Watch electricity usage to track industrial activity in real time

Yet, policymakers and analysts typically rely on lagging indicators. 

The Index of Industrial Production (IIP) has a lag of 42 days, which is being lowered to 28 days; quarterly GDP is reported with a two-month delay; and the Annual Survey of Industries (ASI) is released about two years after the relevant financial year. 

The real-time pulse of our industrial economy, however, lies not in delayed data and indices, but in the silent electric current flowing through our wires, which is updated every second, everywhere.

Also Read: Rahul Jacob: Manufacturing is crying out for a reality check

Electricity consumption is one of the most immediate, granular and underused indicators of industrial output. It powers not just lighting and ventilation, but also looms, furnaces, extruders, assembly lines and cold storage facilities. Where there is manufacturing activity, there is electricity demand, and changes in it offer powerful insights into the industrial sector’s health.

We conducted an analysis using the ASI of 2022-23 to assess the strength of the relationship between electricity consumption and manufacturing output. 

Across a sample of nearly 48,000 factories, the correlation between the logarithm of electricity purchased and consumed and the logarithm of manufacturing output was a striking 0.80 (1 being perfect correlation). A ‘local polynomial smoothed curve’ (or a fitted line) plotted over the data showed a tight upward-sloping trend across the output spectrum, suggesting that electricity usage tracks production volumes with high precision. 

The data strongly suggests that electricity usage can reliably serve as a proxy for factory output across plant sizes and industries.

Also Read: Think ahead: India’s electronics manufacturing must go up the value curve

What stands out is the strong correlation between electricity consumption and manufacturing output across states. Those like Meghalaya, Jharkhand, Punjab and Odisha top the list with coefficients above 0.84, while even large and industrially diverse states like Tamil Nadu (0.80), Maharashtra (0.79) and Gujarat (0.76) show robust associations. 

The pattern holds in less industrialized regions too; Assam shows a correlation of 0.81, and so does Jammu & Kashmir and Ladakh. Only a few states like Nagaland (0.62) and Tripura (0.69) fall significantly below. In all, the broad reliability of this relationship across geographies is clear.

Also, electricity-output correlations remain strong across industries. Energy-intensive sectors like rubber and plastics (0.86), paper and printing (0.85), pharmaceuticals (0.85) and metal and fabricated products (0.84) top the list. Even moderately intensive sectors such as textiles (0.79), electronics (0.77) and food and beverages (0.76) show consistent and robust associations.

Also Read: Going nuclear will be the only way to keep the lights on as AI guzzles ever more electricity

Just because electricity use and manufacturing output move together doesn’t necessarily mean that one causes the other. So, we conducted a more rigorous analysis using an instrumental variable (IV) approach. Again, the findings were striking: a 1% increase in electricity consumption leads to a highly statistically significant 0.26% increase in the value of manufacturing output, even after accounting for other relevant inputs. Hence the relationship isn’t just a statistical coincidence, it’s economically meaningful and causally robust. 

Unlike most macroeconomic indicators, electricity data is available in real time, is objective and less prone to misreporting. And because it can be disaggregated by geographic region or industrial consumer category, it provides granular insights that are ideal for identifying which districts or sectors are struggling and which are progressing. For example, imagine being able to detect a production dip in the textile hub of Tirupur or a slowdown in auto-parts manufacturing in Pune not a month later through the IIP, but within days based on electricity consumption patterns. 

This kind of responsiveness could fundamentally change how we support and promote manufacturing activity, including managing supply chains and preparing for shocks. The IIP, being a national indicator, fails to capture industry-specific trends at the state, district and more granular levels. 

Also Read: Manufacturing versus services: Why privilege one over the other?

India already has the infrastructure to make this transition. What we need now is to properly classify electricity consumption data by type of industry and institution and integrate it with decision-making at the most granular level. The Central Electricity Authority (CEA) already lays down guidelines for information reporting for distribution companies in each state. It should also collect information on the institutional status of each industrial consumer.

We could use such identifiers as the Company Identification Number (CIN) issued under the Companies Act, Factory Serial Number obtained from the chief inspector of factories, Shops and Commercial registration number for partnerships and proprietary entities registered with the state government, and even the GST number, along with the entity’s nature of activity as categorized under the National Industrial Classification system of 2008.

A dashboard showing daily-updated electricity data that integrates IIP as well as GST metrics and also uses the Geographic Information System (including satellite feeds) could revolutionize how we track and support industrial growth in India.

The authors are, respectively, former director general, ministry of statistics and distinguished fellow, Pahle India Foundation; and associate fellow, Pahle India Foundation.