A view is only as good as the process behind it. This page states mine in public, so every analysis on this site can be judged against it — and so you know exactly what you're reading before you read it.
1. Screen. Issuers come from a defined coverage universe — India's power and renewables complex: IPPs, utilities, wind OEMs, transition financiers, and InvITs. I don't chase whatever is moving; coverage is deliberate.
2. Primary filings first. Every note starts with the annual report, quarterly filings, investor presentations, and the rating agencies' own rationales (CRISIL, ICRA, CARE, India Ratings). Secondary commentary comes last, if at all.
3. Spread. Full financial spreading in a workbook: revenue build, margins, working capital, capex, debt schedule. The workbook ships with the note — formulas visible, assumptions labelled.
4. Structure. Where the debt actually sits: holdco vs. opco, security, maturity walls, covenants, refinancing needs. In project-heavy sectors, structure is half the credit story.
5. View. A stated credit view — improving, stable, or deteriorating — with the two or three drivers that would change it, and a revisit date on which I'll publicly score it.
Business risk before financial risk. For energy-transition issuers that means: counterparty quality (who buys the power and do they pay on time), tariff and PPA structure, resource risk, execution track record, and regulatory exposure.
Financial risk through cash flow, not P&L. Leverage (debt/EBITDA), coverage (EBITDA/interest, DSCR), liquidity (cash + undrawn lines vs. near-term maturities), and funding access. Project-level DSCR sensitivities are run against tariff, generation (CUF), and rate scenarios.
My view vs. the agencies'. Each note states where I land relative to the actual assigned ratings and why. Agreeing is fine; disagreeing requires showing the work. Either way, the reasoning is on the page.
A piece is published on this site only if all three tests pass:
1. Primary sources. It's built from filings and documents I actually opened — never from a summary of a summary.
2. My numbers. Every assumption and figure was set or verified by me. AI is part of my workflow — drafting, extraction, cross-checking — and I document that use openly rather than hiding it. But AI-assisted is only legitimate when the judgment is mine; AI-generated-and-unexamined is fake, and it doesn't ship.
3. Defensible live. If I couldn't defend any line of it in an interview without notes, it isn't ready.
And one standing commitment: every view gets revisited on its resolution date, publicly, whether it aged well or badly. A track record you can't check isn't a track record.