Better estimation or overestimation?

A sharp upward revision in GDP growth in 2013-14 in the new GDP series released by the ministry of statistics and programme implementation (MOSPI) on January 30, 2015, has raised questions in some quarters.

Revision in the estimates of GDP is a continuous process. The availability of more updated data and changes in methodology which enlarge the access of data are the two factors that result in revision in the initial estimates of GDP. The base year of GDP is also revised regularly to have it aligned with the structural shifts that the economy witnesses continuously.

The higher figure from the new series seems at odds with other economic indicators such as the growth in bank credit, the Index of Industrial Production and corporate performance. The most significant change in the 2011-12 series is the use of MCA 21 data (from the ministry of corporate affairs). Earlier in the 2004-05 series, private corporate sector (PCS) was being covered using the RBI Study on company finances covering around 2500 companies.

According to MOSPI’s note, “In the new series, accounts of about 5 lakh companies have been analysed and incorporated for the years 2011-12 and 2012-13, while the number of common companies (companies for which accounts are available for 2012-13) is around 3 lakh for the year 2013-14”. Unincorporated enterprises belonging to households with complete sets of accounts have also been treated as quasi-corporations.

Usually, we compare the new series of national accounts with the earlier one by looking at the ratio of the two series in the base year. The NAS series with base 2004-05 which was launched on January 2010 was nearly aligned with the earlier series with base 1999-2000. In the new series (with base 2011-12) launched in January 2015, though the overall ratio of GDP (at factor cost) in the base year (2011-12) is 0.98, inter-sector differences are relatively large -- such as in the case of trade, hotels and restaurants and transport.

This is attributed to the new NSSO survey of unincorporated enterprises and the use of MCA 21 data set. This has also resulted in a sharp decline in the ratio of GDP from trade to the GDP originating from real sectors, comprising agriculture, mining, manufacturing and electricity.

Apart from the absolute GDP numbers, concerns were raised about the growth of GDP, both at the aggregate and at the level of sectors. Comparing the average quarterly growth of GDP/GVA for 10 quarters (until Q2 of 2014-15), we are not able to conclude that the differences in rate of growth of GDP (old series) and GVA (new series) are attributable to production taxes and subsidies.

For instance, the average of quarterly growth of GDP originating from manufacturing at 5.9 per cent as per the new series is significantly higher than a growth of just 0.5 per cent in the old series. MOSPI has said the more complete corporate sector database and greater use of value linked indicators as against IIP (a volume Index) and ASI, which is establishment oriented, helps better describe corporate value addition.

However, the average quarterly growth of trade transport and communication sector was higher in the new series for the 10 quarters (10 per cent, against 4 per cent) –- which implies that the spurt in manufacturing growth was even higher for the ratio between the trade GDP and the real sector GDP to decline during the period of the new series.

Besides the use of MCA 21, MOSPI has attributed this high growth to the shift to a value based indicator of sales tax. But the movement of gross trading index and sales tax index do not differ in any significant manner.

Therefore, we are not able to explain the higher growth derived from the new value based index.

Higher manufacturing growth is also somewhat unexplained. The asset–turnover ratio of Nifty companies which is a measure of capacity utilisation shows a steady decline from 4.41 in 2007 to 2.26 in 2014, a pointer to a deceleration in growth. Business confidence index and credit growth to manufacturing sector also do not validate the buoyancy as suggested by the new series.

Since the use of MCA 21 data base is at the root of higher GVA growth, it is worthwhile to look at corporate performance. However, the data set of the Centre for Monitoring Indian Economy, suggests that the manufacturing growth has continued to remain sluggish.

The MCA 21 data base in the new series has a much wider use in manufacturing, trade, transport, communication, mining and power sector. At current prices, the rates of growth GVA of these sectors and the corporate sector appears to converge in the most recent four quarters, though GVA growth in earlier period is higher. This convergence of growth rates in 2014-15 at aggregate level, though somewhat comforting, once again conceals wide differences across the sectors.

Differences in growth rates at sectoral level in the earlier two years, 2012-13 and 2013-14, are even sharper. Further, the overall higher growth in GVA originating from manufacturing and trade in the new series cannot be ascribed to a large extent only to the unorganised sector. The share of the unorganised sector in both manufacturing and trade has considerably declined.

R Nagraj ( EPW , March 28) has also raised the issue of inconsistency in the savings and investment data of the private corporate sector in the draft and final report of the sub-committee of MOSPI on the private corporate sector including PPPs.

We tend to agree with him that it would be appropriate to have an independent professional review and statistical audit of the entire set of procedures. We also recommend validation of this data set with the data set of the tax research unit of the department of revenue. The use of a larger data set (MCA 21) and shift to a direct method of estimation for the corporate sector is certainly an improvement. We do recognise improvements in methodology now as GDP is available at industry level, demand level and also in terms of factor shares.

The new series at the aggregate level is nearly aligned, though this association gets weaker at sector level. In 2013-14 and 2014-15, overall GDP growth and growth at the sector level shows a significance divergence, often at odds with the other indicators. We advocate additional efforts at data validation and data dissemination.

R Gopalan is a former secretary (economic affairs), Govt of India. Singhi is former senior advisor in the Ministry of Finance. The views are personal

COMMENT NOW