The report of the Committee on Real Sector Statistics appointed by the National Statistical Commission (NSC) was available in the public domain (MoSPI website) since July 25. The sudden spate of media attention and the political debate that this report triggered from August 17 was, hence, surprising. The task assigned to the committee, among other terms of reference, was to link the old and new GDP series using a methodologically robust procedure. This had become imperative since the missing back numbers for the official new GDP series for the 2011-12 base had become a major impediment for time series research and evidence-based macro policy analysis.
The comparison of growth rates during different political regimes was never a consideration for the committee. It is quite understandable that the media has focused on such comparisons which in turn triggered a political debate. As a member of the committee, I found it gratifying that most of the professional commentaries and editorials welcomed the committee’s efforts in generating the back numbers for the new GDP (2011-12 base year) series. It must also be appreciated that the comments that followed from political leaders, including the former finance minister, P Chidambaram, and the current finance minister, Arun Jaitley, and others like Niti Ayog vice-chairman, Rajiv Kumar, mainly focused on the policy issues arising out of the back series without questioning the integrity of the processes followed in generating the series.
One exception is the article by Surjit Bhalla (‘GDP Report — misreading the facts’, IE, August 21) which implied that growth performance during the UPA period appeared to be better compared to the subsequent NDA period because of the methodology chosen for generating the new series. Such innuendo, if intended, is unfortunate.
By way of background, it should be pointed out that base years for GDP estimates are revised from time to time the world over to account for changes in the production structure of economies. Whenever the base years are changed, the statistical authorities also provide comparable back series that enable macro-policy analysis. In India, MoSPI has produced such back series whenever base years have been changed. This time it found it challenging to do so because, among other reasons, the new and much improved MCA-21 data, which is based on lakhs of audited corporate balance sheets, was not available for the period prior to 2006-07. Hence, no back series has been generated for the past three years since the release of the new series in 2015. Nevertheless, generating the back series was essential. Since, the conventional method of GDP estimation was not an option prior to 2006-07 in the absence of MCA-21 database, the committee had to consider the next best option. A back series based on the production shift method was finally presented to the NSC while possible alternative approaches were mentioned for consideration.
In applying this method, the committee ensured that while linking the old and new series, the growth rates of the old series were maintained as far as possible. This was achieved by incrementally adjusting the old series GDP levels over the period 1993-94 to 2010-11 such that the back series exactly matched the GDP level of the old series in 1993-94 and that of the new series in 2011-12. The new growth rates are derived outcomes from the GDP back series so generated.
One issue that Bhalla has raised is about the spike in growth rate of the sub-group “trade, hotels, transport and communications” (THTC) in the back series being sharply higher at 17.1 per cent (average) for the period 2004-08. There are three reasons for such a spike in the sub-sector growth rate. One, the old series also had a similar growth spike in this period, although of a smaller magnitude. Two, in the new series (2011-12 base), there is a change in the composition of the sub-sectors where some of the sub-components of the THTC group have been shifted to the manufacturing sector, thus increasing the share of manufacturing while decreasing the share of the THTC group, with consequent impact on sectoral growth rates. The compounding of the two effects results in a sharper spike in the back series growth than in the old series. Finally, the committee’s main goal was to achieve a smooth linking of the old and new series at the aggregate level. As those familiar with national accounts base year adjustments are well aware, smooth adjustments at the aggregate GDP level may entail some sharp changes at the sector level. In fact, the growth rate changes arising at the sub-group level for manufacturing and THTC because of the compositional change are in opposite directions and off-setting each other, such that the net impact at the aggregate GDP level is marginal.
In attempting to show that the change in sub-group growth rates between the old series and back series has a significant impact at the aggregate GDP level, Bhalla reduces the back series growth rate of THTC by 8 percentage points in each year of UPA-1 to approximately match the old GDP series THTC growth rate, thereby reducing the average aggregate growth for 2004-08 by 1.4 per cent per annum (I guess his resultant growth should be 6.6 per cent and not 5.6 per cent as printed in the article). In so doing, he, in effect, reverses the compositional shift between manufacturing and THTC in the new GDP series that accounts for the spike in THTC growth rates in the back series as explained above. Such arbitrary changes in the sectoral composition are exactly what the committee tried to avoid. The committee’s attempt was to link the old and new series as smoothly as possible at aggregate GDP level without tampering with the compositional and other changes introduced in the official new GDP series. If one were to take such liberties with sub-group composition of the official GDP series, then it is always possible to make such changes so as to generate any growth profile one likes.
Another issue that Bhalla has raised is the identical growth rates for GDP in both old and new series for the year 2013-14. The committee made no change in the level of GDP of the official GDP series or its growth for the period 2012-13 and 2013-14. The back series was generated only upto the year 2010-11 as was required to link it with the official new series from 2011-12 onwards.