Credit Rating Agency: How Ratings Are Assigned and Regulated

A credit rating agency is an entity registered with the Securities and Exchange Board of India to assign opinions on the creditworthiness of issuers and instruments. A rating is assigned through a defined process that ends in a committee decision, and it stays under review for as long as the instrument lives. The registration, the process requirements and the disclosure obligations are published by the regulator at sebi.gov.in, where the current text is confirmed.

Two things sit underneath this role and they pull in opposite directions. A ratingAn opinion on creditworthiness, expressed on a scale the agency itself defines and publishes. is an opinion, formed mostly on information the agency receives from the very entity it is forming the opinion about, and paid for by that same entity. Say that out loud once and it sounds broken. It is not broken. It is the arrangement the market runs on, and almost everything regulated about the role exists to make an opinion formed under that arrangement into something a stranger can lean on.

The shape of this is familiar from ordinary life. A cousin asks somebody to stand behind a loan for a person they have never met. The guarantor cannot look at that person's books, would not know what to look for, and has no way of chasing them for the truth. So the guarantor asks somebody who does that for a living. The answer that comes back is not a promise and not a percentage. It is a view, held by a person whose method can be asked about, based on papers the borrower handed over. A credit rating is that same object, scaled up and registered.

The registration, the process a rating has to travel through, the review that follows it and the disclosures attached to every change are Indian throughout, and all of them come from the rulebook named further down. The one thing a credit rating agencyAn entity registered with the Securities and Exchange Board of India to assign opinions on creditworthiness. is registered to do takes one line to state. Who decides a rating, who pays for it, what happens to it on every day after publication, what has to be said out loud when it moves, and the two things it is not are all settled in the same rulebook.

The case running through this guide is the Vindhya Ceramics raise. Vindhya Ceramics Private Limited, an invented company, raised Rs 40,00,00,000, made up of Rs 25,00,00,000 of equity and Rs 15,00,00,000 of debentures. Ratnakar Deshpande is its finance director. Sahyadri Ratings Limited is the credit rating agency that rated the Rs 15,00,00,000 debenture portion, and Anantpur Trusteeship Services Limited is the debenture trustee standing beside the same portion. One thing is missing throughout: the symbol Sahyadri Ratings assigned. A symbol means nothing away from the scale that defines it, and that is why it stays blank.

What is a credit rating agency registered to do?

A credit rating agency is a registered entity, not a description somebody can adopt because it suits what they are doing. Registration comes first and the activity comes second. Sahyadri Ratings Limited was registered before it ever received a document from Vindhya Ceramics, and the registration is a thing a reader can confirm at the regulator rather than take on trust from a letterhead.

The registration permits something narrow, and the wording is worth stating precisely. The registration permits the entity to assign opinions on creditworthiness, meaning views on how likely an issuer or an instrument is to meet what it has promised to pay. Assigning those opinions is the whole permitted output. The registration is not a permission to advise anybody, not a permission to manage money, not a permission to sell a security and not a permission to certify anything as safe.

The registration licenses the production of an opinion and nothing beyond an opinion. Almost every misreading of a rating is somebody quietly upgrading it into something the agency was never registered to produce.

There is a second half to the registration that gets less attention. Being registered does not only unlock an activity, it attaches obligations that would not otherwise exist: a process the agency has to follow, records it has to keep, a review it has to keep running, disclosures it has to make when things change, and disclosure about its own arrangements and its own conflicts. Sahyadri Ratings may do one thing, and a single sentence states it. The list of what Sahyadri Ratings must do runs a great deal longer.

How does a rating actually get assigned, from information to a decision?

A rating is not typed out by whoever knows the sector best. A rating travels through a defined sequence, and the sequence is prescribed rather than chosen by the agency. Information arrives. Analysts work it up. A committee takes the decision. The decision is recorded with its reasoning. The agency publishes the rating together with the rationale behind it.

The step worth slowing down on is the third one. The person who did the work does not make the call. A rating committeeThe body inside the agency that takes the rating decision, rather than any individual analyst. takes it, and that separation is required rather than a matter of internal preference. The same protective move recurs across registered entities: distance between the person closest to the pressure and the person who signs.

Separating the analyst who prepared the view from the body that takes the decision is the single structural protection inside the assignment process, and it is set in the rules rather than left to the agency to arrange as it prefers. Think of a household deciding whether to lend to a relative. One member gathers the facts, and the household decides together. Nobody in that room thinks the arrangement is an insult to the person who gathered the facts. The separation exists because the person closest to a conversation is the person most exposed to it.

Who pays for a rating, and what does that arrangement create?

Here is the uncomfortable part, and it goes early rather than late because burying it would be its own small dishonesty. Vindhya Ceramics Private Limited supplied most of the information on which the rating was formed. Vindhya Ceramics Private Limited paid Sahyadri Ratings Limited for the rating. And the rating that came out of it is read by people who supplied nothing and paid nothing.

The arrangement has a name. The name is issuer paysThe arrangement in which the entity being rated is the entity that pays for the rating., and calling the arrangement by its name is far more useful than working around it. Two of the three things moving in this picture start at the party being rated. Only the third one, the opinion itself, travels outward to everybody else.

The same shape appears in ordinary life and usually passes without comment. A man selling a second hand scooter pays a mechanic to write a condition report, and the buyer reads the report. Nobody thinks the mechanic is corrupt. Everybody understands, without being told, that the report is worth more if the mechanic works to a known method, writes down what was checked, and has a name that would suffer if the report turned out to be worthless. Method, record and reputation are the entire regulatory answer, written out in a lane outside a scooter shop.

The party being rated selects and pays the party forming the opinion, and the regulation exists because of that arrangement rather than in ignorance of it. Notice what the honest response to this is and what it is not. The dishonest response is reassurance: agencies are professional, reputations matter, it works out in practice. Some of that is true and none of it is the answer. The answer is that the conflict is named openly, that the role is registered rather than merely contracted, that the process ends in a committee rather than an individual, that the method is published so a stranger can see how the view was reached, and that the agency has to disclose its own arrangements and its own conflicts. A structural conflict is handled by building duties and daylight around it, never by insisting it is not there.

What does the agency do with the information it receives, and what must it keep?

Information reaching Sahyadri Ratings Limited from Ratnakar Deshpande arrives in two conditions. Some of it is already public and would be available to anybody who went looking. Some of it is not public at all, and Vindhya Ceramics handed it over on the understanding that it was going into a rating rather than out into the world.

Information that is not public creates an obligation with a sharp edge on it. Information given to an agency for the purpose of forming a rating is used for forming the rating. It is not traded on, not passed sideways, not used to advantage anybody, and not published because it would make the rationale read better. An agency that leaked what an issuer handed it would stop receiving anything worth having within a season, and every rating it published afterwards would be thinner for it.

A decision nobody can reconstruct afterwards is indistinguishable from a decision nobody really took, so the record of how the opinion was reached matters as much as the opinion. So what was received, what was analysed, what the committee considered and what it concluded all get written down and kept. How long they are kept for is set in the rulebook, and intervals move, so a stale interval quoted confidently is worse than no interval at all.

Is a rating finished on the day it is published?

The instinct almost everybody has is the wrong one, so answer the question first. A rating looks like a verdict. Verdicts are delivered and then they are over. A rating is not over when it is delivered.

What happens to a rating between the day it is assigned and the day the instrument ends?

A rating is kept under review. The word for that work is surveillanceKeeping an assigned rating under review for as long as the rated instrument remains outstanding., and it runs from assignment until the Rs 15,00,00,000 debenture portion is redeemed. Information keeps arriving from Vindhya Ceramics. Sahyadri Ratings keeps looking at it. The opinion is revisited, and it is either confirmed as it stands or it moves.

Continuous review is where a small and extremely common reading error does real damage, and the error has nothing to do with finance. Somebody finds a rating quoted in a document, in a presentation, in a summary written by a helpful colleague. The symbol is there. The date is not. And a symbol with no date attached is not an opinion about today at all: it is an opinion about whichever day it was last confirmed, still sitting there looking current.

A rating without the date it was last confirmed is a historical opinion that happens to still be printed, and treating it as a present one is a mistake that costs nothing until the day it costs everything. Think of the fitness certificate stuck to the windscreen of a hired vehicle. Nobody reads the sticker as a claim about the engine this morning. Everybody reads the date on it first. A symbol looks self contained in a way a sticker does not. A rating deserves exactly the same reflex as the sticker and rarely gets it.

What is a rating action, and what has to be disclosed when one happens?

Every time the agency does something to the rating, that is an event with a name. A rating actionA change, a confirmation or a withdrawal of a rating, each of which carries its own disclosure. covers a change, a confirmation and a withdrawal, and each of the three carries a disclosure of its own.

The one people miss is the confirmation. Nothing changed, so it feels like nothing happened. Something did happen: the agency looked again and decided the opinion still stands, and it said so out loud. Without confirmations, an actively reviewed rating and one nobody has thought about in years would look identical from outside, and nothing would separate them. Silence would mean two completely different things at once.

Disclosure on every rating action is what makes the review visible from the outside. Without that disclosure nobody could tell a living opinion from an abandoned one. A withdrawal works the same way and is worth naming separately. A rating that simply vanished with no notice would leave everybody guessing at why, and a rating does not vanish. A rating is withdrawn, and the withdrawal is itself an event that gets said.

What is a rating scale, and why is a symbol meaningless without it?

Every agency publishes a rating scaleThe agency's own published set of symbols together with what each one is defined to mean.: its own set of symbols, and the definition of what each symbol is being used to say. The scale belongs to the agency. It is the agency's definition, published by the agency, and it is the only thing that makes any symbol on it mean anything at all.

Sahyadri Ratings Limited assigned a symbol to the Rs 15,00,00,000 debenture portion, and that symbol is left blank throughout. A mark printed away from its scale has its meaning stripped out. Somebody then carries the mark onward, and by the third retelling it has acquired whatever meaning felt natural. Preventing exactly that is the whole purpose of the scale.

The everyday version is exact rather than approximate. A vendor says the mangoes are sixty. Sixty what. Sixty a kilo, sixty a dozen, sixty for the crate: those are three different transactions and the number is identical in all three. Nobody hands over money on the number alone. A rating symbol means what the agency's own published scale defines it to mean, so a symbol quoted away from its scale is a number with no units attached.

There is a harder version of the same rule that a lot of people get wrong, so it is worth stating plainly. Each scale is the definition of the agency that published it, so two agencies can use symbols that look similar and define them differently. Treating a symbol as though it belonged to the market rather than to the agency is how a comparison that was never valid gets made confidently and repeated.

What is a rating not, however it is presented?

A rating is an opinion on creditworthiness, held on a scale the agency defines. Two other things get read into it constantly, and both are named below without softening.

A rating recommends nothing. Nobody at Sahyadri Ratings Limited is advising anybody to buy the debentures, avoid them, hold them or sell them, and the agency is not registered to advise anybody about anything. The agency has produced a view on how likely the Rs 15,00,00,000 debenture portion is to be paid as promised. Acting on that view is a separate question, answered by a separate person, working under a separate registration.

A rating is also not a probability, and the probability reading is the one that does the quieter damage. A rating is a place on a defined scale, not a number carrying a chance of anything. When somebody multiplies a rating by an exposure to arrive at an expected loss, every step of that arithmetic is correct and the input never meant what the calculation needed it to mean. The dangerous property of the second misreading is that it produces a rigorous looking number out of a meaning the scale never claimed, and rigorous looking numbers travel further and faster than the caveats attached to them.

The everyday version is a doctor at the end of a check up telling a patient they are in good shape for their age. Good shape for their age is a considered opinion from somebody qualified to hold it. It is not an instruction to go running, and it is certainly not a percentage that belongs in a spreadsheet. Anybody who converted it into one would have invented the number themselves and then attributed it to the doctor.

What must an agency disclose about its own method and its own conflicts?

An agency publishes its methodologyThe agency's own published description of how it goes about reaching a rating.: the description of how it goes about reaching a rating for a given kind of issuer or instrument. The agency also publishes the rationale that sits behind a particular rating, and discloses its own arrangements, including how it is paid and what relationships it has with the entities it rates.

Publishing the method does something specific and limited, and both halves are worth stating precisely. A published method lets a reader see how the view was reached rather than only what the view was. A published method does not let anybody arrive at the same view independently. The judgement inside the process remains the agency's judgement, taken in a room the reader was not in, on information some of which is never disclosed.

A published method makes the reasoning checkable without making the judgement repeatable, and confusing those two is how readers end up either dismissing the method as window dressing or trusting it far past what it can carry. Both errors come from the same place: expecting a published method to be a formula. A methodology is not a formula. It is an account of what the agency looks at and how it weighs what it finds.

How is the method behind a rating found?

The route is short and entirely public, and it starts in a place people rarely start. The agency's name is part of the rating and not decoration around it, so note which agency assigned it. Go to that agency's own published material rather than to the issuer's. Find the scale, so the symbol acquires a meaning, and find the methodology for that kind of instrument. Then read the rating action itself, with its date, rather than the symbol lifted out of it.

The issuer can give the symbol and cannot say what it means. Every step of the route therefore starts at the agency, and none of it starts at the company being rated. Ratnakar Deshpande will happily confirm what rating the Rs 15,00,00,000 debenture portion carries. He is not the author of the scale, the method or the action, and nothing about that is a criticism of him.

What does somebody actually do with all of this in a working week?

Meera Vaidyanathan is the investment manager at Neelanchal Asset Managers Private Limited, and the Neelanchal India Growth Fund is looking at the Vindhya Ceramics debenture portion. Her use of the rating in forming a view is set out under credit analysis. Her compliance and records work follows directly from the machinery set out above.

She confirms that Sahyadri Ratings Limited holds a current registration, at the regulator rather than on the agency's own description of itself. A symbol travelling without its agency is a symbol without its scale, so she writes down which agency assigned the rating. She records the date of the most recent rating action rather than the date she happened to read it. She keeps the agency's published methodology and scale alongside her own file. In eighteen months nobody will remember which version was current when the file was made. And she watches for rating actions afterwards, including confirmations. A run of confirmations and a long silence look identical in a summary and mean completely different things.

None of that work is an investment judgement, and all of it is what makes an investment judgement recordable afterwards. The same discipline turns up at a lender's credit desk and, in a much smaller way, in a household keeping the papers from a loan it took: not the decision, but the file that lets anybody reconstruct what was known and when it was known. The five steps above are exactly the ones somebody wishes they had taken when they are asked, two years later, what they were looking at.

Sahyadri Ratings Limited, Vindhya Ceramics Private Limited, Anantpur Trusteeship Services Limited, Trilokpur Capital Markets Private Limited, Neelanchal Asset Managers Private Limited, the Neelanchal India Growth Fund, Ratnakar Deshpande and Meera Vaidyanathan are invented.
Educational material. Not advice on any investment, tax, budget or market position.