NEW DELHI: The credibility of India’s statistics system came under a shadow when a former insider, Arvind Subramanian who has served as chief economic advisor (CEA) in the finance ministry for four years until last year, claimed in a paper earlier this month that there was a significant overestimation of India’s economic growth for the period from 2011-12 to 2016-17 because of methodology changes. The actual annual average gross domestic product (GDP) growth may have been 4.5% for this period instead of the 7% officially estimated, according to Subramanian. Prime Minister Narendra Modi’s Economic Advisory Council (PMEAC), however, rejected the idea in a paper last week, saying Subramanian’s conclusion was technically inappropriate and demoralising to the dedicated team of official statisticians and economists. Subramanian should not have questioned the credibility of India’s statistics system, PMEAC chairman Bibek Debroy says in an interview but concedes that the system is not perfect and the unfolding economic slowdown warrants action. Edited excerpts:
You have not entirely rejected the gaps in GDP estimation that Arvind Subramanian has highlighted, have you?
Dr Arvind Subramanian is a close friend. No personal fights with him. He is not just any individual. He is not any economist. He is a former chief economic advisor. As he is a former CEA, his views are taken a little bit more seriously than the views of an average economist. I think a former CEA should be a little bit more careful before articulating views in the public domain or bringing out a working paper, which is methodologically questionable. The moment you put out headlines that say that India’s GDP was overestimated by 2.5% and India’s real GDP growth rate over that period was 4.5%, it tends to get picked up by international newspapers and magazines. It does great damage to the country and I think any former chief economic advisor, including Dr Arvind Subramanian, should be much more careful. Our paper (PMEAC’s rebuttal of Subramanian’s paper) is signed not just by members of the PMEAC but also by others who have felt uncomfortable with what Subramanian has done. For example, Arvind Virmani (a former CEA) and Charan Singh are not members of the council. However, they felt sufficiently disturbed by Subramanian’s paper to lend their names to this paper. On Subramanian’s methodology, I will not get into details. Based on a bunch of regressions and correlations, Subramanian is saying that as India is an outlier in the relationship between certain indicators and the GDP growth, the figure of 2.5% GDP overestimation is arrived at. Therefore, India’s GDP grew at 4.5%. (Subramanian had pointed out that the correlation between most indicators such as electricity consumption, two-wheeler and commercial vehicle sales, and railway freight traffic, and the GDP growth broke down in the post-2011 period to argue that GDP growth was overestimated.) In that case, one should exactly do the same regression with the other countries to check. Dr Surjit Bhalla, who used to be a member of the PMEAC, has mirrored in the case of other countries exactly what Arvind Subramanian has done. He has done it for 89 countries with the same set of indicators and methodology. In that list of countries, India is at number 28. Jamaica tops the list and Germany comes second. Then there is Belgium and Britain. So if the logic is that if it (the correlation) is off the line, India has cooked up the data, then these countries have cooked up data more. (Bhalla wrote in The Indian Express on Saturday that Subramanian’s method of estimating GDP suggested Germany over-estimates and Brazil under-estimates it the most, while India is only a mild outlier).
The holes in GDP estimation raise questions about the credibility of India’s statistics system, do they not?
The United Nations (UN) switched to a new system of national accounts in 2008, following which India also made the switch. It is not a switch that has happened overnight. The committee that examined it had experts in national income accounting. They recommended under the United Progressive Alliance (UPA) government the transition to a new system of national accounts. In January 2015, the first National Democratic Alliance (NDA) government implemented it.
It was based on consensus among economists and national income experts in various committees. There has been such transition in other countries too. There, too, GDP estimates have changed because of the transition. All countries in the Organisation for Economic Co-operation and Development (OECD) made the transition between 2013 and 2015. In 2013, because of these changes, the US GDP was revised upwards by about 2.5%. That is because one of the things that UN’s 2008 system of national income accounts did was to try and measure the value of intellectual property. So the US constructed a back series to estimate the contribution of Hollywood, which went back to 1929. All countries did various kinds of revisions. Did people in America say their country was cooking up data? They did say that now that you have got Hollywood, the new series is not comparable with the old series. However, no one questioned the whole system. No one questioned the system in the OECD countries because it was the process of moving to a better method of measurement. No one is suggesting for a minute that the new system is perfect. The question is whether it is better.
Are you totally dismissing the negative correlation between GDP numbers and the performance of various industries that Subramanian highlighted?
Do not bring in Arvind Subramanian’s paper and reduce the importance of those questions. We will come to that point. People ask questions about the imperfections in the ministry of corporate affairs’ (MCA’s) database (of companies). (MCA data is used in place of the annual survey of industries (ASI) in the new way of GDP estimation introduced in 2015). The ASI system was not perfect. No one is saying for a minute that the MCA system is perfect. I have statements by the former chief statistician and the former chairman of the National Statistical Commission Pronab Sen saying that it is not very obvious what the imperfections in the MCA system do to GDP. Why should I disbelieve him?
Also, do realise that Arvind’s figures —4.5%—is in real terms but GDP is in nominal terms, not in real terms. We need a deflator to deflate it. If we are looking at the nominal GDP, whether it is the CSO (Central Statistics Office) back series, or whether it is the alternative back series that was earlier done by Sudipto Mundle, there is virtually no difference in the nominal. In nominal (GDP) there is no discrepancy, which is a point, I am sorry to say, that Dr Arvind Subramanian does not even mention. The nominal figures are virtually the same. So the concern is with the GDP deflator (a measure of inflation), not about the nominal GDP growth rate. I want to stress this point. The ministry of statistics and programme implementation (MoSPI) had said in May that they are working on improving the price data and the deflators. Then it is a work in progress.
The economy is cooling down. What does one do about it?
Clearly, there has been a slowdown in the economy. No one denies that. I don’t actually need Arvind Subramanian’s working paper to argue that. A whole lot of people are arguing that there has been a slowdown in the system. It began several years ago, including when Arvind Subramanian was a part of the government. He has talked about that in the economic survey he authored without questioning the data. Most economists have said that this reflects certain structural weaknesses and they have suggested structural reforms to energise growth drivers. The drivers of growth essentially are consumption expenditure, investment, government expenditure and net exports. Obviously, there are limits to government expenditure because of fiscal compulsions. So, it boils down to what needs to be done to stimulate investments, private consumption expenditure, and net exports.
How fast should India’s economy expand to become a $5 trillion economy by 2024—the target set by Prime Minister Narendra Modi—and how do we manage that?
Several ingredients about what this government is likely to do has been set out in the President’s speech in Parliament. Several more things will be set out in the budget speech, which is not far away. So we need to wait a bit. The only thing I will say is whenever you are giving a dollar figure, it is subject to the exchange rate. The country’s GDP is always in the country’s terms. One can make some assumptions about the exchange rate and deduce what the growth rate should be.
Will doubts about robustness of India’s GDP estimation affect foreign direct investment inflows?
No one is saying for a minute the statistical system is perfect. It is constantly being improved. I do not think anyone’s cause is served by undermining the statistical system, domestically or internationally. Criticizing the way I am doing GDP deflator is one thing but masquerading it to question the whole Indian statistical system is unnecessary and uncalled for on the part of a former CEA.
Indian economy has too many problems to grapple with now. In addition to the pile of toxic assets in the banking sector and over-leveraged corporations, now we have liquidity problems in non-banking sector. Trade tensions add more uncertainty to investments and growth. Your thoughts on this?
Wait for 5 July for your answers. There are some problems that are exogenous to the country. You cannot do much about them. When India had very high rates of growth, the external sector was much better and we need similar external environment to get back to high growth rates. There are internal issues also and there is plenty of slack within the system to get respectable growth rate. The figures floating around do not suggest the NBFC problem is as large as the other problems we have known about.