Nothing strange about back series GDP data, says Niti Aayog's Rajiv Kumar
These high growth rates have now been revised downwards after applying the latest data sets.
Changes made by the Central Statistics Office (CSO) in the coverage and methodology of preparing the ‘backseries’ GDP data represent a recalibration of economic activity for compiling the GDP series from 2004-05 onwards with 2011-12 as base.
The gap between GDP as per the 2004-05 base and the 2011-12 one is seen to be increasing over the years. This is due to higher growth rates observed in the 2004-05 series, which was prepared on the basis of data sets available at that time. After using the new survey-based estimates for the 2011-12 base, and implementing the recommendations of the United Nations System of National Accounts (SNA), 2008, the gap between new and old-base estimates was 1.7% in 2004-05, which increased to about 3% in 2011-12.
This indicates somewhat inflated growth rates in the 2004-05 series. These high growth rates have now been revised downwards after applying the latest data sets. Estimation has been mainly used for the unorganised segment, where data comes through periodic National Sample Survey (NSS) figures. There has been depression in growth rates mainly in unorganised trade, communication and financial services.
The survey estimates of unorganised trade were available in 1999-2000 (NSS 55th round), and then in 2010-11 (NSS 67th round).
Those of 2010-11 were found to be lower than the projected estimates prepared based on the 1999-2000 survey and the General Trade Index (GTI). Using data from NSS survey results of 1999-2000 and 2010-11, using the growth of sales tax was found to better reflect the movement and behaviour of trade. In unorganised trade, sales tax index (STI) and the new series of wholesale price index (WPI) have been used for preparing the back-series, where the growth rates are lower as compared to the growth rates in 2004-05 series where constant GTI was used.
In the communication sector, in the old series, benchmark estimates were moved with telecom subscriber growth, where the growth was over 20%. In the new series, minutes of usage has been adopted, which better reflects this sector’s behaviour.
There was a change in the methodology of computing gross value added (GVA) of the financial services as per SNA 2008. The contribution of RBI has been computed by a non-market approach, as compared to considering the banking division of RBI as a market enterprise in the earlier series. Consequently, RBI’s (non-market) contribution has reduced.
The estimates now released differ from those prepared by the National Statistical Commission (NSC) Committee on Real Sector Statistics (CRSS) chaired by Prof Sudipto Mundle. The latter used a rather simplistic econometric model for estimating GDP growth for previous years while changing the base year from 2004-05 to 2011-12.
The methodology followed was essentially a ‘production shift’ approach to prepare estimates in current prices, and then applying a price deflator on each such current price estimate to arrive at the constant price estimate for the back-series. This is not according to the methodology recommended by SNA 2008, which suggests volume extrapolation as a preferred methodology for constant price estimates over price deflation.
It may be seen that the level estimates of GDP at current prices between the 2004-05 series and the 2011-12 series in the base year 2011-12 was different by about Rs 3 lakh crore. This difference was adjusted uniformly by the NSC Committee over a period of 18 years between 1993-94 and 2011-12 on the assumption that the structural changes in the economy were gradual in all the sectors. This assumption is not borne out for the period under consideration, given both the price and volume changes across sectors.
This method, though simple, was an academic exercise that had limitations and did not represent the sectoral changes in the intervening period. It is also worth clarifying that the CRSS was not originally mandated to estimate the back-series and, therefore, the issue of the government accepting its recommendation on this count is somewhat moot.
For the revised series, latest data sources like sales tax, new series of WPI and consumer price index (CPI), and new methodologies have been used for preparing estimates of GDP for the back-series, at the sectoral and sub-sectoral levels, both in the current and constant price series. For preparing the new series, CSO has used volume extrapolation method in some sectors, sector-specific price deflators in some others, and revaluation using base year price and year-wise quantity in some sectors.
So, the methodology used by CSO is not only robust, but it’s also in accordance with international best practices and UN SNA 2008 recommendations.