The revision cycle nobody explains
Every year, India's Ministry of Statistics and Programme Implementation releases an advance estimate of GDP growth, followed by a first revised estimate, a second revised estimate, and eventually a final figure — sometimes three or four years after the reference year. If you have noticed that the advance estimate is nearly always lower than the number that appears in later government documents, you are observing something real, not an optical illusion. Between 2015 and 2023, every single advance estimate for annual GDP growth was revised upward in at least one subsequent release. The average upward revision across those eight years was 0.6 percentage points — modest on the surface, but enough to shift a narrative from 'slowdown' to 'resilient growth' in the media cycle. The reason lies in the base-year methodology and the data sources used at each stage. Advance estimates rely heavily on extrapolation from a limited window of quarterly corporate filings and government expenditure data. By the time the final estimate is compiled, the statisticians have access to complete annual accounts from a much broader set of enterprises, including informal-sector proxies derived from the Annual Survey of Industries. The gap between those two data worlds — the early extrapolation and the later full accounting — consistently points in one direction. That direction is up. This does not mean the GDP figures are manipulated. It means the measurement system is designed for speed in the first release and accuracy in the last. Readers who report the advance estimate as settled fact are working with a provisional tool. Gospregledx will continue to flag the vintage of each figure we cite.