There was even greater incredulity when the back numbers for the new series were made available. The (old) 2004-05 series estimates were revised downward from an average of 8.8 per cent to 7.8 per cent for the period 2003-04 to 2007-08.This meant that average growth during this period of global boom was just 30 basis points higher than what it was in 2014-15 to 2018-19 (7.5 per cent), as against 130 basis points earlier. This revision was done over a decade after the events.
What gave credence to these doubts were calculations done by technical experts in the Committee on Real Sector Statistics, formed by the National Statistical Commission.
These showed GDP growth during the boom years to be actually higher than what was originally estimated under the 2004-05 series.
Rival politicians publicly sparred over these numbers, especially since they overturned received wisdom. This was that while the preceding UPA-II government (2009-14) may have lost the plot, growth during the UPA-I regime (2004-09, but excluding the global crisis year of 2008-09), was nevertheless significantly higher than during the tenure of the succeeding NDA government, as there was a global boom.
The debate over GDP numbers was joined recently by Mr Subramanian, with the publication of two papers in June and July 2019, respectively, following up on some observations made by him earlier.
First, he compared India’s growth trends with those of similarly-based major developing economies. Second, he took stock of the discrepancies between headline GDP numbers and high-frequency indicators.
Illustration by Binay Sinha
On this basis he estimated that India’s growth for the period beyond 2011-12 was about 2.5 percentage points lower than what the CSO had estimated. His second paper reiterated these arguments but avoided quantifying the overestimation.
GDP growth numbers should be broadly consistent with other economic data to be amenable to macroeconomic logic. If these are not, GDP numbers need to be revisited.
One can think of several data sets that have a strong positive correlation with the GDP growth rate. There may be occasions when, for specific reasons, a particular indicator is misaligned over a short period. Taken collectively, it is hard to imagine lead indicators being anything but positively correlated with GDP. All such indicators go into the calculation of the national income, directly or indirectly.
The indicators considered in the accompanying table are the index of industrial production, the index of agricultural production, automobile production, power generation, rail, road and air freight, exports, central tax revenue, corporate profits, employment, bank credit, information technology market size, net foreign direct investment (FDI) and the Sensex. Wherever monetary data sets are involved, such as bank credit, revenue growth and Sensex (asset) inflation, these have been deflated with the average of wholesale and consumer price inflation. This is because comparisons are made with real GDP growth. Dollar-denominated series are taken at face value.
Overall, these data sets point to robust growth between 2003-04 and 2007-08. This was followed by a sharp fall in 2008-09 and 2009-10, coinciding with the global financial and economic crisis. There was an equally sharp recovery that peaked in 2010-11. Growth, however, declined and flattened out thereafter, at levels significantly below the earlier period.
The data also points to a modest uptick between 2014-15 and 2017-18, when a new government had taken charge. Growth, however, fell back yet again in the wake of demonetisation and a flawed (Goods and Services Tax) tax reform.
The data also indicate that overall growth between 2014-15 and 2018-19 was markedly lower than in the period 2003-04 to 2007-08.
It is difficult to come up with alternative authoritative GDP numbers in the absence of a comprehensive review. The data, however, appear to be broadly consistent with Mr Subramanian’s estimation that GDP growth in the post-2011-12 period is overstated by 2-2.5 percentage points.
This means that the average growth between 2014-15 and 2018-19 could be closer to 5 per cent, as against 7.5 per cent estimated by the CSO under the 2011-12 series. It also means that GDP could be currently trending nearer 2-2.5 per cent.
To conclude, if one were to keep the contentious GDP figures aside, and look at other high-frequency data directly correlated with growth, it would appear that there was only a weak recovery following the growth crisis beginning 2011-12. The decline in potential growth from the highs of 2003-04 to 2007-08 would therefore appear to be structural, through a process of hysteresis. The sharper decline over the last few quarters, on the other hand, could possibly be cyclical, but on a lower base GDP.
That being the case, while macroeconomic (fiscal and monetary) policies might raise growth back to 5 per cent, major structural reforms would be required to raise the growth potential back to anywhere near the 2003-04 to 2007-08 levels. These include, in addition to the longstanding suspects, addressing frictions in the banking system, restoring the lost trust between government and economic players, and a well-articulated, predictable forward-looking policy road map that is adhered to.
The writer is RBI chair professor, ICRIER. These are his personal views