Question mark over CSO’s independence

The recent release of the back series of national income has raised a series of controversial reactions. It is understandable for, the gross domestic product (GDP) numbers and the resulting growth estimates provide analytically useful, more powerful than commonly understood, perspectives encompassing not only economic but also social and political spheres.

To begin with, there have been critical comments on the nearly four years of gap between the release of revised 2011-12 series in January 2015 and now the back series in November 2018, unlike in the earlier series.

But it has to be realised that unlike in the earlier revisions, there have been many radical changes in the coverage of sectors as well as the methodology of estimation in the 2011-12 series adopted with a view to implementing, to the extent feasible, the recommendations of the UN System of National Accounts (SNA) 2008. These changes also have had their strong rationale in attempting to improve the data base of the Indian economy. In this context, the general perception that the data base of the Indian economy has been weak only because of the poor data available for the unorganised sectors, is obviously wrong.

We have noticed that the quality and coverage of data in respect of the most important of the organised sectors, namely, the private corporate sector, leaves much to be desired. We single this out for a critical comment. For this sector and for NAS purposes, hitherto only a small sample data was being used. Many authoritative reports had brought out how those sample data had been unrepresentative of the universe of the private corporate sector, at any rate since the early 1990s.

As is widely known, what the CSO has done for the new series is to widely expand the coverage in respect of the private corporate sector – a sample, or almost census, of 5.24 lakh companies (85%) as against 2,500 companies or so earlier (about 16%). By adopting this method, the CSO has achieved two valid objectives: to expand the coverage of companies and to shift to the enterprise approach from the ‘establishment’ approach adopted earlier. Again, by adopting the ‘enterprise’ approach, yet another milestone has been achieved, that is, to account for the production activities of head offices of enterprises.

This method has been adopted for manufacturing and mining sectors in the 2011-12 revision. Earlier, for the manufacturing sector, the Annual Survey of Industries (ASI) data were used both for the organised as well as unorganised manufacturing, whereas in the new series, for the organised private corporate (PC) component, corporate statistics has been used from the E-governance programme of the Ministry of Corporate Affairs (MCA).

Of the 5.24 lakh companies, only 1.35 lakh are under manufacturing and 6,721 under mining and quarrying. The rest of about 4.22 lakh companies are covered under construction and various services sectors. The kind of required information that the MCA data have provided for estimation of GVA for all these companies under all industrial categories, have not been available for the period prior to 2011-12 and hence, the back series have been desired for these components using the 2011-12 revised results as the bench-mark and the same have been back cast here up to 2004-05 using appropriate indicators. In the case of PC manufacturing, for instance, estimates have been derived using the growth rates computed from the ASI data of PC segment, keeping of course the 2011-12 estimates based on the corporate data derived from the enterprise approach, as the benchmark.

The relevant data for the benchmark year (2011-12) are: ₹7,47,638 crore as per ASI and ₹10,98,467 crore as per MCA (the former being 32% less than the latter). Treating the later as the benchmark number, the annual growth produced by the ASI manufacturing data will be applied to the above benchmark number and thus the new series will be built on it backwards.

ASI corporate manufacturing grew by 12.5% in 2012-13 and 9.8% in 2013-14, while the PC segment of the NAS manufacturing grew by 12.0% and 8.7% during the same respective years. But, in the absence of MCA data, the CSO obviously has had no other option but to opt for such surrogates. The SNA literature has strongly approved of using the benchmark-indicator approach which, on the face it, appears realistic. There have been a number of other changes too such as the effective labour input method in respect of unorganised manufacturing and a majority of the services sectors, expansion in the coverage of financial corporations by including a number of financial and capital market enterprises which were not included earlier, the treatment of the entire operations of the Reserve Bank of India as non-market and the value of its output to be derived as the sum of costs, unlike earlier when its functions were bifurcated as market and non-market.

For all of them, while adopting the methodology followed in the new base (2011-12) series, CSO has had to opt for a mix of methods due to the limitations of data availability such as benchmark-indicator described above or splicing method or ratios between relevant variables observed in the base year 2011-12 estimates – all of them seem to stay closer to the results of 2004-05 series except for the RBI operations.

Changes in Sectoral Shares or Growth

Table 1 presents the nature of changes brought about in the sectoral growth rates as between the earlier 2004-05 series and the present back series, as also the sectoral shares. The sectors which have lost in sectoral share or in growth rate are trade, communications, and financial services, where radical changes, introduced in the 2011-12 series, have been applied to the back-series estimation. In respect of the trade sector, the organised sector data have been revised in the new series from RBI data to the corporate sector data as per MCA and for the unorganised trade, from the gross trading income (GTI) to sales tax data.

For the communications sector, the indicator ‘number of subscribers’ has been replaced by ‘minutes of usage’ in the new series. As for the financial services, a major contributory factor has been the revision of RBI output which has resulted in a drastic fall in the new series and thus considerably undermined the implied GVA contribution of RBI to the economic system. These sectors have contributed to a decline in the overall growth rates in the new series.

Interestingly, a few sectors like agriculture and manufacturing have experienced higher sectoral shares in the back series though their growth rates have moderated. Notably, a peculiar reason for the agricultural share to improve as per the back series at constant prices is the improvement in the relative terms of trade for the sector. For constructing back series at 2011-12 prices, the price vector for 2004-05 has been replaced by 2011-12 price vectors which would reflect the higher terms of trade favouring the agricultural sector. As per the NAS data the deflators for the agricultural sector improved by nearly 100% between 2004-05 and 2011-12 as against the corresponding improvement of 54% for industry or 60% for the overall GDP. The reasons for the changes in the manufacturing share have been explained earlier.

Autonomy under cloud

Whatever may be the defence advanced by the NITI Aayog Vice-Chairman Rajiv Kumar, the fact remains that his involvement and his lead role in the release of the back series data along with the Chief Statistician of India, has undoubtedly become a blot on the autonomy, independence and integrity of the statistical system which the community at large had always taken for granted. This has not been just a pious hope. As late in May 2016, the Union Cabinet had adopted the United Nations Fundamental Principles of Official Statistics which is said to “bring professional independence, impartiality, accountability and transparency in methods of collection, compilation and dissemination of official statistics, besides adopting international standards” (Press Information Bureau, May 4, 2016).

All the 10 Fundamental Principles of Official Statistics, as endorsed by the UN General Assembly and adopted by the Indian Government, seek to achieve the basic trust in official statistics for the community at large. Against this background including the lead role played by the present NITI Aayog Vice-Chairman, the latest story revealed in the Indian Express of December 7, 2018 stating that the previous NITI Aayog Vice-Chairman Arvind Panagariya, had objected to the earlier CSO exercise on back series leading to ‘an upward revision of the growth rates for the United Progressive Alliance (UPA) years’ asserting that “we cannot allow it”, appears as one most damaging in this respect. The integrity and credibility of the CSO and its macro statistical exercise results thus stands badly dented. We must hasten to add that this in no way questions the CSO’s competence in statistical exercises.

In this light, we have reasons to suspect that the drastic reductions effected in the back series from the high growth rates of some years 2005-06 to 2007-08 or 2010-11 – experienced as per the earlier 2004-05 series, may have something to do with such arbitrary interventions by the powers that be (Table 2). Our suspicions arises from the fact that methods employed for the back series, namely, benchmark-indicator method, or splicing method, or unchanged growth rates, cannot produce such radical changes in the overall growth rates, that too, for some years and not all. For any keen observer of India’s growth performance during the last decade or so, if the economy grew at an annual average rate of 6.7% during 2005-06 to 2013-14, then any estimation of the average growth for the subsequent period cannot be of any higher order.

The writer is Director, EPW Research Foundation

COMMENT NOW