Assess the role of robust official statistics such as the ASI in evidence-based industrial policymaking. What are the limitations of a factory-register-based frame?

The Annual Survey of Industries (ASI) is India's principal source of data on registered manufacturing. MoSPI conducts it under the Collection of Statistics Act, 2008 [3]. Its 2024-25 round shows employment up 7.19% and GVA up 9.59% [1]. Figures like these can anchor industrial policy, but only if the limits of the frame behind them are understood.

Role in evidence-based industrial policymaking

• Macro planning: ASI data feeds National Accounts Statistics, which are used to compile GDP [1].
• Sectoral targeting: Basic metals, motor vehicles, chemicals, pharmaceuticals and food products give over 45% of manufacturing GVA [1]. This helps direct sector-specific incentives.
• Regional and federal policy: Five States hold over 54% of GVA [1]. Tamil Nadu leads in employment (14.99%) while Maharashtra leads in GVA (15.92%) [1]. This separates labour-intensive from capital-intensive growth models for State industrial policies.
• Labour and productivity signals: Emoluments grew 12.08% against employment growth of 7.19%, and invested capital (+11.10%) grew faster than GVA [1]. These point to rising pay per person but slightly lower returns on capital.
• Credibility: The survey is statutory and response is compulsory. It has run annually since 1960, except in 1972, which gives a long, comparable series [3].

Limitations of a factory-register-based frame

• Out-of-date registers: The sampling frame comes from State Chief Inspector of Factories lists. Closed units stay on these lists as "deleted factories", and new units get missed [3]. About 18% of 2,66,931 registered factories were not in operation [1].
• Narrow coverage: Only units defined as factories under Section 2m of the Factories Act, 1948 are covered. These are units with 10 or more workers using power, or 20 or more without power [3]. Unregistered units, computer services and defence establishments are excluded [3].
• Formalisation shown as job creation: When a unit crosses the worker threshold, it enters the frame for the first time. Its existing workers then show up as "new" jobs.
• Breaks in the series: Coverage depends on the legal definition of a factory. If the Labour Codes change that definition, figures before and after the change will not be comparable.
• Time lag: The 2024-25 results came out about 18 months after the year ended [1]. The 2023-24 round had a similar lag [2].

The ASI therefore remains indispensable, but its register-based design limits what it can show. Useful reforms would be:

• annual clean-up of the factory registers by State CIFs;
• a bridge series overseen by the Standing Committee on Industrial Statistics (the body that approved the 2012-13 sampling redesign) before any change in the definition [3];
• reading ASI alongside PLFS and ASUSE.

Better statistics would support industrial policy that meets SDG 8 (decent work) and SDG 9 (industry and innovation).