Base changes happen periodically, but why has it become such a big issue this time? What was the problem getting back series data?
A: You change base because the structure of the economy is changing. Though the structure is changing daily, you cannot adjust it every day. So you take a time series every 10 years and change the base. It happens globally.
However, when you change the base precisely to capture this change, the absence of a back series makes things difficult for anyone interested in long-term analysis for policy purpose, market projections, etc. In the absence of back series, they are forced to use their own macro modelling. For the past two years, everything was held up due to absence of back series.
When I had requested for the back series, I was told by Mr (T.C.A.) Anant (the previous chief statistician) that they would come up with the data in two or three months. This kept on happening for over a year. At the end of last year, when he was anyway going to retire, he admitted that he would not be able to give the data.
I don't know what the problem was, but I would think it was because the MCA (Ministry of Corporate Affairs) database didn't exist before 2006-07.
You can go further back using CMIE data, but these are very small samples compared to the MCA database, which has audited accounts of lakhs of companies.
The database capturing the Indian economy is improving in many ways, and this (MCA data) is one of the major improvements. But we could not find the back series so we had to find an alternate way of doing it. That's what our committee has done.
There is a lag between when listed and large companies declare their results and when the smaller, unlisted companies do so. How do you reconcile this?
In the MCA data, there are two main limitations. Since it is balance sheet data and consolidated, you cannot tell how much value addition happened in each state (even for bigger companies). The annual survey of industries (the backbone of earlier database) is establishment-based - you have a list of establishments in different states, cities and towns. From that you can do a region- or state-wise value addition calculations.
Another problem is that one cannot do a straight-forward sectoral analysis. The data doesn't have sectoral distinction the way the national accounts data has.
The other lacuna is that even though it has a lot of data on lakhs of companies and it is far better than anything in the past, it still lacks (data on) small enterprises. This has always been a big black hole. You don't know what is happening in half the economy.
How do you tackle this issue?
The NSSO (National Sample Survey Office) does enterprise surveys. First they did this for services, and then for industries, and now they have combined the two. That has become a very important tool for capturing informal sector data.
Traditionally we have been using employment data as a proxy for informal sector data in the national accounts. But since 2011-13, NSSO survey on employment has been discontinued. Now periodic labour force surveys take place. The statistical establishments plan to do these surveys four times a year for the urban sector and once a year for the agriculture sector. We will get the first quarterly numbers in October and the annual numbers in December. Between the two - the labour force survey and the enterprise surveys - we can get a good fix on what's going on in the informal sector data.
While the back series confirmed that GDP during the UPA government grew faster than it was believed to be, it also revealed that the growth composition was different. How did that happen?
In the new series, there are many changes. The biggest one is the MCA database, which is why the back series never happened. In the process of changing, some components have shifted from services (broader group services) to manufacturing. So the weight of services has reduced, impacting the growth rate. When some services that grew slowly were shifted to manufacturing, its growth slowed and that of services increased.
The committee has raised a concern on the quality of agriculture data. How good is that data?
Traditionally, we were dependent on the Anavari system, which was based on the local patwaris guesstimate. Then crop cutting survey was started by the NSSO. A similar survey is also done by the Bureau of Agriculture Cost and Prices based on data from all over the country. Then there is remote sensing. This may be a technical leap but interpreting satellite data is not simple. We have two-three different sources of data but ultimately some judgement would be involved.
Besides, there is also a political angle. States as well as the centre collect agriculture data. These data invariably have differences. Whose would you trust more?
What is your view on the debate that Gross Value Added (GVA) data is a better reflection of the economic growth than GDP data?
It depends on your purpose. If you want to focus on the actual income in the country, then I would look at the national income data, which is at market price. But if I want to compare sectoral performance, then I need to look at GVA data or GDP at factor cost.
What's your view on the dispute between UPA and NDA on whose growth was better?
There is no doubt that in UPA-I, GDP growth was higher than during UPA-II. In the first four years of NDA-II, the growth rate is lower than during UPA-II. But this in itself is neither a compliment nor a criticism of the government. Growth depends on so many factors. In 2008 there was a big crisis, and the government gave fiscal stimulus. P. Chidambaram had admitted in an interview that maybe the third stimulus was not necessary, and that it was more expansionary. In comparison, this government has been very conservative on the fiscal front.
There are different aspects to GDP growth, and political parties will choose the narrative that suits them.