June 29, the birth anniversary of Prasanta Chandra Mahalanobis, is National Statistics Day.The former Reserve Bank of India governor, Raghuram Rajan, recently questioned India’s economic growth narrative, arguing that weak corporate investment and slowing foreign capital inflows were difficult to reconcile with official Gross Domestic Product (GDP) figures showing the economy expanding at more than 7%. Leaving aside the current investment scenario or its lag-factor with actual production, it is worthwhile to examine the growth aspect from the perspective of measurement. We need to look, particularly, at the backdrop of two underlying and major methodological changes – double deflation and non-agriculture unorganised segment methodology – of the current estimation process.Double deflationOn February 27, 2026, the Ministry of Statistics and Programme Implementation (MoSPI) released GDP and other related aggregates on the updated base year 2022-23. While the use of double deflation has been cited, inter alia, as a reason for data revision vis-a-vis the old series, it is also stated that the constant price estimates would be revised after the base-change of Wholesale Price Index (WPI)/Producer Price Index (PPI) and Index of Industrial Production (IIP), which were then under process. WPI and output PPI on base 2022-23 have since been released by the Office of the Economic Adviser. However, the monthly trial input PPIs (Goods) are being published in respect of March 2026 onwards only for manufacturing sector and that too on experimental basis, so that “…the publication of Trial Input PPI would enable the Department to examine the data quality and also receive feedback of stakeholders and users”.This is somewhat confusing. As the input PPIs, even on experimental basis, are available after March 2026, we need clarification on whether output deflators from the WPI series – old or updated – were applied in National Accounts or whether some internal data has been used with regard to inputs for double deflation for 2023-24 and 2024-25. Also, in contrast to output PPIs, the sub-component weights of input PPIs have not been shared so far. Since the computation of real value added is quite sensitive to the choice of deflator, it was hoped that input PPIs would have been put through the rigours these warrant as deflator in National Accounts. This is particularly relevant for manufacturing where input-to-output ratio was quite high and even minor variations in input deflator may result in significant changes in Gross Value Added (GVA). Whether these experimental input PPIs possess requisite precision awaits the attendant details. Prima facie, we feel that it was not advisable to base double deflation on trial input PPIs.In any case, the experience of adopting a new data set, that is the MCA21 in 2011-12 series, whose repercussions on the GDP estimates continue to be debated even today, must have caused circumspection in adopting an experimental index for assessing the economy’s true growth trajectory. In the specific context of manufacturing growth, it is observed that the nominal GVA estimates show minor revisions, vis-à-vis 2011-12 series, in all three common years 2022-23 to 2024-25. However, in real terms, while the change in the growth rate for 2023-24 between the two series is only marginal; 2024-25 growth saw a major upward revision from 4.5% in 2011-12 series to 9.3% now. Digging deeper, manufacturing growth in 2024-25 is seen to be mainly driven by its sub-sector metal products, which, though having a share of 15.8 % in 2023-24 in the GVA at constant prices, alone accounts for 39.5% of the manufacturing growth in 2024-25. In fact, change in GVA in first two years of the new series in Metal Products has been 55% – quite extraordinary in this period when compared to its volume change of 18.3% as per the updated 2022-23 IIP series. Since the requisite GVA details for the year 2025-26 are yet to be released, it is not feasible to comment on the sub-sectoral impacts after 2024-25. Conceptually, the main purpose of IIP is to provide a measure of the short-term changes in value added. However, since it is difficult to collect high-frequency data that accurately measures value added, gross output measures such as value of production or turnover data are more commonly used. But if the year-on-year trend of value added from National Accounts differs so significantly from that of output based IIP, what can be said of the intra-year output-GVA growth fidelity?Non-agriculture unorganised sectorIn most of the unorganised segments, in conjunction with workforce, GVA per worker (GVAPW) of enterprises having at least one hired worker (HWE) in Annual Survey of Unincorporated Sector Enterprises (ASUSE) has been used to estimate sectoral GVAs in the new series. This ignores Own Account Enterprise (OAE) segment altogether from computation. As the latter has significantly lower GVAPWs, the extent of likely overestimation due to the chosen procedure may be gauged from the accompanying Table.Table: Broad sectoral comparison of market units in ASUSE 2022-23SectorSegmentShare (%) ofGVAPW (Rs)EnterprisesWorkersManufacturingOAE87.362.675012HWE12.737.4189309TradeOAE83.969.0106034HWE16.131.0221453OthersOAE85.255.7125257HWE14.844.3204674Source: Computed from unit-level data of ASUSE 2022-23.Besides apparent GVA overestimation owing to use of higher GVAPW from HWEs only, and nearly 75% of ASUSE data remaining unused, there are interpretation issues also since, to smoothen observed fluctuations, it is now proposed to use three-year moving average of GVAPW from ASUSE to avoid sampling fluctuations, wherever necessary. Thus, depending on the segment(s) selected for taking the three-year moving average, changes in the yearly GVA estimates may fail to convey true health of the economy. Also, moot issue is whether the purpose of measurement is to capture real dynamics of the economy or control the so-called volatility. ConclusionThe contribution of the informal non-agriculture sector is likely to be overestimated wherever ASUSE results are used on account of ignoring the contribution of OAEs. In fact, due to application of the three-year moving average, users may not be able to analyse sub-triennial impacts of economic stimuli/shocks.In manufacturing, it is likely that untested trial input PPIs have been used for which no data is available in the public domain prior to March, 2026. More importantly, double deflation, applied partially, has created a varied mixture of sectoral estimates amenable to interpretation in terms of neither the erstwhile single deflation nor the true double deflation. Until complete details are shared in the public domain, there is no clear answer as to whether the GDP growth coming from the extant estimates is a consequence of the input indices available or the actual dynamics of the economy. Thus, defending or questioning the available GDP growth rates amounts to searching the proverbial black cat in a dark room.Sanjay Kumar retired as Additional Director General of the Ministry of Statistics & Programme Implementation. N.K. Sharma retired as Director General of the Ministry of Statistics & Programme Implementation. Views are personal.