The New IIP Series: Necessary But Not Sufficient
The new series of the index of industrial production (IIP) which was launched last week with the base year of 2011-12 provided a much-needed update to a metric that was running on a 2004 base. While the Ministry of Statistics and Programme Implementation clearly states that the two series are not strictly comparable, such a comparison needs to be carried out to confirm that the older IIP estimates were not totally off the mark.
A comparison of the end output as well as comparing analytical methods of the old and the new series can give useful insight on how to calibrate reading of future IIP numbers. Such an exercise is not to establish technical accuracy or the lack of it, but to understand the relative correctness of new measures vis-à-vis the old, and to arrive at a reasonable consistency in measuring macroeconomic factors across time periods.
Divergence From Old IIP Trajectory
The new series shows higher growth than the old series in three-fourths of the monthly readings taken since April 2012. This can be explained in part as the new series includes items which are actively produced, as opposed to certain items in the previous series that were hardly produced on any scale. However, what may be difficult to explain is why the growth trends in the two series are significantly divergent from December 2015 to March 2017.
The new series has 809 items from the manufacturing sector as opposed to 620 in the old series. The weight of the manufacturing sector in the index is as much as three-fourths of the total. 149 new items have been added to the basket for the new series and 120 items from the old series have been removed in the interest of a more relevant reflection of industrial activity.
The author estimates that approximately two-thirds of the weight in the new series is attributable to items which were common to the old series. That would lead to an assumption of high positive correlation between the two.
That positive correlation was strong till the month of December 2015, after which the trend has reversed. With the information disclosed so far, there is no clear explanation of why this happened.
It is possible that the quality of the old series deteriorated with time and the new series is ‘the latest and the greatest’, hence the sudden divergence.
Could it also be that new statistical errors got unintentionally introduced into the new series which replaced the errors of the old series?
Another possibility is that the new items which got added to the new series may have more volatile productive capacity, and have, since December 2015, developed an entirely different growth trajectory than it had in the first three years of the new series. So one cannot rule out a structural break in the data even as the series is launched. It would be helpful if the ministry provides more data or shares an explanation for this trend.
There are three other factors which need to be understood to appreciate how the new IIP series can be read, which calibrate our understanding of the economy. These factors may also help explain the divergence between the new series and the old.
Capital Goods Adjustment: Implementation Doubtful
Capital goods carry around 8 percent weight in the new and the old IIP series. There has been criticism in the past about IIP data on account of the lumpiness of this sector. Units of most capital goods take more than a month to produce so there are months when the goods are not finished, and the IIP is low as a consequence. In months when the production is completed, the capital goods IIP number shoots up. This is a problem with no perfect solution.
However, responding to popular (which need not always be correct) criticism, the approach adopted in the new series is of calculating the operating work-in-progress of capital goods. This method is comparable to the accounting methodology used on the balance sheets of companies in the capital goods space. Companies typically calculate work-in-progress once a year but at a company level. Larger or well-established companies may track work-in-progress on a quarterly basis that is likely to be for internal purposes.
It is difficult to conceive how mid- and small companies will be able to calculate work-in-progress on a monthly basis, that too at a factory level, to furnish the data for IIP estimation.
In short, however well-intentioned the move of using work-in-progress may be, it may not meaningfully improve the quality of the IIP estimate for capital goods.
Factory Frame: Beware The Survival Bias
The new series has a higher number of factories, mostly to account for the additional new items. The Working Group for Development of Methodology for Compilation Of All-India Index of Industrial Production with Base Year 2009-10/2011-12 suggested that apart from factories covered under Annual Survey of Industries, factories covered by other sources may also be added to the new series.
The addition of factories in any new series may be prone to something called survival bias. The new series may have possibly taken companies which were functioning for all years from 2011-12 till date.
Which means that these are stronger than average companies, and are likely to exhibit higher production.
The standard global practice when a country measures its industrial output has been to replace closed factories with functioning factories and during the transition period, the loss in production gets reflected in the IIP number. However, when a new series is launched because of inclusion of stronger companies which remains open for most years around the base year, the production trend may be more positive than the rest of the economy where factories close down regularly. This may be among the reasons why the new IIP series is consistently above the previous series.
Monetary Value Of Production Versus Volume
IIP is expected to measure the physical volume of output in an economy, which is easier said than done. Do note that it is a globally acceptable practice to take the monetary value of those items where volumes are difficult to measure. In such cases, the monetary value is deflated by the relevant inflation measure to estimate ‘real’ production.
Despite such adjustments, it is empirically observed that rising inflation tends to give an upward bias to IIP to the extent it has a higher component of ‘monetary value’ measure.
In the new series, the production value of 109 items will be measured by monetary value - deflated by the wholesale price index (WPI). This number was 53 in the previous series. It is possible that the falling growth trend in IIP in recent months may have to do with the fall in the WPI, which got aggravated by a higher proportion of items measured in monetary value terms.
To conclude, much more effort and investment are required to systematically capture data in a timely fashion to give enhanced credibility to the IIP number. A data refresh or changes in the analytical process may improve the quality only so much, if at all.
Deep Narayan Mukherjee is a financial services professional and visiting faculty of finance at IIM Calcutta.
The views expressed here are those of the author’s and do not necessarily represent the views of BloombergQuint or its editorial team.