Was demonetisation behind India's low growth? Hard to say: TCA Anant

What are the primary factors behind the deceleration in GDP in January-March?

One of the factors which has been influencing growth characteristics has to do with global commodity prices. Global commodity prices had collapsed somewhere around 2014-15. This had led to a fall in WPI (Wholesale Price Index), which had entered into a negative territory.

But, they gradually stabilised back into the positive territory and started rising. Because prices are used as deflators in GDP, growth in this period when prices were low, further went up. This price behaviour was something which every analyst had predicted would happen. It was expected that when prices start rising, it nothing else changes, growth would correspondingly come down as well. To that extent, part of the slowdown in constant price measurement is attributable to simply this long-run behaviour in prices.

When we talk about the slowdown being good or bad, we first need to filter, how much of it is attributable to purely long-run process. For the most part, what we are seeing is that our growth has not fallen off the window. It has dipped, but it would be more or less what you would have expected given this long-term price behaviour.

Also read: GDP expands at 7.1% in FY17, but slows to 6.1% in Q4 as demonetisation bites

Do you see the same trend going forward?

Commodity prices had started rising from the end of 2015-16. The rise that was seen in 2016-17 was because of a low base in 2015-16. In 2017-18 you are not going to see such a high rise in prices unless something happens globally. That will taper off, everything else remaining the same. The dampening effect of growth that was affected by price behaviour will also taper off.

Had it not been a change in the base year of IIP (index of industrial production) and WPI from 2004-05 to 2011-12, how would GDP numbers look like?

It is hard for me to answer that. If you look at the annual estimates of GDP since 2011-12, the effect of a change in IIP and WPI has been small. IIP is particularly muted as it enters GDP in a very limited way. In WPI, the difference in the old and the new series was small to start with. WPI as a whole is used in different segments in GDP compilation. Based on the WPI corresponding to that sector, when you aggregate it, its overall weight depends on the share of different segments. The overall impact on GDP was muted. Having said that, while its impact on annual figures were muted, quarterly data is more affected by the change in IIP and WPI data. IIP plays a bigger role in quarterisation of annual data, partly because we use IIP data when no other indicator is available. In the new series, IIP shows higher first-quarter growth (April-June) and tapering off later.

To what extent is the impact of demonetisation implicit in the GDP data?

GDP is impacted by a number of long-run processes. Commodity prices crashed in 2014-15 and continued to play along the way in GDP is two years later, simply because of the dynamics of the price index. Before you can argue that this is an impact of demonetisation, first you have to separate out all the long term processes out of the trajectory which you are seeing. You need to take out the long term process out and then try to work out things. Statistically, it is simply not possible at the moment. It will take a lot of time-series before somebody is in a position to answer that question.

When can we expect the new employment data to be released?

The fieldwork of that has started. We will be able to release the estimates after a full year of fieldwork is complete. The NSS (National Sample Survey) employment and unemployment data is for a full year. All seasonal attributes also need to be taken into consideration in the data. When we get one year's data, benchmarking and change in indices make sense only after a two-year exercise is over. Otherwise, you will not be able to distinguish between the long-term patterns. We will take out the first set of results in 2018.

When do you plan to release the GDP back series data?

The back series is a statistical challenge. What people are asking is can you re-compile GDP in the past which is statistically equivalent to the manner in which we compile it now. I need a statistical equivalent measure of citing corporate performance analogous to the large number of corporate data that I have now. I do not have corporate data readily processable for this sort of large volume in the past. Electronic data and MCA (Ministry of Corporate Affairs) database is available only from 2011 onwards. Can we project this into whatever that was available in the past, which is statistically equivalent?  That is a more difficult challenge as we do not have as yet large enough time series of full data set to be able to make an assessment going back.