The latest release of the second advance estimate of GDP for 2016-17 may have put him in the line of fire but Chief Statistician TCA Anant shrugged it off by saying that he is “constrained to use publicly observable data” and that one should be grateful that GDP is not compiled based on “people’s beliefs or anecdotes” because then it can go many different ways. In an interview with Aanchal Magazine and Anil Sasi, Anant emphasised that additional data will help in getting a better picture of October-December, the quarter of demonetisation, and it will be reflected in revised estimates next year. “Certainly, in many of these areas we will get more complete information next year in many ways. For example, capital formation, I’ll get more complete information when complete accounts are available,” he said. Edited excerpts:
It’s being said that downward revision in Q3 GDP for FY16 helped Q3 growth this year.
The revised estimate of last year is based on the additional data that became available after the provisional estimate was finalised in the end of May. In this connection, there are many additional data sources, complete accounts of the government but most importantly since you are talking about downward revision of Q3, one data source which is important is final estimate of agriculture. The agriculture ministry’s final estimate came in lower than what the advance estimates had been compiled for 2015-16. The agricultural estimate is for full year. It is distributed across the quarters according to a certain cropping calendar… Its effects have to be reflected because we have to incorporate the final estimates in the revised estimates. Those were done at the time of the revised estimates and whatever are the consequences of Q3, are on account of these data changes, but if you look at it in greater detail, it is principally in agriculture where this reduction has taken place. But there are changes in all segments, this was one of the major segments and the explanation lies there. What’s its consequential effect is on growth rate, is so to speak beside the point. And the fact that agricultural estimates are going to come in lower, if you reflect on it, it was known long before the discussion on the current series took place…The consequences and crop calendar is also known to people, so its implications also would have been decided then. We are simply purveyor of news according to a well described formula and well described methodology.
There is use of proxies such as tax for trade, cement and steel for consumption. For a disruptive thing such as demonetisation, does GDP capture it because we are just extrapolating from the signs?
The logic of the indicator is as follows. If cement and steel sales have gone up, they must have been used somewhere. If somebody wants to provide the counterfactual, then they should point to evidence of stockpiling. At the moment I am not aware of any assessment of that kind. The stories are much more conversational. My neighbour has stopped construction, or something to that effect. The fact is that I don’t know. I am unfortunately constrained to use publicly observable data and you should be grateful that GDP is not compiled based on people’s beliefs or anecdotes because then it can go many different ways.
Does it get more refined as we have more data filtering in?
Certainly, in many of these areas we will get more complete information next year. For example, capital formation, I’ll get more complete information when complete accounts are available. At the moment, we are assessing it from a limited set of indicators. But both for corporate capital formation and for government capital formation, both of which are significant chunks of total capital formation, by next year, we will have complete accounting data which will be much better for assessing capital formation. So certainly as more data becomes available, assessments will improve and that is why we have a system of revision. But I can only improve for which data is available. Anywhere else the indicators will continue to map this, unless somebody can suggest a better indicator.
IIP is being used for informal sector and IIP itself is a questionable indicator. There are problems with IIP and whatever the revisions were proposed for IIP have not happened so far …
That is underway. The revision of IIP because it is compiled from a large number of government source agencies, there is a process being followed and that will happen. It’s not an issue. The limitations of IIP are also well argued. I had pointed them out in earlier conferences and am not dissociating myself from that. At the moment, in the absence of something else, it is the best I have. When I get something better, I’ll certainly use it and that process is underway. However, somewhat problematic is to assert something which we have not seen. For all its limitations and directional characteristics, what you see, you know what it is and you also know what is the nature of the directional biases for the most part. So, I am not quite sure what people would like.
Credit growth after demonetisation has been negative and it has not been captured fully in GDP data …
That’s not true. It is included in GDP data compilation. We use credit, not by itself but in financial services, we assess value added by a composite derived from credit and deposits, and that has been used. In part, a sector’s performance, is a reflection of that growth.
The base year for GDP was proposed to be revised every five years. So it would be up for revision in 2017-18?
It should be revised but as yet, a decision on that needs to be taken as to what would be the base year. Let that process be complete. What will be the year will be announced when it is made. But, yes we should start work on it.
Had demonetisation not happened, probably 2017-18 would have been the new base year?
The choice will not be so much influenced by demonetisation but by data availability. We would ideally like to go as close as possible to five years but we have to commission a number of studies and you would like the studies to be as close as possible to the base year. It’s a decision which has to be based on hard assessments.
When 2011-12 was chosen as base year, it was decided to have back series. There is none so far.
We had said that we will attempt to make it because many people suggested that it should be useful to do it. It is proving to be a major statistical challenge because in the new series, one of the databases that we use from the corporate affairs ministry is proving to be very hard to develop an analog in the past because the time series data which is available for firms is for a much thinner subset. And trying to tweak that to reflect behavioral characteristics of the full set is not as easy as thought earlier. It’s turning into a challenging exercise. We will tell you whether it’s feasible or not.
A meeting was held with NITI Aayog and finance ministry officials recently on it …
Not with NITI Aayog specifically, we hold discussion with technical experts including members of our advisory committee of national accounts. It is a research exercise, like good academics, they will do one set of calculations and then point out loopholes in it.
So the back series is unlikely to be there in the near future?
Yes, it’s not going to happen that quickly.
Looking at the data for trade and manufacturing, is there a possibility of trade channel stuffing pushing up the numbers during the demonetisation period?
I don’t know because I can only talk about what I have numbers for. Outside of that the numbers are there. I am not in a position to speculate. I do not speculate on what lies behind the numbers. At the moment the numbers what I have are numbers on corporate financial performance. It shows they have increased sales. I also have information on trade which shows that trade volumes have gone up in certain areas because sales tax collections have gone up in a certain area. Other than this, what is lying behind them I don’t know. I will work only with numbers. Since at this stage we have only aggregate numbers, any number of speculations can be made.