Sharp revisions in the GDP growth rate tell a tale of uncertain times

Is the growth modeling fraught with errors? If so, how can the policy makers frame policies in such a scenario?

Seven years ago, the 2010-11 Economic Survey authored by then chief economic advisor (CEA), Kaushik Basu, had predicted the gross domestic product (GDP) growth rate at nine per cent, give or take 0.25 per cent, for the year 2011-12.

As the year advanced, this optimism gave way to pessimism and the mid-year review of the economy tempered the expectations by pegging the growth at 7.25-7.75 per cent for 2011-12. The actual growth turned out to be 6.5 per cent for the year, according to the first provisional estimates (based on the old series of 2004-05).

This background amply substantiates a point that a sharp revision in economic growth isn't a new thing in India during times of uncertainty. In those days, it was the deepening euro zone crisis that was the pain point.

However, one might understand the revision from a high economic growth rate of nine per cent predicted that year, but it is more difficult to grasp the sharp revision in the current financial year, when the government's initial target was just seven per cent and the Reserve Bank of India's was 7.4 per cent.

While the government has not formally announced any new number for GDP growth, chief economic advisor Krishnamurthy Subramanian told Business Standard that it will be in the 6-6.5 per cent range. 

The central bank, on the other hand, has since successively lowered its forecast to 7.2 per cent in April, 7 per cent in June, 6.9 per cent in August and 6.1 per cent in October.

The sharp revision in October was due to an over six-year-low economic growth at five per cent in the first quarter of 2019-20 --  which came as a surprise to many -- and high-frequency indicators that followed.

By the time RBI announced its October policy, one of the high-frequency indicators – the Index of Industrial Production (IIP) – still showed a respectable growth of 4.3 per cent for the month of July. But RBI was worried because it was lower than 6.5 per cent a year ago.

“With regard to the domestic economy, the slump in real GDP growth to five per cent in the first quarter of 2019-20 has been followed by generally weaker high-frequency indicators for the second quarter. Industrial production was lower in July 2019 on a year-on-year basis, pulled down mainly by manufacturing,” RBI governor Shaktikanta Das said in his statement at the time of the October policy review.

This IIP growth was later revised to 4.6 per cent for July. However, the same data showed that IIP contracted to 1.1 per cent in August, the sharpest fall in 81 months. Will this lead the RBI to further revise the growth downward? This would probably be dictated by the second-quarter GDP growth numbers that would come by November end, a few days ahead of RBI’s monetary review on December 5.

Awry projections across the board

RBI was not alone in cutting India’s growth rate. Take the International Monetary Fund (IMF), for instance. It originally pegged India’s economic growth rate at 7.2 per cent for FY20 in its April outlook. However, a few months later in July, it cut the grwoth rate to 7 per cent. Now, in its October World Economic Outlook, the Fund has further scaled it down to 6.1 per cent, in line with RBI’s estimate.

The World Bank's downward revision was the sharpest, at six per cent from its earlier estimates of 7.5 per cent. OECD scaled the growth rate down to the lowest, at 5.9 per cent from its earlier forecast of 7.2 per cent.

Other agencies, such as India Ratings, Moody’s Investors Service and Crisil, also cut India’s growth rate sharply for 2019-20.  

The moot issue is whether these revisions point to errors in growth modeling. How can the policy makers frame policies in such a scenario?

N R Bhanumurthy, professor at the National Institute of Public Finance and Policy (NIPFP), said he is not sure about the weakness in growth modeling. He had projected India’s economic growth rate for the current financial year at 6.1 per cent in March this year, at an event in which UNESCAP pegged it at 7.3 per cent. Also, in August, 2018, he had projected the growth rate to be 6.7 per cent for 2018-19, but was criticized by many for being pessimistic, he said. Ultimately, the growth rate turned out to be 6.8 per cent for FY18.

He partly blamed the wrong GDP forecasts to weakness in the back series data given by the Central Statistical Office (CSO), now known as the National Statistical Office (NSO).

Back series data to blame?

To put things in perspective, a panel that had been appointed by the National Statistical Commission (NSC) came out with its back series data. Bhanumurthy was part of that committee. Later, however, the CSO junked the data and came out with its own back series data, drawing criticism from many quarters.

The committee’s data showed that the economy grew in double digits twice during the two tenures of the Manmohan Singh government -- by 10.23 per cent in 2007-08 and by 10.78 per cent in 2010-11, against earlier numbers of 9.32 per cent and 8.91 per cent, respectively. The new series has the base year of 2011-12 and the earlier series had 2004-05.  

There are some years for which the new figures showed deceleration in growth as well. For example, the growth stood at 4.15 per cent according to the new numbers, against the earlier 6.72 per cent in 2008-09 when the global financial meltdown had a ripple effect on India. That year also belonged to the previous United Progressive Alliance government.

After scrapping this data, CSO came out with its own estimates which showed that the GDP grew at an average of 6.7 per cent in nine years of the UPA regime, 2005-06 to 2013-14, in comparison to 7.3 per cent growth under the current National Democratic Alliance government. GDP growth for all years prior to 2010-11 had been revised downwards.

Bhanumurthy said he used the committee’s estimates when he forecast the growth rate for 2018-19 and 2019-20.

What about policy making?

To a query on whether the sharp revisions would affect policy making, Bhanumurthy said that is precisely his point. Talking again about back series data, he said any such data should not disturb the macro economic relationship. He said the CSO’s back series data sharply revised growth rates downwards for the period 2005-06 to 2008-09, when investment rates were high. Whereas there was not much change in the committee’s back series data and the previous calculations on the 2004-05 base year.

Former chief statistician Pronab Sen said these sharp revisions in the economic growth rates indeed point to lacunae in the growth modeling. 

Explaining this point further, he said the growth modeling assumes continuity and is incapable of measuring discontinuity. Both, the demonetisation announced in 2016 and the goods and services tax (GST) introduced in July 2017 were discontinuities that these modeling could not estimate. Most analysts were surprised by an over 8 per cent growth rate in 2016-17, the year of demonetisation, he said.

When asked why are we revisiting the decisions on demonetisation and GST, which were taken two or three years ago, he said because of the simple reason that these shocks started showing on the economic data that time. 

Sen asserted that sharp revisions in GDP growth affect policy-making. However, the point to note is that GDP is only an outcome model in which the underlying causes are not addressed. The aim of policy making is to correct those causes, not the symptoms, he said.