,
17 tweets,
6 min read
Read on Twitter
Was indeed a fascinating discussion, and quite a respectful exchange (the kind that we need more in 'New India')... the #greatIndianGDPdebate has certainly not ended, and my takeaways so far are the following #Thread #IPF2019 (1/n)
I thought @arvindsubraman (AS') presentation yesterday was much more sharper, more thought-through, and better articulated compared to his previous paper on this issue (2/n)
He made it clear that he was NOT re-estimating the GDP series but was merely trying to validate using 'demand' side metrics (investment demand, exports etc.) and the dissonance between these indicators and new series was sharp (3/n)
Pronab Sen (PS) argued that this dissonance can be explained by productivity (/technological advancement) and even if that is not showing up in profits, it is showing up in wages (4/n)
The problem with PS' argument is two-fold: One, it relies on the aggregate values from the same MCA-21 dataset that is being questioned by critics (and by an NSC report) for gaps and inconsistencies (see this: livemint.com/politics/polic…) (5/n)
Second, it does not seem to tally with either ASI or a cleaned up subset of the the MCA-21 database (Prowess) which we had analyzed sometime back, and which show wage compression rather than increase (livemint.com/Companies/X3Oz…) (6/n)
Sebastian Morris' presentation was an extension of his previous @IIMAhmedabad working paper (livemint.com/news/india/ind…) where he sort-of-extends the old GDP series forward to find it to be relatively lower than official estimates (7/n)
Morris' method by the way seems to come close to one of the three methods suggested by the Mundle committee on real sector stats set up by the NSC which was published last year in back-casting/reconciling the two series (mospi.gov.in/sites/default/…) (8/n)
I have some sympathy for Morris' view that in trying to expand coverage (at a point in time), the CSO may have ended up compromising consistency across time. Note that this is a technical critique, not an ad hominem one. (9/n)
There is one aspect on which Morris and PS agreed: that use of formal sector proxies to capture informal sector growth was problematic, and I am glad PS has come round to this view. It was a joint NSC-ACNAS comm chaired by him that had recommended this in the first place. (10/n)
When I had first raised this issue of inappropriate proxies to capture informal sector growth in 2015, I got a counter-reply from CSO saying that this is the best we have, and if you have a brighter idea, why don't you suggest a better proxy! (11/n)
I had to remind that person (a senior statistician, who shall not be named) that my job was to ask questions, and his was to provide answers. Manna (ex-CSO, now NSC member) too has a paper on this issue, which we had written about here: livemint.com/news/india/the… (12/n)
... in the morning session yesterday, which he chaired, he simply refused to engage with the questions on either data quality or autonomy raised by several scholars. His non-response to those questions, and his admonition to researchers.... (14/n)
... that they must not always think of 'bashing the government' but help improve the system does not seem to have gone down well. Several economists think they are only trying to help by pointing out the flaws in the current system that may be causing.... (15/n)
... the dissonance between official numbers and other indicators. If academic research is seen as 'bashing', and is to be conducted with a 'no criticism clause', we may as well turn universities into govt offices! (16/n)
Finally, I thought the organizers @ncaer did a great job in having a diverse set of presenters, and while I understand that it may take time to put up the recorded sessions, it would be helpful if they could put the presentation slides up. (17/n)
