The goods and services tax regime came into force on 1 July 2017. Measured against the Income-tax Act 1961, or against the century of authority behind the Indian Contract Act 1872, that is yesterday. Researching a GST dispute therefore means working in a field where the settled answer often does not exist yet, where a proposition can be accepted in one State and rejected in another, and where the first genuinely well reasoned judgment from any High Court travels across the country within months.
This changes the shape of the research. In a mature practice area you look for the controlling authority and stop. In GST you are usually assembling a field: what has been said, by whom, on what statutory text, and how far it has travelled. The advocate who has surveyed every High Court on the point is not being thorough for its own sake. On a young statute, comprehensiveness is the argument.
It also changes what counts as material. GST research pulls in a class of documents that ordinary case law research would never touch: advance rulings that bind nobody but the applicant, circulars that are executive instructions rather than law, notifications that are amended and withdrawn at short notice, and a very large volume of writ petitions decided at interim stage. Knowing the weight of each of these is most of the skill.
Why the Jurisprudence Is Still Thin
Three structural facts explain the state of GST case law, and each of them has a research consequence.
First, the regime is genuinely new. Every dispute that reaches a High Court today concerns a provision that has existed for under a decade, has often been amended in the meantime, and has usually never been construed by the Supreme Court. Where the Supreme Court has spoken, it has spoken on a handful of questions, and the remainder of the field is High Court authority that binds only within its own territory.
Second, the legislative architecture is dual. The Central Goods and Services Tax Act 2017 sits alongside a State Act for each State and Union Territory, and the operative provisions are, for the most part, mirrored. A judgment construing a State provision therefore usually reads directly onto the corresponding central provision and onto the equivalent provision of every other State. This is a considerable advantage to a researcher who knows to look for it, and a considerable blind spot for one who searches only under the central Act.
Third, for a long stretch after the regime began, the appellate tribunal contemplated by the statute was not functioning. An assessee aggrieved by an appellate order in that period had no tribunal to go to, and the natural consequence followed: an unusual volume of GST disputes was carried straight into High Court writ jurisdiction under Article 226, notwithstanding the ordinary rule that a writ court will not entertain a petition where an efficacious alternative remedy exists. That has left a body of High Court authority on questions that, in a settled regime, would have been resolved at tribunal level and never reported.
The first good judgment travels
In a young jurisprudence, a carefully reasoned judgment from any High Court is read everywhere, because there is nothing else. A researcher who confines the search to a single High Court is not merely missing persuasive material. On many GST questions the persuasive material is the field, and the first court to reason the point well tends to set the terms on which every later bench argues about it.
What Counts as Authority, and What Only Looks Like It
GST research surfaces a wider range of documents than ordinary case law research, and they are not of equal weight. Getting this hierarchy wrong is the commonest failure in the practice area, because a document can look official, be issued by a statutory body, and still be worth nothing as precedent.
| Source | Weight | How to use it |
|---|---|---|
| Supreme Court judgment | Binding on all courts under Article 141 | Where it exists on your point, the enquiry is largely over. Confirm the proposition is the ratio and that the statutory text has not since been amended. |
| Your own High Court | Binding on you and on the authorities within its territory | This is the first thing to find and the last thing to be surprised by. A contrary view from your own High Court decides your matter regardless of how the rest of India has ruled. |
| Another High Court | Persuasive only | On a young statute this is the bulk of the useful material. Weigh the reasoning, note whether the provision construed was the central or a State counterpart, and check whether any court has declined to follow it. |
| Appellate tribunal orders | Persuasive, and subordinate to the High Court | Useful for the department's arguments and for how a point is usually run, but never a substitute for High Court authority on the same question. |
| Advance ruling by a State Authority | Binds only the applicant and the concerned officer, in relation to that applicant | Intelligence, not precedent. Cite it only to show how a question has been approached, never as authority, and expect the other side to point out that it binds nobody. |
| Appellate Authority for Advance Ruling | Same limited binding effect, on appeal from the ruling | Slightly better reasoned as a rule, and often the place where conflicting State positions become visible. |
| Circulars and instructions | Executive instructions binding on the department, not on the assessee or the courts | Powerful when they favour the assessee, because the department cannot resile from them. Worth nothing against the statute where they purport to add to it. |
| Notifications | Delegated legislation, valid to the extent of the power conferred | Always check the version in force on the relevant date, and whether it has since been amended, superseded or withdrawn. |
Advance rulings: useful, and routinely misused
Advance rulings deserve their own paragraph because they generate more citation errors than anything else in the practice. The statutory scheme is clear enough: a ruling is binding on the applicant who sought it and on the officers concerned with that applicant, and it does not operate as a precedent for anybody else. States run their own authorities, they reach different conclusions on materially identical facts, and nothing in the scheme resolves the conflict.
That does not make them useless. A ruling tells you how a revenue authority has read a provision, what facts it thought mattered, and where the department's argument will start. Where several States have ruled the same way, that is a real signal about administrative practice. Where they have split, that split is itself worth placing before a court considering the question for the first time. What you cannot do is put a ruling in a written submission as though it were authority, and a bench that notices will discount the rest of your citations along with it.
The Writ Overhang, and How to Read It
Because so much GST litigation has run through Article 226, a researcher has to be unusually careful about what a High Court order actually decided. Writ jurisdiction produces a great many orders that are not adjudications of the substantive question at all: an order setting aside a demand for breach of natural justice decides nothing about whether the demand was right; an order remitting a matter for a fresh hearing decides nothing about the credit in dispute; an interim order protecting an assessee against coercive recovery is not a finding that the assessee is correct.
The procedural categories where this matters most are the ones that recur: cancellation and revocation of registration, refund claims and the conditions attached to them, the conditions on which input tax credit is allowed and the consequences of a supplier's default, blocking of the electronic credit ledger, transitional credit, detention of goods and vehicles in transit, the sufficiency of a show cause notice, and limitation for passing an order. On several of these, High Courts have taken visibly different positions, and on several the answer within a State has changed as the same court has revisited the question.
The practical discipline is to classify every order you intend to rely on before you rely on it. Was the relief procedural or substantive? Was the writ entertained because an exception to the alternative remedy rule applied, and if so which one? Did the court record that it was not deciding the merits? A short note against each authority saving that classification will save you a bad afternoon later.
Mirrored Provisions: One Judgment, Many Statutes
The mirroring of central and State provisions is the single most exploitable feature of GST research, and it is routinely left unexploited. A judgment of one High Court construing a State provision is, in substance, a construction of language that appears in identical or near identical terms in every other GST statute in the country.
- Search the concept, not the enactment. A search confined to the central Act will miss a squarely relevant judgment delivered under a State Act using the same words.
- Compare the texts before you argue the transfer. Near identical is not identical. Set the two provisions side by side in your note, and be ready to show the bench that the language is the same.
- Watch the numbering. Corresponding provisions usually carry the same section numbers across the central and State statutes, but this is a convenience rather than a guarantee, and amendments can disturb it.
- Note which statute the parties argued. A judgment construing a State provision will express itself in the vocabulary of that statute. Quoting it without explaining the equivalence invites an easy objection.
Pre-GST Analogues: When Old Authority Still Persuades
The regime that GST replaced generated decades of authority on excise, service tax and State sales tax and value added tax. Some of that authority is dead on arrival, because the statutory scheme it construed has gone. Some of it is highly persuasive, because the concept it construed has been carried forward in substance.
The distinction turns on the language and the structure, not on the label. Where a general principle of taxing statutes is in issue, the older authority is as good as it ever was: how an exemption notification is construed, what happens when the charging provision and the machinery provision do not align, the treatment of a levy without statutory authority, the limits of delegated legislation, the effect of a circular that conflicts with the statute. Where the question turns on a scheme that GST redesigned, the old case law can mislead precisely because it is familiar. The credit mechanism is the obvious example: the vocabulary of input credit survived the transition, and the conditions attached to it did not survive unchanged.
A pre-GST authority is worth what the similarity of the statutory language is worth. Set the two texts next to each other. If you cannot show a bench that they say the same thing, do not lead with it.
A Method for GST Research
The sequence below is written for a real dispute with a real deadline. It runs from the narrowest binding material outward, and it front loads the two checks that most often invalidate an otherwise good argument: the text in force, and the position in your own High Court.
Fix the date and the text
Establish the date on which the liability, the transaction or the impugned act arose, then reconstruct the statutory text, the rules, and any relevant notification as they stood on that date. GST provisions and rates have been amended frequently, and an argument built on the current text about a transaction from an earlier year is not an argument. Record the version you relied on and where you got it.
Read your own High Court first
Whatever the rest of the country has held, a decision of the High Court to which your authority is subordinate governs. Search it exhaustively, including on the mirrored State provision, and including orders that disposed of writ petitions without detailed reasons, because the department will cite them.
Survey every other High Court on the concept
Run a single search that reaches all 25 High Courts, and search by concept rather than by fixed phrase, because benches describe the same dispute in different words. One court frames a question as blocked credit, another as denial of credit for a supplier's default, and a keyword search that matches one will miss the other. This is the step that builds the field.
Classify what each judgment actually decided
For each result, record whether the relief was procedural or substantive, whether the writ was entertained despite an alternative remedy and on what basis, and whether the court expressly left the merits open. Half of what looks like a favourable line of authority turns out, on this test, to be a line of remands.
Pull the administrative layer
Find every circular, instruction and notification touching the point, and check each for later amendment or withdrawal. A circular favourable to the assessee is often the fastest route to relief because the department is bound by it. A circular that goes beyond the statute is vulnerable, and saying so is a substantive argument rather than a complaint.
Use advance rulings as intelligence
Collect the rulings across States on the same question, note where they conflict, and use them to predict the department's case and to show a court that administrative practice is inconsistent. Keep them out of your table of authorities. If you refer to one, say expressly what it binds and what it does not.
Test the pre-GST analogue
Where the point has no GST authority, identify the closest provision in the earlier indirect tax statutes and set the two texts side by side. If the language and the scheme carry over, the older line of authority is genuinely persuasive and should be led as such. If they do not, say so yourself before your opponent does.
Check what has happened since
For every judgment you intend to cite, check whether it has been followed, distinguished or doubted, whether a larger bench or another High Court has taken a different view, and whether a special leave petition or appeal is pending against it. On a young statute the position moves quickly, and an authority that was the leading case last year may now be under challenge.
Coverage beats depth, early
On a settled question you read three authorities closely. On a GST question you first need to know how many exist and where. Survey the whole field before you decide what deserves a full read.
Weight before content
Classify every document by what it binds before you quote from it. A ruling, a circular and a judgment can all say the same thing and carry entirely different force in court.
Dates are substantive
In a regime amended this often, the date of the transaction is a legal fact. Every authority you cite has a text behind it, and that text has a version history.
Where CourtMesh Fits, and Where It Does Not
The hardest mechanical step above is the third one: surveying every High Court on a concept rather than a phrase. CourtMesh runs a single unified search across the Supreme Court, all 25 High Courts, the District Courts and Tribunals, drawn from official government portals, with roughly 310 million records keyword-searchable. Structured filters for court, year, case type and date range let you narrow that field once you have it. Semantic retrieval, which covers roughly 2 million judgments rather than the whole corpus, is what catches the decision that is squarely on your point and written in another bench's vocabulary.
Related-case discovery and citation relationships help with the last step, showing how a judgment has been treated since. That analysis covers a subset of the corpus and is expanding rather than complete, so the absence of a flagged relationship is never proof that nothing has happened. It narrows where you look. It does not conclude anything.
What no research tool will do for you
It will not tell you whether a State provision is materially identical to the central one, whether a pre-GST authority survives the change of scheme, or whether a favourable order was a remand dressed as a victory. Those are readings, and they are the work. Whatever a tool displays, the record of the issuing court is the authoritative one, and the statutory text should be confirmed against the Gazette and the official portals before it goes into a submission.
Survey the whole country, then argue
GST rewards the researcher who sees the field first. One search across the Supreme Court, every High Court, the District Courts and Tribunals, semantic retrieval that catches the differently worded judgment, filters for court, year and case type, and citation signals to check what has happened since. Build the all India picture on your provision, classify each source by what it actually binds, and take the reasoning to the bench with the dates and the text version already nailed down.
Explore CourtMesh
