Examine the use of GST administrative data in official statistics. What safeguards are needed to make sure such data is reliable?

India's GST network produces monthly sales data for every registered business. MoSPI's trial Index of Services Production (ISP) (base 2024–25) is among the first official statistics to use it. The index treats outward supplies reported in GSTR-1 as a proxy for services output [1]. The data is fast and wide in coverage, but it is only reliable if safeguards are built in.

How GST data is used

• Output proxy: sales are recorded by service code (SAC). They are mapped to NIC codes, added up and deflated to get real output [1].
• Privacy: MoSPI uses only aggregated data, never unit-level records [1].
• High frequency: monthly indices for 19 sub-sectors, covering about 60% of services [2]. For example, retail trade grew 18.5% in July 2026 [3].
• Faster tracking: it gives a read on services between quarterly GDP estimates.

Limitations

• Formal-sector bias: only GST-registered entities are counted. Providers below the turnover threshold are left out [1][4]. The TAC confirms this built-in limit [5].
• Exempt sectors: health and education are outside GST [4]. They will be added through ASISSE, a separate MoSPI survey [2]. However, ASISSE also samples from the GSTN register [4].
• Compliance vs. output: more firms filing, or firms declaring sales they used to hide, can look like real growth.
• Deflator mismatch: banking and insurance are deflated by the general CPI [1]. The approach paper prefers producer price indices (PPIs) [4].
• Misclassification: wrong SAC–NIC mapping assigns output to the wrong sector.

Safeguards needed

• Separate the effect of new registrations from growth among existing filers.
• Better deflators: use the Service PPIs launched in June 2026 for banking, insurance, railways and air travel [6]. These are quarterly, so monthly values need interpolation.
• Cross-validation: check results against administrative data (railways, aviation) and surveys. Keep the trial phase and the annual revision policy [2].
• Informal coverage: add survey-based estimates such as ASUSE (the survey of unincorporated enterprises) for unregistered units.
• Data governance: a formal GSTN–MoSPI data-sharing framework, confidentiality rules, and published metadata and revision rules.
• Expert oversight: continued technical review on the model of the TAC [5].

GST data can modernise Indian statistics, turning tax records into timely economic signals. To be reliable, it needs sound deflators, cross-checks and a way to capture informal-sector activity. Built this way, the ISP can be a credible counterpart to the IIP. It would also support evidence-based policy and the UN Fundamental Principles of Official Statistics.