Skip to content

alirezarezvani/claude-skills / finance/skills/stock-analysis

stock-analysis

Unverified

Analyze, research or value a listed stock or company (US, global, NSE or BSE) with verified data, ratios, scores and valuation. Use whenever asked if a stock is a buy, what it is worth, or how a company looks.

FileKindSize
SKILL.mdskill md32.7 KB
evals/evals.jsonconfig13.2 KB
evals/fixtures/synthetic-ar-excerpt.mdmarkdown5.7 KB
evals/fixtures/synthetic-drhp-excerpt.mdmarkdown3.4 KB
examples/README.mdmarkdown1.9 KB
examples/forensic-analysis-example.mdmarkdown8.8 KB
examples/standard-analysis-example.mdmarkdown15.6 KB
references/01-data-sourcing.mdmarkdown38.5 KB
references/02-core-factors.mdmarkdown49.3 KB
references/03-earnings-quality.mdmarkdown40.1 KB
references/04-balance-sheet-and-cashflow.mdmarkdown56.6 KB
references/05-returns-and-dupont.mdmarkdown49.8 KB
references/06-valuation.mdmarkdown53.7 KB
references/07-forensic-red-flags.mdmarkdown57.5 KB
references/08-governance.mdmarkdown67.0 KB
references/09-risk-and-macro.mdmarkdown52.7 KB
references/10-peer-set.mdmarkdown45.5 KB
references/11-scoring-rubric.mdmarkdown36.8 KB
references/12-report-template.mdmarkdown26.6 KB
references/13-situations.mdmarkdown60.6 KB
references/14-accounting-comparability.mdmarkdown48.6 KB
references/15-document-diligence.mdmarkdown76.1 KB
references/16-market-mechanics-and-tax.mdmarkdown48.6 KB
references/17-process-and-epistemics.mdmarkdown59.6 KB
references/18-forensic-mode.mdmarkdown16.5 KB
references/19-ipo-mode.mdmarkdown15.6 KB
references/20-challenge-pass.mdmarkdown14.1 KB
references/21-data-integrity-tools.mdmarkdown7.2 KB
references/sectors/_index.mdmarkdown67.8 KB
references/sectors/auto.mdmarkdown47.5 KB
references/sectors/aviation-hotels.mdmarkdown49.3 KB
references/sectors/banks.mdmarkdown38.6 KB
references/sectors/biotech-clinical.mdmarkdown47.8 KB
references/sectors/chemicals-cement.mdmarkdown46.2 KB
references/sectors/exchanges-payments.mdmarkdown41.2 KB
references/sectors/fmcg-consumer.mdmarkdown35.7 KB
references/sectors/holdco-assetmgr.mdmarkdown41.4 KB
references/sectors/infra-capitalgoods.mdmarkdown43.0 KB
references/sectors/insurance-brokers-services.mdmarkdown49.9 KB
references/sectors/insurance.mdmarkdown38.0 KB
references/sectors/it-saas.mdmarkdown44.3 KB
references/sectors/metals-mining.mdmarkdown49.4 KB
references/sectors/mortgage-reit-specialty-finance.mdmarkdown43.5 KB
references/sectors/nbfc.mdmarkdown40.3 KB
references/sectors/oil-gas.mdmarkdown43.5 KB
references/sectors/people-businesses.mdmarkdown46.3 KB
references/sectors/pharma-healthcare.mdmarkdown40.6 KB
references/sectors/rail-freight.mdmarkdown66.2 KB
references/sectors/realestate-reit.mdmarkdown44.3 KB
references/sectors/retail-ecommerce.mdmarkdown48.3 KB
references/sectors/semiconductors.mdmarkdown43.2 KB
references/sectors/shipping-logistics.mdmarkdown47.2 KB
references/sectors/telecom-media.mdmarkdown43.1 KB
references/sectors/utilities-power.mdmarkdown44.9 KB
references/sectors/waste-environmental.mdmarkdown46.5 KB
scripts/benchmarks.jsonconfig200.5 KB
scripts/lint_report.pyscript47.8 KB
scripts/ratios.pyscript127.4 KB
scripts/score.pyscript96.8 KB
scripts/valuation.pyscript37.9 KB
scripts/verify_data.pyscript72.5 KB

references/sectors/biotech-clinical.md

View on GitHub
# Clinical-stage biotech and pre-revenue drug developers — sector playbook

Use this when: the company has no approved product and no product revenue — a pre-clinical or clinical-stage drug developer, a platform company monetising only through partnerships, a de-SPAC'd or reverse-merged research entity, or a demerged innovator R&D arm carved out of a profitable pharma parent. Any recognised revenue is collaboration, grant or milestone income, not product sales.

The entire market capitalisation is a probability-weighted claim on a regulatory approval that has not happened. There is nothing on the income statement to value, nothing on the balance sheet except cash and lab equipment, and the assets that matter — trial data, patent life, regulatory relationships, a chemistry platform — are expensed as incurred and never appear. Two facts govern everything below: **the operating loss is a plan, not a failure**, and **the distribution of outcomes is bimodal, so a single expected-value number describes a share price that will never be observed.** If the company has product revenue, or is a biosimilar, generics, API, CDMO or CRO business, stop and use `references/sectors/pharma-healthcare.md` instead — the economics are unrelated.

## Contents

- [Why the generic ratio set fails here](#why-the-generic-ratio-set-fails-here)
- [The metrics that actually matter](#the-metrics-that-actually-matter)
- [How to value companies in this sector](#how-to-value-companies-in-this-sector)
- [Peer set construction](#peer-set-construction)
- [Sector-specific red flags](#sector-specific-red-flags)
- [Cycle and structural context](#cycle-and-structural-context)
- [India vs global notes](#india-vs-global-notes)
- [Checklist](#checklist)

---

## Why the generic ratio set fails here

Do not compute these. If a screener returns them, state in the output that they are undefined for a pre-revenue developer rather than printing them with a caveat — a printed number gets ranked, and ranking a biotech on any of them produces an inverted conclusion.

**OPM, gross margin, EBITDA margin, net margin — undefined by division.** There is no revenue. Where a small collaboration or grant line exists, the denominator is an accounting artefact of one contract's recognition schedule, and margins computed on it swing by hundreds of percent between quarters with no change in the business. A company that signs a partnership prints a "profitable quarter"; that is deal timing, not economics.

**P/E, PEG, EV/EBIT, EV/EBITDA — undefined and, worse, wrong-signed.** Earnings are negative by design and are expected to stay negative for years. A negative-denominator multiple is not a small number, it is a meaningless one. EV/EBITDA is doubly broken: EBITDA is negative *and* EV is frequently distorted or negative because cash exceeds market capitalisation.

**P/S and EV/Sales — undefined.** No sales. Applying them to collaboration revenue values the company on its licensing calendar.

**ROCE, ROIC, ROE, asset turnover — negative by construction and inverted in direction.** All are negative, and all *improve* when the company shuts down research. A developer that halves its R&D budget shows a less negative ROE and a screener reads it as improving quality; in reality it has just stopped building the only asset it has. Capital employed is a cash pile awaiting deployment, not capital at work.

**Operating cash flow and free cash flow — negative by design, and a plan rather than a result.** The correct question is never "why is OCF negative" but "is the burn being spent on the assets that create option value, at the rate management guided, and does the cash on hand reach past the next value-inflecting readout". Positive FCF at a clinical-stage company almost always means trials have been paused. FCF yield is meaningless; a DCF of enterprise free cash flow cannot be built, because there is no cash flow until an approval that may never occur.

**Book value, P/B and net asset backing — technically computable, economically empty.** Book value is overwhelmingly cash and marketable securities. P/B is therefore a rearranged statement of market-cap-to-cash: it measures how much premium the market pays for the pipeline, which is an *input* to the analysis, not a valuation. It says nothing about intrinsic worth, and it collapses mechanically as cash burns whether or not the science improved.

**D/E, net debt/EBITDA and interest coverage — mis-signed.** Most clinical-stage companies are net cash, so leverage ratios screen as pristine while the company is three quarters from insolvency. The binding solvency constraint is runway against catalyst timing, not gearing. Conversely, where debt does exist it is usually a venture-debt facility with covenants tied to trial milestones, or a convertible whose real risk is dilution, not default — neither is captured by D/E.

**Current ratio, quick ratio, working capital, inventory and receivable days, cash conversion cycle — uninformative or undefined.** There is no inventory and no trade receivable. Current ratios of 5–15x are routine and tell you only that the cash has not yet been spent; a company can carry a current ratio of 8x and nine months of runway simultaneously.

**Revenue and expense "growth" — not growth.** Rising R&D expense is the cost of running more or larger trials. It is a commitment, not an achievement, and treating an expanding expense base as a growth rate is the single most common misreading of these filings. The corresponding real growth metrics are pipeline stage advancement and risk-adjusted value created per rupee or dollar burned.

**Dividend yield, payout, buyback capacity — structurally zero and correctly so.** Any capital returned by a pre-revenue developer is capital not spent on the pipeline.

**DCF with a terminal growth rate — structurally invalid.** Drug cash flows end at loss of exclusivity. A perpetuity assumption on a patented molecule capitalises revenue that legally cannot exist. Model to the exclusivity cliff and let revenue collapse; never bolt on a terminal value.

---

## The metrics that actually matter

All ranges below are **indicative only**. They shift with the funding cycle, therapeutic area, modality, listing venue and period, and any of them can be wrong for a specific company in a specific year. A tight peer set and the company's own disclosure history override every absolute band printed here.

| Metric | Definition / how to compute | Indicative healthy range | Why it matters |
|---|---|---|---|
| **Cash runway in quarters, and the funding cliff** | (Cash + equivalents + short-term marketable securities + committed, non-contingent milestone receipts − debt due within the runway window) ÷ average quarterly net cash used in operating activities *plus* capex, using the last two to four quarters and adjusting for known step-ups (a Phase III start can double burn). Express in quarters and mark the calendar quarter cash runs out. | >8 quarters comfortable; 6–8 adequate; 4–6 means a raise is being prepared now; <4 means the terms will be set by whoever is willing to fund. Guidance must fund **past** the next value-inflecting readout with two to three quarters to spare, not "into" it. | This replaces every solvency ratio in the generic set. The funding cliff, not the science, determines who captures the value of a good result: a company forced to finance before a readout raises at the market's price and hands the upside to the new money. Runway also sets negotiating leverage in any licensing deal — a partner reads the cash line before it reads the data. Language matters: "into the second half of next year" is a cliff, "through the topline readout and twelve months beyond" is a plan. |
| **Quarterly net burn and its trajectory** | Net cash used in operations + capex, per quarter, for 8–12 quarters. Decompose the trend: rising because trials enrolled and scaled (good), rising because headcount and facilities grew without programmes advancing (bad), falling because a trial completed (neutral), falling because programmes were paused (a hidden pipeline cut). | Burn trajectory that maps cleanly onto disclosed trial starts, enrolment and completions. A step-change in either direction with no corresponding pipeline event is the flag. | Burn is the sector's only real income-statement signal, and its *shape* is the tell. Trials cost money in a known pattern — start-up costs, then per-patient costs scaling with enrolment, then a tail. Burn that does not match the stated trial calendar means either enrolment is behind guidance or spend has quietly moved somewhere the deck does not mention. |
| **R&D as share of burn, and R&D by programme** | R&D expense ÷ total operating expense; and, where disclosed, R&D split by programme and by external (CRO, CDMO, trial site) vs internal (personnel, facilities). Track G&A per employee and G&A as % of burn separately. | R&D 65–80% of opex is typical for a company genuinely running trials; sustained below ~55–60% means the cost base is corporate rather than scientific. G&A rising faster than R&D for more than a year is a governance signal. | This is the quality-of-burn metric. Two companies with identical runway are not equivalent if one spends three-quarters of it on trials and the other on public-company overhead, investor relations and executive compensation. Programme-level R&D disclosure, where given, is also the honest map of management's real priorities — an asset featured on slide three of the deck but carrying almost no spend is being deprioritised. |
| **Pipeline map — asset × phase × indication × modality** | A table with one row per asset: programme code, target and mechanism, modality (small molecule, mAb, bispecific, ADC, peptide, ASO/siRNA, cell therapy, gene therapy, vaccine, radioligand), indication and line of therapy, current phase, trial identifier, enrolment status, next milestone and date, ownership (wholly owned vs partnered vs in-licensed), and the royalty or milestone burden owed to any licensor. | Concentration is the variable, not a level. A lead asset carrying >70–80% of risk-adjusted value makes the company a binary instrument regardless of how many preclinical rows sit beneath it. | This is the balance sheet, and none of it is capitalised. Two companies with identical financials can differ tenfold in value on phase mix, mechanism validation and catalyst proximity alone. Build this table from the trial registries (ClinicalTrials.gov, EU CTIS, CTRI in India) and the latest corporate deck side by side — assets that appear in one and not the other are the most informative rows in the file. |
| **Probability of success (PoS) by phase — industry base rates** | Assign each asset a cumulative probability of reaching approval from its current phase, then adjust for mechanism validation, effect size seen so far, trial design quality and competitor read-across. | Widely cited industry base rates (Phase-transition studies covering roughly 2011–2020): Phase I→II ~50–55%; **Phase II→III ~28–32%** (the real filter); Phase III→filing ~55–60%; filing→approval ~88–92%; cumulative Phase I→approval ~7–10%. By area, cumulative Phase I→approval has run near ~5% in oncology, ~6% in CNS, ~14% in infectious disease, ~17% in rare disease and ~20–25% in haematology. | These are **base rates, not estimates for this asset**, and they vary very widely by indication, modality, line of therapy and sponsor experience. Use them as the prior you must argue away from, and state explicitly in the output what evidence justifies any deviation. Anchoring on the sponsor's own confidence instead of the base rate is how the sector is systematically over-valued. Note the asymmetry: filing-to-approval is high, so the binary risk sits overwhelmingly in Phase II and Phase III readouts. |
| **Risk-adjusted NPV (rNPV) per asset** | For each programme: bottom-up peak sales × PoS-weighted revenue ramp, net of gross-to-net, less cost of goods, sales and marketing and residual development spend (each cost line weighted by the probability it is actually incurred), taxed, discounted at 10–14% (15–20% for pre-clinical, single-asset or first-in-class), with the exclusivity cliff modelled explicitly and revenue collapsing after it. Deduct royalties and milestones owed to licensors. **No terminal value.** | Not a range — a construction standard. Every rNPV must state its four load-bearing assumptions: peak sales, PoS, launch year and discount rate. | This is the core method and everything else is a cross-check. The discipline of building it forces the analysis to be explicit about what is being assumed, which is precisely what a multiple hides. Run sensitivity on all four inputs together, not one at a time — peak sales and PoS are correlated (a bigger claimed market usually means a harder trial and more competition), so a one-at-a-time sensitivity understates the true range. |
| **Catalyst calendar** | Every dated event over the next 24 months: interim analyses and DSMB reviews, topline readouts, conference presentations (ASCO/ESMO/AHA/AAN and equivalents), IND/CTA clearances, End-of-Phase-2 meetings, NDA/BLA/MAA submissions, filing acceptances, PDUFA dates, advisory committee meetings, CHMP opinions, partnering deadlines and option-exercise windows. Mark each as company-controlled or externally timed. | At least one value-inflecting catalyst inside the funded runway. A 24-month window with nothing but "data expected" and no dated events is a warning in itself. | Value in this sector is realised in discrete jumps on known dates, so position sizing, entry timing and the financing question are all functions of this calendar. Note the standard clocks: US PDUFA goals run ~10 months from filing acceptance for standard review and ~6 months for priority review (add ~2 months for the filing decision); a CHMP positive opinion is followed by a European Commission decision roughly two months later. A slipping catalyst is data, not scheduling. |
| **Trial design quality** | For each pivotal or value-defining trial, score: randomised vs single-arm; blinded vs open-label; active comparator vs placebo vs external/synthetic control; primary endpoint (hard clinical outcome vs validated surrogate vs unvalidated surrogate vs biomarker); powering assumptions — the effect size the trial is powered to detect vs the effect size actually seen earlier; sample size; alpha allocation and multiplicity handling across secondary endpoints and interim looks; analysis population (ITT vs mITT vs per-protocol) and handling of missing data; enrolment velocity (patients per activated site per month) vs guidance. | Randomised, double-blind, adequately powered against an active or placebo control on a regulator-accepted endpoint is the reference standard. Anything less requires an explicit haircut to PoS. | This is where PoS is actually determined, and it is knowable *before* the readout. An open-label single-arm trial cannot separate drug effect from patient selection, natural history and placebo response — it can support accelerated approval in a high-unmet-need setting but converts to a confirmatory-trial liability later. An unvalidated surrogate endpoint risks a regulator accepting the statistics and rejecting the clinical meaning. And a Phase III powered on the point estimate from a small Phase II ignores regression to the mean, which is the most common mechanical cause of a failed pivotal trial. |
| **Addressable population and realistic penetration** | Build the funnel bottom-up: incidence or prevalence in each target geography → diagnosed → tested/biomarker-eligible → meeting the trial's own inclusion criteria → treated with any drug → reachable by the launch salesforce or partner → share captured given order of entry → duration of therapy and adherence → doses per year × net price. | Rare-disease funnels lose most of their patients at "diagnosed"; oncology funnels lose them at biomarker eligibility and line of therapy. Peak share for a third entrant with no differentiation is usually a fraction of what a corporate deck assumes. | Peak sales is the input rNPV is most sensitive to and the one most casually asserted. "A $10 billion market" is almost never the addressable population — it is the prevalence pool before every attrition step. The single most valuable check is to re-derive peak sales from epidemiology and compare it to the number in the deck; a gap of 3–5x is common and it flows straight through to valuation. |
| **Net price and reimbursement exposure** | Assumed list price benchmarked to approved comparators in the same indication; gross-to-net deduction; payer mix by geography; whether the drug is a physician-administered benefit or a pharmacy benefit; HTA exposure (NICE, G-BA/IQWiG, HAS, PBAC, CADTH) and whether the trial generated the comparative evidence those bodies require; US Medicare negotiation exposure under the IRA. | US gross-to-net commonly 30–55% for a specialty launch. Europe realises materially below US net price. A launch price assumption above the nearest approved comparator needs an evidence-based justification, not a slide. | An approval that cannot be reimbursed is not a commercial product. HTA bodies ask a different question from regulators — not "does it work" but "does it work better than what we already pay for, per unit of cost" — and a single-arm trial against no comparator answers the first and not the second. The IRA also creates a modality asymmetry: small molecules become eligible for US Medicare price setting materially earlier in their commercial life than biologics, which shortens the effective revenue window in an rNPV and has already shifted where capital flows. |
| **Competitive landscape and order of entry** | Every competing asset in the same indication and mechanism, by phase and sponsor, from the registries — not from the company deck. Estimate probable order of market entry, and identify who reads out before you. | Being third or later into a class with no differentiation on efficacy, safety, dosing or route typically caps realistic share in the low tens of percent at best. | Order of entry drives peak share, price and the speed of the launch ramp, and it also creates read-across risk: a competitor's failure on the same target can destroy value here without a single event at the company itself, and a competitor's success can compress the commercial opportunity while validating the mechanism. Both directions must be in the scenario set. |
| **Partnering economics and retained rights** | For each deal: upfront cash; equity component and the premium paid; development, regulatory and commercial milestones separated (near-term and probable vs distant and improbable); royalty rate and tier structure and the sales base it applies to; territory split; which rights were retained; cost-sharing vs full funding; opt-out, reversion and change-of-control clauses; who books sales. | Judge the **upfront as a share of headline "biobucks"** — a low ratio means the partner bought an option cheaply. Retained US or major-market rights signal genuine internal conviction and are usually where the value is. | Headline total deal value is a marketing number that assumes every milestone is achieved; it should never enter a valuation. A partnership with a credible large-pharma counterparty is the strongest third-party validation available for a platform or a mechanism, because the partner ran its own diligence. The inverse is the single most informative negative signal in the sector: **a partner returning rights** means someone with better information walked away. |
| **Dilution history and instrument structure** | Shares outstanding (basic and fully diluted) each year for 5–10 years, with every raise, its price and the discount to market; plus the current overhang — ATM programme capacity and utilisation, warrants (count, strike, expiry, any cashless or ratchet terms), convertible notes and their conversion mechanics, pre-funded warrants, PIPEs, and royalty/revenue-interest financings. | Share count growth that is clearly less than the risk-adjusted value added by pipeline advancement over the same period. A structure with reset-priced or floor-less convertibles is a category risk, not a pricing detail. | Per-share value is what an equity holder owns, and in this sector the share count is the most reliably growing number in the filing. Assess every raise against what it bought: three programmes advanced a phase on a 40% share-count increase is value creation; a doubled share count with the same pipeline is value transfer. Convertibles whose conversion price resets downward as the stock falls create a self-reinforcing decline and have destroyed more small-cap biotech equity than failed trials. Also always model **expected future dilution** into the valuation — the raises required to fund the plan are as certain as the burn. |
| **EV/cash, implied pipeline value, and implied PoS** | EV = market cap + debt − cash and marketable securities. Compute EV ÷ cash; the implied pipeline value (EV itself, plus PV of unallocated G&A); and the reverse test — implied lead-asset PoS = (market cap − net cash − rNPV of other assets) ÷ unrisked NPV of the lead asset. | Negative EV means the market ascribes less than zero to the pipeline — worth investigating, but it is usually a judgement on burn and governance, not a free option. | The reverse test is the most disciplined valuation cross-check available here, because it converts an opinion about price into a falsifiable statement about probability. If today's market capitalisation implies a 65% probability of success for an asset entering Phase III, compare that to the ~55–60% base rate and to this trial's design quality, and the investment question becomes concrete. Run it in both directions: also compute the share price implied by success and by failure, since the outcome distribution is bimodal. |
| **Regulatory designations and the exclusivity stack** | List every designation held and what it actually confers: Fast Track (more frequent FDA meetings, rolling review eligibility), Breakthrough Therapy (intensive guidance and senior FDA engagement), RMAT for regenerative medicines, Orphan Drug (US market exclusivity of seven years on approval plus fee waivers and clinical trial tax credits; EU ten years), Priority Review, Accelerated Approval (surrogate endpoint plus a mandatory confirmatory trial), Priority Review Vouchers where the statutory programme is currently authorised. Separately map the exclusivity stack: patent expiries by family, regulatory data exclusivity (US five years for a new chemical entity, twelve years for biologics; EU eight plus two, with a possible extra year), paediatric extensions. | Designations should be read as time-saving and access-improving, never as evidence of efficacy. | Companies market designations as quasi-approvals and retail narratives treat them that way. **None of them lowers the evidentiary bar for approval** — Breakthrough and Fast Track change the process, not the standard, and a meaningful share of Breakthrough-designated programmes still fail. Accelerated Approval is the one with real economic teeth in both directions: it can bring revenue years earlier on a surrogate endpoint, and a failed or unfinished confirmatory trial can withdraw it. The exclusivity stack, meanwhile, sets the length of the revenue window in the rNPV and is often the second-largest value driver after peak sales. |
| **Ownership and management's prior clinical track record** | Insider ownership and recent transactions (Form 4 in the US; SEBI PIT continual disclosures and quarterly shareholding pattern in India); holdings and changes among specialist healthcare funds and crossover investors (13F/13D-G); free float and index membership. For management and the scientific board: prior INDs filed, trials run, approvals actually obtained, and companies previously led — including how those ended. | Meaningful insider ownership with purchases rather than sales; a stable roster of specialist biotech funds rather than only generalist or retail holders. | Specialist funds employ scientific staff and run their own diligence; their entry and exit is the closest thing to informed third-party opinion available in the public filings, and a broad specialist exit ahead of a readout is worth more attention than any sell-side note. On management: drug development is a craft, and a team that has taken a molecule from IND to approval before executes trials, regulatory interactions and manufacturing scale-up materially better. A chief medical officer or head of research departing shortly before a pivotal readout is one of the highest-signal events in the sector. |

---

## How to value companies in this sector

The task is to value a portfolio of options on future cash flows that do not yet exist. Only methods that make probability, timing and the exclusivity cliff explicit are usable.

**Primary — rNPV, built asset by asset, then summed.** For every programme in the pipeline:

> Asset rNPV = Σ<sub>t</sub> [ PoS × (peak-sales-derived net revenue<sub>t</sub> − COGS<sub>t</sub> − SG&A<sub>t</sub> − royalties owed<sub>t</sub>) × (1 − tax) ] / (1+r)<sup>t</sup> − Σ<sub>t</sub> [ P(reaching that stage) × remaining development cost<sub>t</sub> ] / (1+r)<sup>t</sup>

Then: **Company value = Σ asset rNPV + net cash + PV of committed non-contingent milestone receipts − PV of unallocated G&A over the development period − expected financing dilution cost.** Model revenue only to loss of exclusivity and let it collapse; no perpetuity, no terminal growth. Risk-adjust the cost lines at their own probability of being incurred, not at the approval probability — a Phase III that will only run if Phase II succeeds should be weighted at the Phase II success rate. Discount at 10–14% for a diversified clinical-stage pipeline, higher for single-asset, first-in-class or pre-clinical situations; do not additionally haircut the discount rate for scientific risk, which is already in the PoS (double-counting risk is the most common technical error in these models).

**Sum-of-the-pipeline plus cash is the reporting format.** Present a table of assets with PoS, unrisked NPV, rNPV and per-share contribution, then net cash and central costs as separate lines. This makes the valuation auditable and shows immediately how concentrated the value is — which is usually the most important output.

**Comparable transaction multiples per asset.** Benchmark against precedent licensing and M&A deals for assets at the same phase, in the same indication and modality: upfront payment per Phase I / Phase II / Phase III asset, total acquisition value per programme, deal value as a multiple of risk-adjusted peak sales, and royalty rates struck for comparable mechanisms. Deal comparables are the market's revealed price for exactly this kind of risk and are far more relevant than any equity multiple. Adjust for territory, retained rights and whether the acquirer was strategically motivated by a specific patent cliff.

**Reverse valuation — solve for the implied probability.** Strip net cash from market capitalisation, subtract the rNPV of secondary assets, and divide the residual by the unrisked NPV of the lead asset. The result is the probability of success the market is paying for. Compare it to the phase base rate and to the trial's design quality, and state the comparison explicitly in the output. This converts "expensive or cheap" into a testable claim and is the single most useful discipline in the sector.

**Scenario valuation, because the distribution is bimodal.** Compute the share price in the success case (rNPV at PoS = 1, less remaining dilution) and in the failure case (usually net cash per share, less continuing burn and any restructuring cost, plus residual value of the remaining pipeline). An expected value between the two describes a price the stock will never trade at after the readout. Report the pair and the probability weight, not just the mean.

**Real options / decision-tree framing where staging matters.** A pipeline is a sequence of abandonment options: management pays for the next stage only if the current one succeeds. A decision tree over phase transitions, or a simple binomial option framing on a single asset, captures the value of the right to abandon that a static rNPV understates. Useful mainly for high-volatility, early-stage or platform situations; do not let it become a black box.

**Secondary cross-checks.** EV/cash and negative-EV screening; EV per clinical-stage programme; price to risk-adjusted peak sales; the cost and time it would take a competitor to replicate the pipeline from scratch; and, for platform companies, the PV of the partnership economics already signed as a floor on platform value.

**Do not use, in any form:** P/E, forward P/E, PEG, EV/EBITDA, EV/EBIT, EV/Sales, P/S, P/B as a valuation anchor, ROCE or ROIC screens, dividend yield, FCF yield, or a DCF with a terminal growth rate. Each is either undefined or actively inverted here, and reporting any of them — even with a caveat — invites a comparison that cannot be made.

---

## Peer set construction

A valid comparable shares **phase, modality, therapeutic area and funding position**. "Biotech" is not a peer set; it is a listing category containing companies whose risk profiles differ by an order of magnitude.

Splits that must never be mixed:

- **Clinical-stage vs commercial-stage.** The moment there is product revenue, the entire framework changes and the company belongs in `pharma-healthcare.md`. A pre-revenue developer and a newly commercial one are different instruments.
- **Phase band.** Pre-clinical/Phase I, Phase II, and Phase III/filed are three distinct risk classes with roughly 8%, 15–30% and 50–60% cumulative success respectively. Comparing EV per programme across them is meaningless.
- **Modality.** Small molecule, monoclonal antibody, bispecific, ADC, peptide, oligonucleotide, cell therapy, gene therapy, vaccine and radioligand differ in development cost, manufacturing complexity, COGS at launch, regulatory pathway, exclusivity length and IRA treatment. Cell and gene therapy in particular carry manufacturing and durability risks that no antibody peer shares.
- **Therapeutic area.** Oncology, CNS, rare disease, immunology, metabolic and infectious disease have materially different base rates, trial sizes, placebo-response problems and payer dynamics. A CNS asset and a haematology asset at the same phase are not equally risky.
- **Single-asset vs platform.** A single-asset company is a binary event with a date on it. A platform company is a portfolio whose value depends on whether the platform itself is validated — ideally by a partnership someone else paid for. Never blend them in one multiple.
- **Partnered vs wholly owned.** A partnered asset carries less financing risk and less upside; the royalty and milestone structure changes the economics so much that the same molecule is worth very different amounts in the two structures.
- **Funding position.** A company with eight quarters of runway and one with three are priced on different things — the second is priced on financing terms, not science. Do not read the second as "cheap" relative to the first.
- **Origin of the asset.** Internally discovered vs in-licensed. In-licensed assets carry upstream royalty and milestone obligations that reduce the economics materially and are easy to miss.
- **Listing venue and shareholder base.** Nasdaq/NYSE biotech, European listings, and NSE/BSE innovator entities differ in disclosure requirements, specialist-investor depth, index inclusion and the availability of follow-on financing. A US-listed peer's multiple does not transfer to an Indian-listed R&D entity or vice versa.
- **IPO'd vs reverse-merged/de-SPAC'd shells.** Different governance quality, different overhang structures, different warrant and earn-out complexity.

Also align on catalyst proximity — a company two months from a Phase III readout trades on a different basis from an identical company two years out — and on whether the valuation is stated pre- or post-money for any announced but unclosed financing. Peer sets here should be small and honest: **three to six genuine comparables beats fifteen tickers sharing a sector tag**, and where no clean peer exists, precedent licensing transactions for similar assets are a better benchmark than any listed company.

---

## Sector-specific red flags

**Data presentation and trial conduct**
- Topline press releases with adjectives and no numbers: no effect size, no confidence interval, no p-value, no control-arm result, no discontinuation rate. "Met its primary endpoint" without the magnitude is not a result.
- Statistical significance on a large sample with a clinically trivial effect size, or significance achieved only after the analysis population changed from ITT to mITT or per-protocol.
- Primary endpoint, analysis population or statistical analysis plan amended mid-trial, especially close to unblinding. Check registry version history against the protocol — the registry keeps the audit trail the press release omits.
- Subgroup rescue after a missed primary endpoint, presented as a path forward. Post-hoc subgroups are hypothesis-generating, not evidence, and regulators treat them accordingly.
- Cross-trial comparisons against historical or published control data, especially where the company's own trial is single-arm and open-label. Patient selection differences swamp drug effect.
- Unvalidated surrogate endpoints, or a surrogate that a regulator has previously declined to accept in that indication.
- A pivotal trial powered on the point estimate from a small, positive early-phase study — regression to the mean is the standard mechanism by which these fail.
- Enrolment materially behind guidance, sites activated but not recruiting, or a trial completion date repeatedly pushed in the registry without comment.
- A DSMB "recommendation to continue as planned" marketed as positive news; it is the neutral outcome.
- Undisclosed pipeline pruning: assets that quietly vanish between corporate decks. Diff consecutive presentations.
- Clinical holds, refuse-to-file letters, complete response letters, or GCP/GMP inspection findings disclosed late, minimised, or characterised as procedural.

**Financing and capital structure**
- Runway guidance that ends at or just before the next major readout — the company will be raising into its own binary event.
- Convertible instruments with reset, ratchet or floorless conversion mechanics, or an equity line drawn continuously at a discount to market. These convert a falling share price into more shares and more selling pressure.
- Large warrant overhang, particularly warrants issued alongside discounted PIPEs, and pre-funded warrants used to keep reported share counts optically low.
- An at-the-market programme being used continuously rather than opportunistically, especially into good news.
- Reverse stock splits to maintain listing compliance; going-concern language in the audit opinion; royalty monetisation or revenue-interest financing that is economically debt but not always presented as such.
- Share count growing faster than the pipeline advances, across multiple years.

**Governance and disclosure**
- Insider selling ahead of a readout, or a 10b5-1 plan adopted shortly before one (India: check SEBI PIT trading-plan and continual disclosures, and any promoter pledge).
- Departure of the chief medical officer, head of clinical development or chief scientific officer close to a pivotal readout; auditor resignation or change; repeated CFO churn.
- Management whose prior ventures ended in delisting, liquidation or serial pivots, or a scientific advisory board with no operating drug-development experience.
- Paid stock-promotion campaigns, retail-focused IR spend, or a communications cadence that outpaces the clinical cadence.
- Related-party CRO, CDMO or laboratory services arrangements; research funded through entities connected to management.
- Reduced disclosure over time — programme-level R&D withdrawn, trial timelines removed from the deck, registry entries left un-updated. In this sector, disclosure quality deteriorates before the numbers do.

**Strategy and pipeline**
- A partner returning rights, declining to exercise an option, or writing the programme down. This is the strongest available negative signal.
- Breadth substituting for depth: a dozen pre-clinical programmes, no clinical data, and a platform claim resting on animal models.
- Orphan, Breakthrough or Fast Track designation presented as a de-risking event or as evidence of efficacy.
- Expanded access, compassionate use or investigator-initiated anecdotes used as evidence of activity.
- Pivoting the lead indication to a smaller or rarer population after a setback, without acknowledging the corresponding cut to peak sales.
- Peak sales assumptions derived from a prevalence pool rather than a treated, eligible, reachable population — and a price assumption above the best approved comparator with no comparative evidence to support it.

---

## Cycle and structural context

**The dominant cycle is the funding cycle, and it is exogenous to the science.** Clinical-stage biotech is the longest-duration cash flow in public equities, so its valuation is more sensitive to the discount rate than almost any other sector. Rising rates compress the entire cohort regardless of trial outcomes; falling rates and open IPO windows re-rate it just as indiscriminately. Track the sector index (XBI/NBI or local equivalent), IPO and follow-on issuance volume, and the number of listed companies trading below net cash — a large negative-EV cohort marks a capitulation phase and is usually followed by consolidation, reverse mergers, liquidations and returns of capital rather than by a recovery in the same names.

**The funding cycle propagates with a lag.** A financing winter shows up in reduced trial starts within a year, in thinner late-stage pipelines two to three years later, and in CDMO and CRO order books with a similar delay — the cross-read into `pharma-healthcare.md` runs in both directions.

**M&A is the exit, and large-pharma patent cliffs set the demand.** A concentrated wave of loss-of-exclusivity events across the major innovators creates a structural bid for de-risked late-stage assets, particularly in oncology, immunology, neurology and metabolic disease. Takeout probability is a real component of value for a company with a differentiated Phase II or III asset, and it is highest where the acquirer's own revenue gap aligns with the asset's launch year. Do not model it as a base case, but do state it.

**Regulatory and policy shifts to underwrite explicitly:**
- **IRA Medicare price negotiation** and its small-molecule/biologic timing asymmetry — this is now a modality-selection input at the point of capital allocation, not a late-stage commercial detail.
- **Accelerated approval reform** — confirmatory trials increasingly required to be underway at the time of approval, and a faster withdrawal process. This raises the cost and lowers the optionality of the surrogate-endpoint pathway.
- **Dose-optimisation expectations in oncology**, which have made early development longer and more expensive by requiring randomised dose comparison before pivotal trials.
- **Cross-border licensing flows.** A large and growing share of in-licensed assets originating from Chinese biotechs has compressed the scarcity value of undifferentiated Western me-too programmes; check whether a company's lead asset faces a cheaper, equally advanced licensed competitor.
- **Manufacturing and supply constraints** for cell, gene and radioligand therapies, where capacity and isotope supply can gate a launch more tightly than approval does.
- **Modality waves.** Capital rotates violently between platforms — antibody-drug conjugates, radiopharmaceuticals, obesity and metabolic, cell and gene therapy, AI-enabled discovery. A company's multiple can move entirely on which wave it is filed under, which is a reason to distrust peer multiples and prefer rNPV.

---

## India vs global notes

The most important India-specific point is a routing one. **There are very few genuine pre-revenue innovator listcos on the NSE/BSE.** The Indian industry is built on generics, API, biosimilars, vaccines, CDMO and CRO — all of which have revenue and belong in `pharma-healthcare.md`. Where Indian innovator R&D exists it usually sits either inside a profitable generics parent (in which case consolidated ratios are a blend of two incompatible economic models and the R&D arm must be valued separately in an SOTP) or in a demerged, separately listed research entity (in which case this playbook governs it and the generics parent does not). Say which structure applies before computing anything. The second point is an accounting one: **Ind AS 38 permits capitalisation of development costs once technical feasibility is established, while US GAAP (ASC 730) mandates expensing** — so an Indian innovator can carry clinical spend as "intangible assets under development" and report a shallower loss than an identical US peer. Restate to a fully expensed basis before any cross-border comparison, and check the capitalised balance against actual programme progression.

| Dimension | India (NSE/BSE, Ind-AS) | US / global (SEC, GAAP/IFRS) |
|---|---|---|
| Product regulator | CDSCO / DCGI, with Subject Expert Committees; New Drugs and Clinical Trials Rules, 2019 govern approvals, ethics committees and compensation; RCGM (DBT) and GEAC for recombinant and genetically modified products; ICMR ethics guidelines | USFDA (CDER/CBER) with IND/NDA/BLA pathway and PDUFA goal dates; EMA/CHMP with the EU Clinical Trials Regulation and CTIS; MHRA, PMDA |
| Trial registry | **CTRI** — prospective registration mandatory for trials conducted in India | **ClinicalTrials.gov** and **EU CTIS**; both keep version histories — use them to detect protocol and endpoint changes |
| Primary filings | Annual report, quarterly results, investor presentation, **earnings concall transcript**, and SEBI LODR Regulation 30 material-event announcements (trial outcomes, approvals, partnerships) | 10-K / 10-Q / **8-K** for material events, S-1/S-3 for financings, 424B prospectuses; proxy (DEF 14A). Disclosure is faster, more granular and more standardised |
| R&D accounting | Ind AS 38 allows post-feasibility capitalisation — check "intangible assets under development" every year and reconcile to programme progress | ASC 730 mandates expensing; IFRS filers face the same IAS 38 option as India, so IFRS-reporting European developers need the same restatement |
| Financing instruments | QIP, preferential allotment, **promoter/investor warrants (25% upfront, balance on conversion within 18 months)**, rights issues, FCCBs. No ATM-equivalent shelf mechanism; dilution is lumpier and more visible | ATM programmes off an effective shelf, PIPEs, registered directs, pre-funded warrants, convertible notes, venture debt, royalty and revenue-interest financing |
| Ownership disclosure | Quarterly **shareholding pattern** (promoter, promoter pledge, FII, DII, public); SEBI PIT continual disclosures for designated persons and trading plans; SAST for large stakes | **Form 4** insider transactions and 10b5-1 plans; 13F quarterly institutional holdings; 13D/G for 5%+ stakes — the specialist-fund register is far more legible |
| Governance overlay | CARO reporting, auditor qualifications, SEBI LODR related-party approvals, promoter pledge levels | SOX 404 internal control opinion, critical audit matters, audit committee independence, Nasdaq/NYSE listing standards including minimum price rules |
| Domestic commercial value | **DPCO/NLEM** price control and low willingness to pay mean an India-only innovative launch rarely supports an innovator economic model; Indian developers must target US/EU markets or out-license to capture value | IRA negotiation, 340B, Medicaid rebates and commercial payer mix in the US; HTA-set prices (NICE, G-BA/IQWiG, HAS, PBAC) in Europe and elsewhere |
| Support and incentives | BIRAC and DBT grants, PRIP scheme support for innovation, Section 35(2AB) weighted R&D deduction now at 100%, state biotech park incentives — treat grant income as non-recurring, never as revenue | Orphan drug clinical trial tax credits, NIH/BARDA grants and contracts, Priority Review Vouchers where currently authorised — again, non-recurring |
| Units and reporting | Rs crore / lakh; fiscal year April–March; quarterly results. Burn is often best read from the cash flow statement's half-yearly disclosure, which is less granular than US quarterly data | USD millions; calendar year common; full quarterly cash flow statements and an explicit runway statement in the liquidity note of the 10-Q |

---

## Checklist

- [ ] Confirm there is no product revenue. If there is, route to `pharma-healthcare.md`; if the entity is an R&D arm of a commercial parent, separate the two and value them independently.
- [ ] Delete P/E, EV/EBITDA, EV/Sales, P/S, OPM, ROCE, ROE, D/E, FCF yield and current ratio from the analysis, and state in the output why they are undefined here.
- [ ] Compute cash runway in quarters, name the calendar quarter cash runs out, and compare it to the next value-inflecting catalyst date.
- [ ] Chart quarterly net burn for 8–12 quarters and check the trajectory reconciles to disclosed trial starts, enrolment and completions.
- [ ] Split burn into R&D and G&A; flag G&A growing faster than R&D, and check programme-level R&D against the deck's stated priorities.
- [ ] Build the pipeline table from the trial registry and the corporate deck side by side; investigate every asset that appears in one and not the other.
- [ ] Assign PoS from published phase base rates, state them as base rates, and justify explicitly any deviation for indication, modality or design quality.
- [ ] Build rNPV asset by asset with bottom-up peak sales, risk-adjusted costs, an explicit exclusivity cliff and no terminal value; sum, add net cash, deduct PV of central costs and expected dilution.
- [ ] Re-derive peak sales from the epidemiology funnel and compare it to management's number; report the gap.
- [ ] Score trial design: randomisation, blinding, control arm, endpoint validity, powering vs observed effect size, multiplicity, analysis population, enrolment velocity. Haircut PoS for single-arm, open-label or unvalidated-surrogate designs.
- [ ] Build the 24-month catalyst calendar with dated events, and mark which are company-controlled.
- [ ] Assess reimbursement: comparator-benchmarked net price, gross-to-net, HTA evidence requirements, payer mix and IRA modality exposure.
- [ ] Map every competing asset by phase from the registries; state probable order of entry and the read-across risk from competitor readouts.
- [ ] For every partnership, separate upfront from biobucks, and record royalty tiers, retained rights, opt-outs and reversion clauses. Treat any return of rights as a first-order negative.
- [ ] Chart 5–10 years of share count with every raise's price and discount; inventory ATM capacity, warrants, convertible mechanics and any royalty financing.
- [ ] Run the reverse test: solve for the lead asset's market-implied PoS and compare it to the phase base rate.
- [ ] Report the bimodal outcome pair — success-case and failure-case value per share — not only the probability-weighted mean.
- [ ] Cross-check against precedent licensing and M&A economics for same-phase, same-modality assets; check EV/cash for negative-EV situations.
- [ ] List regulatory designations and state precisely what each confers; confirm none has been presented as evidence of efficacy.
- [ ] Map the exclusivity stack — patents, regulatory data exclusivity, orphan exclusivity, paediatric extension — and use it to set the revenue window.
- [ ] Check insider and specialist-fund ownership changes, and management's prior record of INDs filed and approvals actually obtained.
- [ ] India: check Ind AS 38 capitalised development costs and restate to a fully expensed basis; check promoter holding, pledge and warrant issuances; read CTRI entries directly.
- [ ] Build a peer set of 3–6 matched on phase, modality, therapeutic area, ownership structure and funding position — or state that no valid peer exists and use precedent transactions instead.
- [ ] State where the sector sits in the funding cycle and what that does to the cost and availability of the company's next raise.
add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope.Stars: 115kUpdated Claude CodeCodexVerifiedInstall
fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks.Stars: 115kUpdated Claude CodeCodexVerifiedInstall