What the Index of Industrial Production numbers mean for economy
The Index of Industrial Production (IIP) has contracted for two consecutive months now. Annual contraction in IIP was 1.4% in August and 4.3% in September. For the quarter ending September, IIP has contracted by 0.4% on an annual basis.
Are the IIP numbers an indication of what is to be expected in industry growth for the September quarter? The Gross Value Added (GVA) component of industry grew at 2.7% in the June quarter, while manufacturing grew at just 0.6%. IIP growth in the June quarter was 3%. With IIP growth turning negative, it is reasonable to expect that the industry component of GVA will grow at a much slower pace in the September quarter than what it did in the June quarter. A good way to test this hypothesis is to look at how industry growth and IIP growth have been related in the past.
Chart 1 plots quarterly growth in IIP and industry component of GVA since June 2012, the earliest period for which quarterly GVA data is available for the 2011-12 GDP series. As can be seen, the correlation is pretty weak. This suggests that caution should be observed in placing too much reliance on the IIP numbers to predict industry growth in the September quarter.
What explains this mismatch?
Ideally, both IIP and GVA in industry should show the same trend. In fact, this used to be the case until the GDP series was revised. This can be seen from Chart 2, which plots both IIP growth and GDP component of industry for the 2004-05 series. The correlation between the two variables is pretty strong.
These comparisons suggest that the current mismatch between IIP and industry component of GVA trends might be because of the change in methodology which happened with the change in base year. Among the biggest reasons for this could be the beginning of use of MCA-21 database in calculating corporate sector output for GDP statistics. A 2016-17 report of the National Sample Survey Office had found that as much as one-third of the companies listed in the MCA-21 database, were untraceable or wrongly classified, giving rise to suspicion that they could have contaminated the GDP numbers. This was first reported by Pramit Bhattacharya in a Mint story (see https://bit.ly/2NDPu1A for details). To be sure, IIP statistics have their own share of problems, especially high volatility (see https://bit.ly/33Mk8vq for details).
However, the fact that IIP growth and industry growth in GDP statistics had a better correlation in the old series than the new series shows that our statistical problems in using high-frequency indicators to predict macroeconomic trends have only increased. Given the fact that the Indian economy is in the middle of a protracted slowdown – GDP growth has been going down for five consecutive quarters – such lacunae in the statistical system only add to the problems of informed policy making.