 Hello and welcome to NewsClick. Today we have with us Dr. Pranab Sain who has been the chief statistician of India and headed the National Statistical Commission. We are going to discuss the back series Pranab with you and the back GDP series which has been highly controversial at this stage. You had earlier defended the new method which the CSO had used for calculating the GDP and the fact that the GDP therefore changed from what was the old basis of calculating GDP. Now you have started calling this the latest back series which has been produced as a DTIO series. Now why are you sort of shall we say changing your views on the method or the back calculation in this case? Well you know I haven't really so much questioned the method. I have questioned essentially the modality through which it appears to have been done. Nithya Iyog seems to have played a rather important role in it. There is no justification that other than that that I can think of for Nithya Iyog not only sharing the same stage but actually taking the lead in the press conference. Nithya Iyog has nothing to do with the statistical office of the government. In fact the whole idea of inventing the National Statistical Commission was to provide a buffer between the rest of government and the statistical system. Which is it has to be independent. It has to be independent of the rest of the government. And that's because you don't want the policy to be informed by shall we say the usual things that happened the government to make the government. Well you see if you think about evidence based policy it should move from evidence to policy and not from policy to evidence. And you think in this case this will happen. Well you know that is the suspicion which arises and it's a very unfortunate suspicion because we spent since the 1950 we've been working assiduously at maintaining a distance between the statistical system and the political system. I think just with one stroke that has been damaged and damaged quite seriously I think. So this is an institutional damage of the kind that afflicts both the RBI and the CBI at the moment. Well you see as far as RBI and CBI are concerned they are both in essence instruments of the government which means instruments of the political system. So political involvement in that I think is understandable one can debate it at length. This whole business of RBI independence is really a fairly recent invention. But independence of the statistical system has been there right from the beginning and for good reason that is if the data that you are using has been doctored the any policy that you are going to make will be driven purely by political considerations rather than by evidence. Now coming to the back series it seems to have quote unquote deflated the good years as per the old GDP series is concerned and also seems to have brought up Mr. Modi's period in so far as a comparison can be made the comparison seems to go all one way making the current period look better and making the say the previous period look worse. So is this just methodology or is selective? Well let me put it this way what has happened is that the current period has remained exactly what it was that hasn't changed. It's the past which has been changed so if you look at the data you can break it up into three parts which are of relevance which is 2004 5 to 11 12 11 12 to 14 15 and 14 15 and after. 11 12 to 14 15 and 14 15 and after have not been changed at all they are exactly what they were earlier what has been changed is the 4 5 to 11 12. Now that is the back casting problem. The question is that what is the appropriate methodology to use for back casting and how do you what does one go about thinking it through? So the argument has been that the MAC 21 basically the corporate sectors returns which were changed what they have to submit gave a whole bunch of new data and therefore held the change in methodology but this data didn't really exist for the past and therefore the problem of trying to estimate a back series when the forward series in some sense has been calculated based on a new source of data. That's right. Now Professor Mandel has done a different kind of back casting than what the CSO seems to have done as you said under the leadership of the Nithya Iyoga unfortunately but Mandel's calculation was really based on essentially econometric methods so it was simpler and tried to do something which was in that sense made economic or econometric sense. How is this method? Let me let me try and sort of explain this in a little more detail because I think it's necessary. See if you think about GDP growth it can come about from four broad factors the first is new products and services which didn't exist earlier. Not captured therefore. No didn't exist therefore not second is the growth in production of goods and services that exist earlier and it continue to exist now. The third is the productivity of producing the existing goods and services and the fourth is improvement of quality. So there are four broad causes why GDP growth happens. If you think about these four in terms of the indicators that pick them up those are very different. The physical volume is the relatively easy one. You're actually just measuring you know the number of cars the tons of steel the tons of grain and so on. For which we have statistics always. Always have good statistics. As far as new products are concerned everything depends upon how your statistical system is able to pick up a new good. Is there a methodology where if a company comes in producing a new good you pick it up. That's not a foolproof area. It's reasonably foolproof for manufacturing for services not at all. Now as far as productivity is concerned productivity changes have always existed and we've been measuring it in the past but it has always been done in the years where the base has been changed. So you estimated productivity in 2014-5 you estimated productivity in the middle you didn't in the middle when you were for going from 4-5 to 11-12 you're assuming productivity state constant. Then 11-12 comes and you find a much higher figure and much of that growth is because of productivity and quality improvement. So the story that comes out from the 11-12 4-5 comparison is that 11-12 shows much higher productivity than 4-5. So the back series should have to take this into account. Now this is the thing. Now if you go so let me now go to the way back casting is done. The one thing you cannot change is the estimate of 4-5 and 11-12 at current prices because you're actually measuring it in that year. So that is fixed. If you look at the new back casted series the current price series is almost exactly the same as it was. It's almost exactly the same because the old series when we had estimated forwards had overestimated the current price. So when you got to 11-12 we found the GDP at current prices 2 percent below what had been estimated by the old series. So a 2 percentage point correction had to be made that's been made otherwise it's perfect no problem. So that's the current price series. The real trick is what do you do with the constant price series. 4-5 is at 4-5 prices. 11-12 is at 11-12 prices. Now when you're doing back casting you want 4-5 at 11-12 prices. Now there are three ways of doing this. Three broad ways there are more. One way is to say I will take the 11-12 price index and I will simply divide the current price series by those price indices and get 4-5. That's one. The second which is what they have they said they have used is you take the volume growth over the period and you go backwards. So let's say 10-11 to 11-12 let's say was 5 percent growth. So you take the 11-12 and deduct 5 percent from that and then you work back. The big difference between the two is that whatever productivity differences are happening in the first method if you go by prices the entire productivity difference is attributed to production. If you go by volume the entire productivity change gets attributed to prices. What the Mandel committee did is the third way of doing it which is to say that you've got all these non-volume sources of growth happening you know new products, quality, productivity. We don't know exactly when they came into existence. So what we do is we assume that these started at some distant past. So they chose 93-94 as the distant past and they essentially did a smooth curve fitting between 93-94 and 2011-12 of the difference not the total but the difference and then just added it back to what the original numbers were. So that's a hybrid between those two. That assumes that the productivity or the quality improvement would have taken place over a long period of time. Over a long period of time. And not a sudden adjustment. Not a sudden adjustment. So if you actually think about it each of these things are appropriate depending upon what your take is on that particular sector. Now in certain sectors you may not have had very high productivity gains and this you actually know from the data that you have in which case the volume index is the right way to do it. But if the bulk of the growth is coming from productivity increases then that's the wrong way to do it. Or new services. Or new services. It cannot be done. So there has to be a judicious selection. Applying one method across the board is a bad idea. The Mandil method which is again it's the same method for across all sectors has the advantage that all you're doing is redistributing over time. You're not taking a call on anything. You're saying I'm observing a difference. I assume that this difference in 11, 12 didn't exist at let's say 93, 94. And I'm just distributing it evenly over this period. So that's where we are. So at least it is a cross check on the shall we say that any other method you use because if you differ very significantly from what this is doing then you could be accused shall we say at least conservatively of having been selective in your choice of method. Yes absolutely and you know this discrepancy is rather stark. So if you look at what the Mandil committee had estimated for as the constant price GDP in 2004-5 and you look at the new series that CSO or Nitya Ayog has come out with for the same year the latter is 11 and a half percent higher. 11 and a half percent higher. That's a lot. So now if I take that 11 and a half percent and distribute it over a six-year period you're going to knock off a percent and a half each year compounded. Right. Which is exactly what you are seeing. So if we do not want to shall we say attribute motives just for the sake of argument then what could be what we see is really try and take different sectors and then select a method which is favorable to what you want to say at the end and as we know that this is the classical overfitting problem as it were you already have something in mind and you then choose the statistics that support it could we say that this Well that's one the alternative explanation is that perhaps you didn't take the trouble to really go sector by sector and ask what was the what were the principal determinants of the growth was it value but why was it productivity was it quality improvement was it not. The question comes up to why would they not because you have headed the CSO so you know how it works. Well it depends look you know this this thing has been going on for for many years now see 2015 the new series got released in I think 2016 CSO had a set of estimates this was discussed in the Niti Ayuk under the previous vice chairman who said that this would he would not allow this data to be published. This is under Professor Pankaria. And he said I am not going to 2016 so this was already the back series calculation had already been done but not this one not this one okay so it then went into hibernation it came out of hibernation because the national statistical commission set up the Mandel committee and Mandel committee actually came out with it and this became came into the public domain okay so suddenly you had a you had a set of figures and if my memory is right more or less matched to what CSO had come up with in 2016. This is not your time. I was that was the fag end of my tenure as chairman of the national statistical commission so before I had a time I I admitted office. Now this happened what two months ago less than two months back the Mandel committee report so now there was pressure because the congress had already taken this up and politicized it saying that you know we are just so much better than you guys you know we had double digit growth our time. So within two months this new series gets generated now the thing is that we are talking about a four year period no two year period people have changed now whether CSO had the time to again go back and look at it sector by sector and take a call on which is the appropriate methodology or did they say let's just go to the extent possible with a volume correction that's an open issue only an insider would be able to tell but my suspicion is that that's the easiest way to do it and they may have just done that and easiest way to maybe satisfy what the powers that be wanted and that's why well let me put it this way if you have a growth process which has been driven by productivity and quality improvement and new products then using a pure volume approach would almost inevitably give rise to much lower growth rates and that's what we see and that's what we see the last question you know that you said that pangaria stopped the estimates from coming out but was it within its jurisdiction to do so or was it in that sense the CSO accepting something that should have been accepted no you see the thing is that the it is entirely within the prerogative of the government and it should be about release of data you can or cannot release or the timing of the release that is within the purview of the government and that's how it should be because ultimately you're releasing it as government what is unacceptable is government getting into the production of the data so what panagariya did in a technical sense is okay he said i will not allow this to be really okay he didn't say change the data but you know you know nithya iog is also supposedly an advisory body it's really not the government today of course unlike the planning commission you know so it's in fact the governments think tank if you will unlike the planning commission which had a role which was far beyond that of just advice so in this sense it again goes beyond what supposedly is a nithya iog role well you know the nithya iog role isn't very clear i mean you know i think statements have been made that they are the single largest user of this data i would probably contest that because i have yet to see any nithya iog publication which has used gdp data in any meaningful way and today any investment banking house uses gdp data possibly far more intensively than nithya iog does so if you go simply by the the heavy users of the gdp data just bring in the investment bankers they'll tell you what to do so what you're saying is nithya iog really is not today the primary agency which looks at look nithya iog doesn't have a model i have not seen any statement which has been made on the basis of an analysis of the gdp data all of that had it been done so if there was in house sort of ability to handle the data to be able to identify anomalies in the data there would be some excuse it's still not an excuse because the nithya iog may be advisory but is inherently a political body appointments are political appointments they are not civil service appointments it's also interesting that rajiv kumar i'm not going to contest his economic credentials but is sort of taken upon himself the role of being the chief defender of the government on economic matters and contesting shall we say the leading economic figures in the world today including amr tussan and saying they don't understand the economy well you see if you think about it as the head of the government's think tank that's a perfect legitimate position for him to take because after all he does had what the government has acknowledged as the think tank now as i said that also requires you to have established your credibility in the eyes of the general public as a think tank now forget the tank at least thing yeah so well well the tanking may have happened as well but you know but that that's that's at the heart of it you know in the planning commission for instance and that's the big difference the planning commission had i think every right to say we're the largest user of statistical data you had a large-scale model we were running every policy through those models so and these were being updated every time a new data point came out so in planning commission you had in-house understanding of the data i would be surprised if nitya yok has anybody who has that understanding including the chief tanker okay let's get that going thank you very much pranab to be with us and we will continue coming to you for explaining what data and statistics means for the indian economy and for all of us sure this is all the time we have today for this episode of news click please do watch news click we'll continue to follow this and other issues