India: GDP data revision - Eat your cake, don’t ‘cook’ it
Also published on Bloomberg, December 14, 2018
- In just a matter of months, a number of developments involving the Indian government have eroded India’s policy credibility on several fronts. All eyes are currently focused on the unexpected resignation of Urjit Patel as governor of the Reserve Bank of India (RBI) and his replacement by mr. Shaktikanta Das, who will probably be much more compliant with government demands.
- One might almost forget that the controversy surrounding India’s GDP back-series is not solved either and this discussion will probably surface any time soon again as well.
- The final GDP back-series published in November show that economic growth was much lower during the Congress-led United Progressive Alliance (UPA) government than the old series. The opposition now accuses the Bharatiya Janata Party (BJP)-led government of Prime Minister Modi of manipulation of GDP data.
- The politicized discussion over the GDP back-series is undermining the credibility of Indian statistics. This is worrying, especially for emerging economies such as India, since it can easily result in waning support of international investors.
- The single-best option for the government to solve the matter would be to attract an independent arbitrage team of international statisticians to re-assess the back-series. This would indisputably break down any suspicion of a government mingling with statistics for political gains.
Has the government been overplaying its hand?
In the Netherlands we have a saying: trust comes on foot, but leaves on horseback which basically means that trust is hard to gain but easy to lose. I am sure there is an Indian version of the saying, but I doubt if the government knows of its existence. In just a matter of months, a number of developments involving the Indian government have eroded India’s policy credibility on several fronts. Of course, all eyes are currently focused on the unexpected resignation of Urjit Patel as governor of the Reserve Bank of India (RBI) and his replacement by mr. Shaktikanta Das, a government-minded bureaucrat with a background in history, who will probably be much more compliant with government demands. In the heat of the moment, one might almost forget that the government already had another credibility issue that has not been resolved yet, and I am referring to the GDP back-series here. Although the RBI-government spat will likely also surface again in the coming weeks, the GDP back-series issue is an equally important one to watch. Like with cooking, it is the combination of ingredients that matters for the quality of the meal, which in this case refers to India’s policy credibility.
For those unfamiliar with the discussion, on 28 November the Central Statistical Office (CSO) released the official back-series on India’s GDP. In 2015 the CSO revised the Indian GDP data, using FY2011-12 as the base year. The rebasing operation in 2015 resulted in Indian GDP series only being available from 2011 onwards. The GDP back-series encompass the revised GDP data for years prior to 2011.
In June, a special committee assigned by the National Statistical Commission already released GDP back-series showing that economic growth in fact exceeded the old figures. The opposition seized the opportunity to attack the government, underlining that under Congress-led United Progressive Alliance (UPA) governments, India topped 10 percent growth not once, but twice. Ultimately, the July back-series were labelled as being ‘preliminary’ and more work had to be done. Now, the final GDP back-series show that growth was much lower during the UPA regimes. As expected, the opposition again rushed to accuse the Modi government of manipulation of GDP data.
Why would the government want to rig GDP statistics?
The big question is if the government has been pulling strings in order to rig the GDP back-series. The government has been pretty relentless in pressuring the RBI to relax loans restructuring and accept the installation of a panel which is going to re-assess the capital framework, which resulted in the exit of governor Patel. The question is: if the government is willing to undermine the independence and credibility of the central bank for political purposes, why wouldn’t it be prepared to undermine the independence and credibility of India’s statistical office?
What initially started as government interventions to tip the scale in the BJP’s favour in the run-up to the general elections next year, now seems to be backfiring. In the most recent five state elections, the BJP lost control of three. This could stall progress on the implementation of the BJP’s business-friendly reform agenda, which will likely result in levelling off of potential growth as well.
Especially if this could tip the scale in the BJP’s favour in the run-up to the general elections next year. Allegedly, the BJP is likely to win sufficient seats for a majority with its National Democratic Alliance (NDA) partners, but it will but have to resort to forging coalitions much more than he has done so far, perhaps even with non-NPA parties. This will definitely stall progress on the implementation of the BJP’s business-friendly reform agenda, which will likely result in levelling off of potential growth as well (Figure 2).
The proof of the pudding is in the eating (and not in the cooking)
There has been an intense debate in the media whether the methodology to calculate the back-series is flawed or not. Some have raised questions how data from the Annual Survey of Industries was replaced with financial company data from the MCA-12. Others state that the capital productivity miraculously has risen during the Modi period. But there also have voices defending the new back-series (e.g. here). We expect the debate to flare up in the coming weeks again.
Personally, I would like to refrain from speaking my mind on any methodological issues. For starters, I am not a statistician, but an economist. Just as doctors prescribe medicine in their day-to-day work, economists work with statistics intensively as well and perhaps have more knowledge than the average person on the mechanics of data compiling, the structure of the National Accounts, and so forth. However, an economist that criticizes the way statistics have been constructed does not differ much from a doctor telling the pharmaceutical industry that the components of a certain medicine are inappropriate. Although there undeniably are doctors with a degree in both medicine and chemistry, this most likely will be the exception rather than the rule.
In general, what doctors can do is assess whether medicine itself is doing its job based on empirical observations, e.g., patient recovery. And that is also what I will do here using a very simple test. First, I have estimated a quarterly GDP model for the period 2012Q1-2018Q3 based on growth of four economic drivers: car registrations, oil consumption, foreign direct investment and industrial production. I have experimented with other drivers as well, such as loans growth, electricity production, tourism, but these additional factors did not yield any improvement of the model. This GDP model is subsequently used to back-cast economic growth for the period FY2005/06-FY2012/13 (figure 3).
For fiscal 2009/10 and 2010/11, the model outcomes are in line with the final back-series rather than the old series. However, for 2005 to 2009 the model predicts even higher GDP growth than the old series. This at least indicates that the new back-series are not aligned with the development of some other key economic data. But I have to stress here that the GDP model fit is still widely variant, which surely underlines how difficult it is to economically capture what is going on in India for anyone.
Investor sentiment is like China in your hand…
As far as the quality of statistics in concerned, India has be careful not to go down the same route as that other large fast-growing Asian emerging economy: China. In 2007, premier Li Keqiang, then the party chief in Liaoning, told the US ambassador that regional GDP figures were ‘man-made’. Since then, many economists have been adopting different methodologies to gauge China’s economic performance. Rabobank too has developed a composite indicator based on several indicators, such as freight transport and electricity usage. So which GDP measure do you find more credible (Figure 4)? Keep this question in mind when answering: which line resembles a country struggling in recent years to keep up economic growth and fires on the credit growth cylinders each time the economy starts cooling? I guess President Xi Jinping not only loves a good game of football, but also favours the occasional can-kicking…
The bottom-line here is that unreliable statistics are worrying, especially for emerging economies who can easily lose the support of international investors. In the upcoming years, India and China remain heavily reliant on foreign direct investment if they want to avoid the middle income trap and live up to their growth strategy, which conveniently has the same name in both countries as well (Make in India/Made in China 2025).
What should the Modi government do?
Back to the GDP back-series. We have no reason to doubt the GDP back-series were produced according to the best methodologies available and the team of statistical professionals involved have been conscientious in their way of operating. The problem for the Modi government is that the discussion has been politicized. So what should be done to limit the damage? There are three options. First, the Modi government can withdraw the back-series, something which has been publicly advocated by even one of the leading newspapers in India. This would definitely result in loss of face and could even be framed by the opposition as a confession of foul play. Definitely a no-go. Secondly, Modi and other government officials can keep defending the back-series, as has been the case so far. The problem is that this won’t solve anything, as Modi will not be able to put the genie back in bottle. Many people will continue to question the government’s credibility in this matter, especially since it has already lost a great deal of credibility in the spat with the RBI. Third, and this is the best option: Modi should hire an international team of top statisticians to re-assess the back-series together with the Indian experts. Sure, there have been consultancy sessions with international experts, but these apparently have been insufficient to convince the public of the robustness of the current back-series. The international arbitrage team should be hired from the one country that produces the single best statistics in the world: the United States. In exchange for some more Indian geopolitical support, I am pretty sure that President Trump will be more than happy to send over a team of his top statisticians to look at those ‘phony numbers’.