How Indian Scholars Should Read a GST Notification Before Modelling Tax Incidence

Indian scholars who model tax incidence, firm behaviour, or consumption effects frequently cite “the GST rate” as if it were a single, stable parameter. It is not. The Goods and Services Tax regime is administered through a continuous stream of notifications, circulars, and rate schedules issued by the Central Board of Indirect Taxes and Customs and the GST Council. A notification can change the effective rate on a narrow HSN/SAC slice, introduce a new exemption with conditions, alter place-of-supply rules, or modify the input-tax-credit chain without touching the headline rates published in the press. Treating the latest newspaper summary as the legal instrument produces mis-specified regressions, overstated pass-through elasticities, and teaching slides that collapse under the first student who opens the actual gazette.

This briefing is written for finance, public-policy, and commerce scholars who need to move from a notification number to a defensible incidence model. It assumes you already know the broad architecture of GST—CGST, SGST, IGST, and the dual levy—but need a practical protocol for reading the instrument itself before you code a variable or lecture a class.

Why a GST Notification Is Not a Rate Change

The most common error is to equate any GST notification with a change in the statutory rate. Many notifications are clarifying, procedural, or conditional. Notification No. 12/2017-Central Tax (Rate), for example, lists exemptions for services; subsequent amendments add or remove entries, attach conditions, or redefine the scope of “educational institution.” A rate notification under Section 9 may alter the percentage for a specific tariff item while leaving the residual rate untouched. When scholars code a binary “post-GST” dummy or a single rate variable for the entire period after July 2017, they average across heterogeneous treatments.

Incidence models are especially sensitive to this averaging. The economic incidence of a tax depends on the relative elasticities of supply and demand, the market structure, and the completeness of the input-tax-credit chain. A notification that exempts a intermediate good can raise the effective tax on the final good if credit is blocked; the same notification can lower the effective rate if the exemption is accompanied by a refund mechanism. Without reading the exact operative text, the researcher cannot sign the predicted direction of the price or quantity response.

The Anatomy of a GST Notification: What to Extract First

Open the PDF or the official gazette version, not a secondary summary. The first page usually contains the notification number, the date, the section of the Act under which it is issued, and the short title. Immediately note three items:

• The legal basis (for example, Section 11(1) for exemptions, Section 9 for rates, or Section 15 for valuation).
• The effective date, which may be prospective, retrospective, or conditional on a later event.
• Whether the notification amends an earlier notification or stands alone. Most rate and exemption changes are issued as amendments to the principal notifications of 2017.

Next, locate the table or the schedule that is being altered. GST notifications almost always work by inserting, substituting, or omitting a serial number in a table. The serial number points to a description of goods or services, an HSN/SAC code, a rate, and sometimes conditions. Extract the exact wording of the entry before and after the amendment. If the notification refers to “as amended,” you must reconstruct the chain of prior amendments; the current consolidated text is available on the CBIC website but must be verified against the individual notifications if your event study requires precise dating.

Also record any footnotes, explanations, or provisos. These frequently contain the definitions that determine whether a firm’s product falls inside or outside the changed category. A change that looks narrow on the serial number can be broad once the explanation redefines “processed food” or “information technology software.”

Distinguishing Rate, Exemption, and Procedural Notifications

Scholars should classify the notification before deciding how to code it.

Rate notifications alter the percentage of tax payable. They are issued under Section 9 and typically amend the schedules to Notification No. 1/2017-Central Tax (Rate) or the corresponding IGST/SGST instruments. For incidence work, these are the cleanest shocks, but only if you can map the HSN codes to the products in your data.

Exemption notifications remove the tax entirely or reduce it to nil under specified conditions. They are issued under Section 11 and amend Notification No. 2/2017 (goods) or 12/2017 (services). An exemption can increase or decrease the effective tax on downstream products depending on whether input credit is available. Model this as a change in the credit chain, not simply as a zero rate.

Procedural and valuation notifications change how the tax is computed or reported—rules for e-invoicing thresholds, place of supply, reverse charge, or the composition scheme. These rarely produce a clean price response but can affect compliance costs and the measured tax base. Do not code them as rate changes.

Building the Incidence Model from the Notification, Not the Headline

Once the operative text is extracted, construct the treatment variable from the legal instrument. For a simple difference-in-differences design, define the treated group as the set of firms or products whose HSN/SAC falls inside the amended entry and whose transaction date is on or after the effective date. The control group should be products that remain under the residual rate or under an unchanged serial number.

For pass-through estimation, the relevant price is the consumer price inclusive of tax, but only after confirming whether the notification alters the credit available to the supplier. If the supplier can claim input tax credit on the exempt intermediate, the net tax wedge may be smaller than the statutory change. Document the credit position in the paper; do not assume full credit or zero credit without checking the conditions attached to the entry.

When the notification is an amendment, the pre-period should end on the day before the effective date of the amendment, not on the date of the original 2017 notification. Many event-study papers inadvertently treat the entire post-July 2017 window as treated, which confounds the specific shock with the general introduction of GST, demonetisation aftermath, and other concurrent reforms.

If you are modelling firm-level outcomes such as investment or employment, recognise that a rate change on an intermediate input affects the user cost of capital only if the input is capitalised or if the credit chain is incomplete. A pure consumption-tax change on final goods may leave the user cost unchanged. State the mechanism explicitly.

Common Errors Scholars Make When Coding GST in Firm-Level Data

Three errors appear repeatedly in working papers that use Indian firm data:

• Using the announcement date instead of the effective date. Markets may react on the GST Council meeting date, but the legal incidence begins only on the date specified in the notification. For quantity or compliance outcomes, use the effective date; for announcement effects on stock prices, use the Council decision date and note that the notification may still be modified before it is issued.
• Ignoring state-level variations. Although GST is largely uniform, certain exemptions or rates can be adjusted by state notifications or by the GST Council’s recommendations that are implemented with a lag. For studies that use state-level price data, verify whether the central notification was mirrored by the corresponding state instrument.
• Treating composition-scheme or reverse-charge changes as ordinary rate changes. These alter who remits the tax and whether credit is available to the buyer. The measured tax wedge for a downstream firm can move in the opposite direction from the headline rate.

A fourth, quieter error is the failure to archive the version of the notification that was in force on the event date. CBIC consolidates texts; the consolidated PDF may contain later amendments. Download and store the individual notification PDF with its date stamp.

Teaching and Citing the Notification Without Overclaiming

When you teach the episode, show students the first page and the amended serial number, not a news graphic. Ask them to identify the legal basis, the effective date, and whether the change is a rate, an exemption, or a procedure. Then ask them to write one sentence that states what the notification does and one sentence that states what it does not do. This exercise prevents the common classroom claim that “GST was reduced on consumer goods in 2019” when the actual instrument covered only a subset of items under specific value thresholds.

In papers and theses, cite the notification by its number, date, and the specific serial or entry that is relevant. A citation such as “Notification No. 1/2017-Central Tax (Rate) as amended” is insufficient for an event study; the reader must be able to locate the exact amendment. If you rely on a secondary compilation, state that you verified the operative text against the original gazette or CBIC PDF.

Further reading on DrDPKlass: the earlier briefing on reading an RBI circular and the briefing on reading a Union Budget annex follow the same discipline of instrument over headline. The same caution applies when the instrument is a GST notification.

Desk Checklist Before You Code or Teach

• Have you downloaded the original notification PDF (not only a news summary or a consolidated schedule)?
• Have you recorded the notification number, date, legal basis, and effective date?
• Have you identified whether the instrument amends an earlier notification and reconstructed the pre- and post-text of the relevant serial number?
• Have you mapped the HSN/SAC codes in the notification to the product or industry codes in your data, and noted any residual mismatch?
• Have you classified the change as rate, exemption, or procedural, and adjusted the incidence mechanism accordingly (including the credit chain)?
• Have you chosen the pre-period and post-period relative to the effective date of this specific amendment, not the general GST introduction date?
• Have you archived the version of the text that was in force on the event date?
• In the paper or lecture, do you cite the exact entry and state what the notification does not change?

A GST notification is a precise legal instrument. Scholars who read it as such produce incidence estimates and classroom explanations that survive the next amendment and the next referee. Those who code the headline produce results that cannot be replicated once the serial number is examined. The difference is thirty minutes with the gazette and a checklist.