Two figures for the same month, both correct, both from official sources, differing by more than two percentage points. That is not a contradiction to be resolved; it is what a rebasing looks like from the outside, and India’s core-sector data has just been through one.
The break, precisely
On 20 July 2026 the Office of the Economic Adviser at DPIIT released the Index of Core Industries on a 2022-23 base and discontinued the 2011-12 series. Four things changed at once. Iron ore entered the index, taking the count from eight sectors to nine. Steel moved from net to gross production data, to match the Index of Industrial Production. Coal middling and washed coal came out, leaving raw coal alone, ending a double count. And the weights were rebuilt from the IIP 2022-23 weights, pro-rated to 100.
The Department published a linking factor of 1.47 for the overall index, defined as the geometric mean of the monthly old-series index for 2022-23 over the geometric mean of the monthly new-series index for the same year, and left the sectoral factors to be derived from the same formula. It also placed the new back series from April 2023 on its portal.
New series, base 2022-23 : What the instrument now reads
| Period / Index | Index Value · Growth |
|---|---|
| ICI, June 2026 (provisional) | 119.6 · +5.0% |
| ICI, May 2026 (final) | 120.1 · +3.2% |
| ICI, June 2025 | 113.9 |
| April–June 2026 (provisional) | 118.3 · +3.6% |
| April–June 2025 | 114.2 · +1.0% |
| Annual 2025-26 | 115.2 · +3.0% |
| Annual 2024-25 | 111.8 · +4.3% |
Three consequences, costliest first
First, any series stored before 20 July 2026 is now the wrong series. A dashboard that pulls the latest core-index print and compares it to a stored 2011-12 level is comparing two instruments. The error does not announce itself; it produces a plausible number.
Second, the sectoral linking factors are not published and must be derived. The aggregate factor of 1.47 does not apply to a sector. Anyone splicing a steel or coal series across the break must compute that sector’s own factor from the geometric-mean formula — and because steel changed definition from net to gross and coal changed composition, a spliced series in those two sectors carries a definitional break as well as a base change. That is a caution, not a prohibition, and it belongs in a footnote of any note that uses them.
Third, the release day has moved to a fixed date. The provisional index lands on the 20th of the following month, or the next working day. For anyone building a data calendar, that is a firmer commitment than the old practice and worth writing into the calendar as a hard date.
What the weights tell an investor
The nine weights repay a minute’s reading. Electricity 30.932, refinery products 22.572, steel 17.584, crude oil 7.430, coal 5.596, iron ore 4.905, cement 4.410, natural gas 3.841, fertilizers 2.731.
Three consequences follow, and this desk has computed each from those weights. The top three carry 71.088 points — the index is, to a first approximation, a power-refining-steel index. The hydrocarbon-and-fertiliser block that contracted in June carries 36.574 points, so a soft quarter in that block will pull the headline down even when everything else is expanding. And iron ore, the new entrant, carries only 4.905 points but grew 43.9 per cent in June, contributing about 2.15 points of the 5.0 headline on this desk’s arithmetic — a small weight can dominate a month when its growth rate is large enough. An analyst reading the headline without the weights will attribute June’s strength to the wrong place.
Iron ore carries under five points of the hundred and delivered more than two of June’s five. Weight is not the same as influence.
For the data desk
• Replace stored ICI history with the 2022-23 back series from April 2023.
• Overall linking factor 1.47; derive sectoral factors from the published formula.
• Flag steel (net → gross) and coal (composition change) as definitional breaks, not merely base changes.
• Release date: 20th of the following month.
• Nine sectors now, not eight — check that field mappings did not silently drop iron ore.
A word on what a provisional figure is
June 2026 at 5.0 per cent is a provisional estimate; May 2026 at 3.2 per cent is final. The convention is standard and the distinction is real: provisional figures move as source data completes. A note that treats the two as equally settled will read a revision as a change in the economy.
Offered only as corrective advice, to help the work move faster: the Department has done the difficult part well — it rebased, it explained each methodological change in plain language, it published the linking formula and it put the back series on a public portal. What remains is small and would be disproportionately useful: publish the nine sectoral linking factors as a table, and mark in the back-series download which months are provisional and which are final. Both are one column of work, and both remove a class of error that will otherwise be made repeatedly, in good faith, across the profession.