Postscript to India’s GDP growth controversy

Had a good chat for half an hour with a good friend. He drew my attention to the remarks by Pronab Sen, former Chief Statistician of India. Pronab Sen had said that Arvind Subramanian’s estimate of annual average 4.5% GDP growth was based on volume indicators and the ‘unexplained’ 2% to 2.5% growth could be due to productivity, quality, etc. Or, it could be an error, of course. That would be a fairer statement to make than to call it an overestimation categorically.

You can read the article in ‘Business Standard’ that carries the comments by Pronab Sen.

These further comments by Pronab Sen are interesting:

“The paper’s finding means that there has been an equivalent improvement in productivity,” he told Business Standard. “But this would also mean that the older series underestimated growth in the years prior to 2011-12,” he added, since it was more robustly based on volume indicators.

That is the problem. The debate over India’s GDP growth calculation would have remained apolitical and technical but for the latest revisions in the last six months that bumped up growth estimates for 201-17 and for 2017-18 and lowered the growth estimates for the period from 2004-05 to 2011-12 in a ‘backcasting’ exercise.

Those changes – which defied logic and did not conform to anecdotal evidence – have what pushed the debate into the political arena.

That is why it is all the more important to take the debate back to the technical realm.

Arvind’s work smells right

There are two people whose timing has been questioned recently. No, this is not about timing of cricket strokes by the batsmen – the timing of the impact of ball on bat. If it is perfect, there is no need to apply force. The ball speeds to the boundary. But, this is about the timing of their announcements.

Yuvraj Singh announced his retirement from cricket even as a World Cup cricket campaign was on. Sharda Ugra who had helped him write his story gently hinted that legitimate questions could be asked of his timing. One can say that he wanted a fraction of the attention on himself. Let us leave it there.

Our interest in this blog post is on someone else.

The other person whose timing has been questioned is that of the former Chief Economic Advisor, Arvind Subramanian. Earlier this week, he wrote an article in Indian Express summarising his findings on the econometric work he had done on India’s GDP growth estimates. His results suggest that, on average, India’s GDP could have been overstated by a magnitude of about 2.5% points every year since 2011 amounting to a cumulative 19-21% during the whole period: 2011-18. However, for estimation, he excludes data for 2017-18 and 2018-19.

He has been criticised by commentators on both sides of the political divide in India. Those who are opposed to BJP and Modi criticise him because he did not publicise this before the election. It might have strengthened their hands. Or, so they think. Unlikely. But, that is their grievance.

The pro-BJP and pro-Modi commentators are upset that this paper has spoilt the ‘feel-good’ air that they are basking in, post-election victory. There is also the unstated anger that AS’ results suggest that the over-estimation of GDP growth, if broken down on an annual basis, is more in the post-2014 years than in the pre-2014 years (up to 2011) because official growth estimates between 2011 and 2014 are lower than the annual growth estimates post-2014.

Since both camps are unhappy, he must have done something right.

In fact, one must applaud his timing. By releasing his study after the elections were over in India, he had shown himself to be apolitical. The only politically somewhat loaded statement he has made in the paper is this: “that also requires us to reject the more recent estimates for the post-2011 period because they may not simply reflect the technical changes.

The statement put out by the Prime Minister’s Council of Economic Advisors does not indicate an awareness of the spirit and the discussion that are needed. In the interests of the credibility of the government and that of the Indian Statistical System, the data and the methodology need to be put in the public domain, instead of hiding behind technical arguments.

The other charge that is laid at AS’ door is that he didn’t he do this when he was in CEA. Now, we don’t know, honestly, if he did not do this exercise inside the government. He might have given the output to them and it might have been rejected. That could be one reason why he left six months earlier than he needed to. Who knows?

But, let us accept his statement that he did not have the time and space to do so when he was the CEA. I assume that by ‘space’ he means ‘mental space’ and nothing else.

To a large extent, I can vouch for that, even in my limited experience of being the Dean of a Business School. One simply does not have the energy at the end of a long day to write serious blog posts. It is hard to catch up even with one’s reading let alone writing. My blog posts came down drastically since October 2018 – when I took up my assignment – for no other reason but that I did not have the time and energy to be as productive as before.

If some were to attribute it to political correctness, I can point out that even my private emails to my mailing lists came down dramatically let alone public blog posts.

So, I am prepared to take him at his face value that he did not have the time and space to engage in a serious data analysis and investigation. It is one thing to write casual op.-eds., and it is another thing to undertake a serious investigation of data and do some statistical analysis on them. After all, the Chief Economic Advisor was engaged in multiple other projects – planning for the introduction of the Goods and Services Tax, annual Economic Surveys, etc.

I have gone through Arvind’s paper carefully more than once. His conclusions appear reasonable. He has merely invoked the spirit of ‘difference in difference’ approach. The estimation method is simple statistical regression. Did the difference in methodology adopted since 2011 make a difference only to India? That is ‘two differences’ there.

His first two pictures – Figures 1 and 2 – are specific to India. They do no invoke cross-country analysis. In other words, he has not just engaged in an ‘unusual exercise’ of cross-country regressions. He has also analysed India independently.

I had done a similar exercise in 2016 itself. See here and here. These two papers show that the GDP growth figures do not accord with growth in many other indicators of the real economy. The analysis was done using several indicators and the conclusions were clear. I did not undertake an econometric investigation. So, my result was not quantifiable. It was descriptive in nature. I concluded – without any statistical estimation – that India’s real GDP growth rate – might be overstated by 1.0% to 1.5%. In other words, the new GDP growth data did not pass the smell test. They still do not.

Arvind’s points about import and export growth and GDP growth in the post-2011 period and the import elasticity of demand are as important as his point about growth in manufacturing not being correlated with growth in manufacturing exports.

One could ignore his cross-country and panel regressions and undertake a ‘India-only’ exercise but there is not simply enough data, if one used annual data and tried to estimate a structural break in the data either by splitting samples or by introducing dummy variables. But, the evidence is clear in his Figures 1 and 2 and in my works cited above. There is no point in being in denial about it.

Pre-2008, India experienced an economic growth boom due to massive capital inflows, credit growth, capital investment boom and export growth. Many other nations did so too. Post-2008, global trade volumes have slumped. Countries – developing and developed – briefly recaptured their growth rates of pre-2008 years due to stimulus. Some of the stimulus was unsustainable as India found out in 2013.

That is why most of the countries have struggled with economic growth since then. In fact, post-2014, leading emerging economies have suffered lower growth than pre-2014. Only India has been an exception in spite of the introduction of many potentially growth-unfriendly and demand-unfriendly structural but long-term positive and essential structural reforms. Does not pass the smell test.

Harsh Gupta has written an interesting and thoughtful op.-ed., in Indian Express drawing attention to the fact that India’s tax collections have grown robustly since 2014 and if Arvind’s growth estimates (mid-point of 4.5% between the range of 3.5% to 5.5%) were accepted, the tax/GDP ratio would be much higher. He is right unless the tax data were fudged too. Rather unlikely. Tax collections are realised in cash and the government has to transfer a portion of it to States too.

My responses to Harsh Gupta’s article are as follows and not just to his argument on Tax/GDP ratio:

(1) The tax collection could well be the problem. That the tax collection mechanism became so active and coercive might be one of the contributors to the private sector uncertainty and investment funk. Indeed, even many sympathisers of the BJP and the Modi-led government were of the view that ‘tax terrorism’ was pursued with greater vigour post-2014 than it was in the pre-2014 years.

(2) For example, between 2014 and 2016, when the price of crude oil slumped, the Government did not pass on the decline in the price of crude oil to the Indian public. Instead, it loaded up additional taxes on the crude oil and ensured that the final retail price of petroleum products was only marginally lower, if at all.

(3) Productivity improvements would show up in profitability data. They have not. Second, export growth would be better if there is a renaissance of productivity. That is why despite the slump in global trade volumes, countries like Bangladesh and Vietnam have done better than India with export growth.

Some friends have pointed to articles such as this and have asked rhetorically if these improvements are captured in GDP data. Two responses are in order:

(1) They should be captured in profitability data, eventually.

(2) Ceteris is not paribus:

(i) Even as GST related improvements are happening to the movement of goods in India, the power sector (Independent Power Producers) has gotten into trouble due to lack of raw materials availability and at the expected prices and due to low price realisation from buyers or non-evacuation of power at agreed prices.

(ii) Then, the non-banking financial sector has gotten into trouble and has stopped lending.

(iii) What has grown rapidly even amidst overall tepid bank credit growth is growth in credit to households including credit card or revolving credit. That is for consumption purposes. That is not the stuff of productivity improvement.

Only if other things remained equal and then GST-triggered productivity gains occurred at the margin, one can speculate on their positive impact on growth.

In the final analysis, it is important to note two things from Arvind’s analysis (and mine):

(i) Post-2008, emerging economies are not the same. It is futile to keep talking about the 2008 crisis as the North-American or the Atlantic crisis. Emerging economies – including India – are as, if not more, worse off as developed nations are. It is not just about India and about economic growth in the UPA or NDA years.

There is a need to rethink the facile conclusions about the momentum of economic activity and the balance of economic power shifting inexorably towards the East and prepare accordingly to negotiate and deal with the West.

Such realism is also necessary to tackle the impact of Artificial Intelligence and Robotics on employment, the challenges of climate change and the looming water scarcity. All of these threaten both economic growth and economic and social stability.

(ii) This cannot be viewed as a partisan political battle. There is much at stake. Instead of arguing about it, it is far easier and more effective (lasting effect) to make available the CSO data and methodology to independent statisticians for verification. Only then will the controversy end.

(iii) It is quite possible that some concerns were raised in the government by some about growth numbers. But, that they were not either heeded or smothered. That should not be the case. The lone dissenting or a different voice deserves all the respect, if not more, as the conforming voices.

The government could have easily asked the CEA or the Department of Economic Affairs to come out with such an exercise as soon as the controversy about GDP growth estimates erupted in 2015. That would have made both monetary and fiscal policies more responsive and reasonable in the last four years. The failure to accept a lower growth rate has actually prolonged the growth stagnation and even resulted in the previous government taking a chance with demonetisation, perhaps at the wrong time.

In short, Arvind Subramanian has done a rather useful service to the discourse on India’s official economic data. It is better to pay heed to the message even if its specifics are open to challenge. It is wrong, self-destructive and counterproductive to shoot the messenger.

Between economic pessimism and market optimism

Through Andrew Batson’s blog (yes, I caught up with quite a bunch of his posts today), I chanced upon this Greg Ip article written in January 2019 at the time of the meeting of the World Economic Forum. Greg Ip concludes that central bankers have to acknowledge that the world economy was not strong enough to withstand the return to a normal level of real interest rates and therefore, must do the following:

Nonetheless, central banks need to proceed carefully: in a low-growth world, a little bit of monetary tightening can go a long, and painful, way. [Link]

Yesterday, while searching for something on the net, I stumbled upon the ‘Weekly Global Economic Update’ put out by Deloitte. The latest edition (14th May 2019) is here. It highlights an important financial stability risk that raises questions over Ip’s preferred monetary policy:

The US Federal Reserve has warned about the risks to the financial system from the sharp increase in corporate debt.15 In the Fed’s periodic report on financial stability, it noted that “borrowing by businesses is historically high relative to [GDP], with the most rapid increases in debt concentrated among the riskiest firms amid signs of deteriorating credit standards.” It drew attention to the fact that, although the volume of corporate debt increased 4.9 percent in 2018, leveraged loans increased by 20.1 percent, thereby boosting the risk profile of the corporate sector. It also said that asset values are relatively elevated and that investors continue to exhibit a high appetite for risk.  ….. It said a downturn could hurt the financial system because of “the rapid growth of less-regulated private credit and a weakening of underwriting standards for leveraged loans.” [Link]

They have quoted from the ‘Financial Stability’ Report of the Federal Reserve Board.

This is the all-important question that monetary policymakers have avoided confronting. If one keeps policy loose for the sake of real economy vulnerabilities, how does one prevent that from feeding through to irrational financial market optimism that does not really square with the real economy weakness or stagnation?

Financial markets do not worry about the numerator (cashflows). They are happy if the denominator (discount factors) is low. In other words, it will worry about the numerator if stock markets are discounting mechanisms. They are not. In casinos, fundamentals do not matter. Only luck and liquidity matter.

With their ostensible purpose of supporting the real economy – for which there is very little long-term evidence – central banks end up lowering risk premiums in financial markets. That, in turn, boosts asset prices, turn them into bubbles, stoke wealth and income inequality (one way to boost short-run stock prices is to cut wages to pay interest on debt taken to finance share buybacks!) and makes the economy even more vulnerable. Then, commentators again call for low interest rates while all that it has done is to boost the financial economy while doing zilch to the real economy.

When will this farce end? It is inconceivable that either commentators or policymakers are unaware of the reality. Are they captured or what else is the reason?

Fear is the key

I had left the following comment below a Andrew Batson blog post of January 2019 on fear being the key to economic development.

This is very interesting. Applies to India as well. Whether it is ‘fear’ or ‘crisis’, the biggest pre-requisite for economic development is ‘fear’ or ‘crisis’. They go together. Two leaders stand out for invoking them successfully. One is Andy Grove of Intel. He used to say that the only paranoid survive. Another leader who invoked this and maintained a constant fear of being swallowed up or being overtaken in economic terms using it as a lever to propel his country into materially higher standards of living was Lee Kwan Yew.

In a way, both India and China are failing to harness the potential of ‘fear’ or ‘crisis’ by overstating their economic growth numbers. If only the economic growth numbers reflect the true growth, without being embellished (intentionally or otherwise) by statistics or artificially spruced up by orchestrated credit growth (as distinct from voluntary commercial decisions of lending institutions), then they would serve as a wake-up call for political and industrial leadership or even labour leaderships in the countries.

You are right, therefore, about the prospect of China succeeding with its environmental clean-up rather than with continued economic growth beause a sense of crisis (or, fear) is driving changes there.

India is failing to harness these fears either with respect to economics or with environment. The danger is that easier, off-the-shelf populist answers are being thought of (e.g., interest rate cuts).

India’s growth statistics – part 2

Pramit Bhattacharya has written a very high quality piece on what has cripped India’s statistical system. It is classic turf war. It is beautifully written; flows well and the chronological sequence is maintained well. It is accessible to the non-technical audience too.

Notwithstanding our reservations on the input-output tables of P.C. Mahalanobis, we must applaud his approch towards ensuring reliability and integrity of India’s statistical system. Angus Deaton’s comments too must be taken seriously.

Couple of very minor quibbles:

(1) On the role of Pronab Sen:

Earlier in the piece, he wrote:

The objections in both cases seem to have come from the CSI, who according to former NSC members, stalled efforts to empower the NSC. India’s first CSI, Pronab Sen, said he objected to the draft Bill, as he felt that a body working “outside the government” would never be able to get other departments to share data. The twin power structures in the new statistical architecture in fact left plenty of room for ambiguity and conflict, and contributed to the bad blood between the CSI and the NSC.


In March 2010, Sen modified the composition of the Standing Committee on Industrial Statistics (SCIS) without consulting the NSC, earning the NSC’s ire. “There has to be clarity on authority and accountability,” said Sen. “The CSI is accountable to the government, the NSC is not.” (The NSC is accountable to Parliament). Sen said he does not recollect the SCIS incident but added that he thought the NSC had no business to decide on the SCIS.”

But, when Pronab Sen became the Chairman of the NSC,

The NSC-CSI tensions were at their lowest ebb when the former CSI, Sen became the NSC chairman. Sen says there was no conflict with MoSPI when he was at the NSC.

In a way, Pramit has not fully explored Sen’s potential or possible role in stymiing NSC when he was CSI and then not ensuring systemic improvements when he was the Chair of NSC (he might have ensured no conflict due to personal rapport with his former colleagues at MoSPI).

(2) The second quibble is very minor. The correction I propose would have made the second sentence clearer:

“A draft statistical policy was released for public comments by MoSPI in 2018 that envisaged a larger role for the NSC, incorporating some suggestions of the Menon committee. However, that (draftpolicy has been scrapped and a revised policy is being prepared, according to two people familiar with the development.”

This appears to be the real issue:

NSC was supposed to be the Regulating body for MoSPI and CSI. CSI was the Secretary to both. No wonder NSC was stifled and stymiied. How could the regulated be the Secretary of the Regulator?! The NSC ought to be an independent regulatory body headed by someone of equal or higher stature than that of the CSI at the MoSPI.

So, it appears to be a turf battle. Why couldn’t the political mandarins see through these obvious design flaw in the creation of the National Statistical Commission (NSC).

Further, based on his article, I get the impression (not sure if it is correct) that the reservations with the GDP growth data seem to be more due to the CSO’s obduracy in not sharing its data and methodology and relatively less due to politics.

In other words, has the NDA II government botched its optics without actually perpetrating anything?

In any case, these are my personal views:

(1) GDP growth data remain overstated from 2012-13 onwards at least if not from 2011-12. That includes UPA II years. 

(2) Growth data for 2015-16 and 2016-17 are somewhat more excessively overstated. Many other indicators do not back the 8.2% growth in 2015-16 nor the 7.1% growth in 2016-17. The GDP data do not pass the smell test.

(3) Based on revised data, India is the only country that seems to have grown faster between 2014 and 2018 than between 2009 and 2013 among seven leading emerging economies (China, Brazil, Russia, Turkey, India, Indonesia and South Africa) as per the IMF World Economic Outlook Database, Oct. 2018 data on Real GDP growth.

(4) Downward growth revisions for the years ending March 2006 to March 2012 do not appear correct. The downward revisions appear excessive.

Given the change in government in 2014 and that these revisions are downward for the earlier regime, the suspicion of political bias is, prima facie, reasonable. CSO data and methodology transparency alone can dispel them. Failing which, the credibility of both the government politically and that of the statistical agency will remain under a cloud.

India’s growth statistics – part 1

India’s GDP growth statistics do not pass the ‘smell test’. That is the only real issue. Everything else is noise.

What NSSO had put out yesterday is that they could not find some of the firms that were listed on the MCA21 database. They may be misclassified or they may have been closed as the parent companies that set up shell companies may not want them anymore because the purpose for which they were set up might no longer exist.

So, the NSSO conclusions on MCA21 database with respect to companies listed as firms operating in ‘services’ is irrelevant to the issue of credibility of GDP growth data. It may under or overstate GDP attributable to manufacturing and services. But, aggregate level impact is likely small, if at all. Therefore, this story is a distraction and merely adds heat rather than shedding light on the GDP growth data.

(1) The issues with respect to GDP growth have been around ever since the new base year and the new methodologies were unveiled in 2015.

(2) The CSO has not put out its methodology and data in the public domain so that their methods could be replicated. So, suspicion remains.

(3) More recently, CSO revised growth figures for the period 2005-06 to 2011-12 based on the new methodology. That brought down the growth rates at constant prices for this period which was predominantly the UPA era.

Further, these new findings were released by NITI Aayog. That was completely unnecessary and poor optics management. Raises questions on credibility of the data. Why should NITI Aayog release it to the public? It is the job of the Chief Statistician.

(4) Further, growth rates for 2016-17 (which includes four months post-demonetisation) and for 2017-18 were revised higher.

So, (3) and (4) combined have raised questions on ‘partisan’ data. 

(5) Further, as I wrote in MINT on Tuesday, India is the only leading EM nation  (I looked at six others) that shows higher GDP growth rate in 2014-18 than in 2009-13 period. It is not credible. It does not square with other pieces of evidence.

(6) The other pieces of evidence are not just bank credit growth. I have looked at overall credit (bank, non-bank, External Commercial Borrowings, etc.), auto sales, home sales, mobile phone sales, corporate profitability, etc. Growth rates of upwards of 7% do not appear credible.

When we have formal businesses struggling with bad investments, when we have both banks and non-banks saddled with non-performing assets, when export growth is stuck at low levels, when informal businesses are struggling with consequences of demonetisation and GST and trying to adapt to these changes, it is not possible that the Indian economy grew at 8.2% in 2015-16 and 7.1% in 2016-17.

(7) Finally, in my view, all the growth numbers below do not pass the smell test – of being in accord with other indicators of the economy. The period below encompasses both UPA II and NDA II years. These growth numbers below are overstated:

March of   Real GDP growth in % (2011-12 prices)

2013         5.5
2014         6.4
2015         7.4
2016         8.2
2017         7.1
2018         6.7

Natural enemies and man-made enemies

It is my piece in MINT on Tuesday. I enjoyed writing this but it was also one of the most difficult pieces to write. I spent several hours on it. But, the end result is gratifying.

A few weeks ago, I saw a news story that the eco-sensitive zone around the Bannerghatta National Park would be reduced by 100 sq km. This news was covered in a small way in the national newspapers. Since then, a campaign has been mounted to prevent this proposed reduction from happening. This story reminded me of an interaction with Meghna Krishnadas of Yale University early in November.

In the paper Weaker Plant-enemy Interactions Decrease Tree Seedling Diversity With Edge-effects In A Fragmented Tropical Forest, written with Robert Bagchi, Sachin Sridhara and Liza S. Comita (Nature Communications, Vol. (9), article number: 4523 (2018)), she tested the hypothesis that natural enemies—insect herbivores and fungal pathogens—help shape plant diversity, especially in the context of forest fragmentation. These enemy effects or their absence are more pronounced in forest edges than in the interiors.

In plain English, if some plants are not regulated by natural enemies, they will tend to grow unregulated and uninhibited. That would reduce plant diversity. So, applying pesticides and destroying insect herbivores and fungal pathogens destroy plant diversity too. In other words, the fragile balance between humans and nature has to be nurtured carefully. If we don’t, we will not be able to sustain biodiversity, especially in fragmented forests. That is why the decision to reduce the eco-sensitive zone around the Bannerghatta National Park deserves the scrutiny it is getting.

Many people feel that there is a trade-off between short-term economic growth imperatives and the need to preserve the ecology and environment. Western countries could burn hydrocarbons without a worry when they were developing countries. Developing countries have to be mindful of carbon emission and their commitments to international climate accords.

However, these commitments are not merely a case of being good global citizens. They are necessary even to maintain the health of local citizens. Without a healthy population, there is no sustained economic growth. So, sometimes, these trade-offs are more imagined than real.

The fragile balance between nature and humans was also the subject matter of the recently released Rajinikanth-starrer 2.0. After having feted technology in his earlier films, director Shankar reminds himself and his audience that technology and seeming technological progress are, more often than not, only mixed blessings. The message to be sensitive to the need for the winged population to survive is neither a luxury nor a concern of developed societies. The movie reminds the audience that by preying on insects and worms, birds maintain plant health and obviate the need for the application of pesticides.

Juxtaposing the message of the paper with the message of the movie gives us a beautiful insight. Birds are natural enemies for insects and worms. Without birds, we will have too much of them. Without them, we will have too little plant diversity. Nature has arranged itself well.

We do not understand it and frown upon any effort required to preserve its fragile balance as a hindrance. We clothe our laziness and our short-termism in intellectual terms, arguing that economic growth and poverty alleviation require relegating environmental considerations to the background. We do so at our own peril.

We cast our interference with and trampling upon natural arrangements as the triumph of human intellect. I view them with trepidation. For example, Financial Times featured an article recently on embryo selection (Profiling For IQ Opens New Uber-parenting Possibilities, 22 November 2018). The article briefly mentions personal and social costs of such embryo selection without going into details. It is fraught with immense danger.

It will be polarising at a social level. It will add yet another dimension of inequality to the ones we know. At a personal level, it will add immense stress as competition will be intense among the so-called “super kids” of which there will be plenty. There is a reason for nature’s bell curve distribution of many things. Consequences of extremely thick fat tails are unknown unknowns.

Indian cricketer Cheteshwar Pujara had said, “When you start playing shots [during a testing spell], that means your game is not capable enough to play the Test format. You are trying to survive rather than understand the situation and play accordingly.” He is right. When someone wishes to rush through a situation that requires deliberation, they are fearful and doubtful of their staying power. That is how humans are reacting to the complexities of the world, some of which may be self-inflicted. When Seth Klarman told the audience at Harvard Business School in October that one of society’s most vexing problems was its relentless short-term orientation, he was echoing Pujara. Short-termism betrays lack of confidence in long-term staying power.

Finally, the conclusion that natural enemies are useful for biodiversity is readily transferable to societies. Natural enemies are useful for diversity of opinions and ideas. So, the more we shut down opposite views (enemies), the less intellectually vibrant the society becomes. Just as biodiversity is beneficial, diversity of views is also beneficial. For that, one needs natural enemies. Therefore, common sense and self-interest dictate that we don’t smother natural enemies.

V. Anantha Nageswaran is the dean of IFMR Graduate School of Business (KREA University). These are his personal views.

Comment are welcome at