:

• Count. How many items are actually in the register — filings pending, contracts signed, concessions awarded. This is the one number management most likes to quote, and the least informative on its own.
• Probability of clearing the gate. Not every filing is approved, not every awarded project reaches financial close, not every signed contract survives to delivery. Each item has an approval or completion probability, and it is rarely uniform — a first-to-file application, a project with land in hand, a contract with a creditworthy counterparty all clear more surely than their opposites.
• Timing. When the revenue arrives matters as much as whether it does. Approvals stagger across years; construction has a commissioning date; contracts have a start. A pipeline that all lands in year three is worth less, and is riskier, than the same pipeline spread evenly.
• Expected value per item once it lands. What each approval is actually worth after the erosion the sector imposes — for generics, the price that falls as competitors enter; for a contract, the tariff net of costs over its tenor. The value at launch is almost never the value implied by today's price.

Multiply them and a register of pending items becomes a probability-weighted revenue curve: mostly nil today, rising in later years as items convert and launch. The leading indicator is not the count; it is this weighted, dated expectation.

Judge the pipeline by the weighting, not by whether any single item hits — a strong process will see some high-probability filings rejected and some long shots clear, and the discipline is in the expected value across the whole book, not in scoring the one result you can see.

The arithmetic is deliberately plain. Take a composite generics maker's US pipeline — five pending filings, each read on its own row rather than as a lump. All figures below are composites, marked [illustrative].

Three things fall out of the table that the headline count hides. The gross potential (₹520 cr) overstates the expectation nearly threefold, because probability and price erosion both bite. A single filing — the first-to-file, low-competition A — carries most of the weighted value, so the pipeline's worth is concentrated, not spread across the count. And the revenue is dated: it arrives across three years, so a model that books it all at once is wrong on both size and timing. The count said five; the honest read is "about ₹190 cr, front-loaded on one product, spread over three years, contingent on a plant." [illustrative]

The method is constant; the register changes. Which public pipeline is the forward book differs by sector — and pointing at the wrong one leaves you reading a company blind to the very database that determines its future. For a generics maker it is the stack of US filings; for a generator, the signed power contracts; for a road or transmission developer, the awarded concessions; for a telecom operator, the spectrum and licences it holds; for a device or new-drug maker, the specific product approvals in train. Same arithmetic — count, probability, timing, value — read off a different government record.

The inversion is the point of the section. In power and infrastructure the regulatory pipeline is contracted revenue — a PPA or a concession fixes what will be paid. In telecom the equivalent register, spectrum and licences, is the opposite: a cost incurred and a capacity ceiling, not a revenue stream. Read a telecom's spectrum holding as if it were a power company's PPA book and you will count a boundary as a booking.

Take a composite mid-cap generics maker, Kaveri Formulations illustrative, whose US business looks flat in the accounts — revenue up 4%, nothing to excite. [illustrative] The pipeline tells a different story, and you can read most of it without the company's help. The regulator's public register shows the company's pending applications and their filing dates; a third-party database shows how many rivals are filing on the same molecules; the Orange Book shows which reference products lose exclusivity and when. From these public sources alone you can rebuild the table above: how many filings, how contested each is, roughly when each clears, and what price competition it will meet.

Do that and Kaveri's flat accounts sit on top of a filling pipeline — a first-to-file application on a large molecule due to clear next year, several me-too filings behind it worth little, and one complex product two years out that would matter if its plant clears inspection. The pending-approval pipeline leads the reported US line by two to three years, so the flat current number is the past pipeline arriving, not a verdict on the future one. An investor reading only the accounts sees a stalled business; one reading the register sees where the next three years are already forming — and sees, too, that the whole thing hangs on one plant inspection, a single binary the P&L cannot show.

The register tells you what has been filed and granted; it cannot tell you four things that decide whether the forward revenue is real.

It cannot give you the probability with certainty. Approval and completion odds are estimates — informed by first-to-file status, competitor count, plant history, counterparty strength — but estimates, and a pipeline weighted with flattering probabilities is optimism in a spreadsheet's clothing. The honesty of the read is only as good as the honesty of the weights.

It cannot fix the timing. Regulators run late, inspections slip, construction overruns, financial close drags. A pipeline whose value depends on a specific year is exposed to a delay that changes nothing about the eventual revenue but everything about when it arrives and what it is worth today.

It cannot price the erosion. An approval is worth what the product earns after the competition an approval invites — and for generics that price falls year after year, so a launch two years out meets a lower price than the one you would model from today. The value at the gate is not the value on the register.

And it cannot see the binary that sits above the whole book. A plant warning letter, a licence cancellation, a policy reversal, a counterparty default — a single upstream event can freeze or void an entire probability-weighted pipeline at once. The line-by-line weighting assumes the site can supply, the licence holds, the state keeps its word; when that assumption breaks, the careful arithmetic beneath it goes with it.

The first trap is counting the filings as revenue. "Forty ANDAs" or "a ₹1,200 cr order pipeline" is a headline built from the raw count and gross potential — the two numbers before probability, timing and erosion have taken their cut. Management quotes the count because it is the largest true-sounding number available; the weighted expectation is a fraction of it, and the gap is where the investor is fooled.

The second is treating public as priced-in. Because the register is public, it is tempting to assume the market has already read it. It usually has not — the filings sit in a database most investors never open, and the weighting is real work most never do. The count is the headline everyone repeats; the probability-and-timing-weighted expectation is the signal almost no one computes, and that is precisely why the pipeline still repays reading.

The third is ignoring the binary above the book. An investor can weight forty filings beautifully and still be wrong if the plant that makes them is under a regulator's cloud, or the counterparty on the PPA cannot pay, or the concession's land is not in hand. The upstream gate voids the downstream arithmetic; a pipeline read without its enabling condition is a sum that assumes its own denominator.

The fourth is matching the wrong register to the sector — reading a telecom's spectrum as a booked revenue stream, or an infrastructure firm's revenue off its P&L rather than its concession awards. The forward book lives in a different government record for each sector, and pointing at the wrong one means reading the company blind to the database that actually governs its future.