How do you take policy decisions without the facts?
Given how so many critical datasets are so riddled with problems, it is just as well that a major revamp is on the anvil of both the IIP as well as the WPI. How critical this revamp is was best brought out by then RBI Governor D Subbarao when he said that the advance estimates of GDP for FY10 were initially 6.8%; this was later changed to 7.7% in the revised estimates in May 2010 and finally to 9.1% in the February 2011—a monetary policy that looked appropriate given the initial GDP would have looked loose given the final GDP numbers. Sadly, some of the biggest problems have been with the IIP, which is most critical for RBI. When the CSO released the new IIP—this is what is being attempted now, with a FY12 base in place of the FY05 one—in June 2011, this doubled the FY08 IIP and halved the FY10 one. GDP-industry grew 3.5% in FY12 according to the estimates made in January 2013 and by 7.8% according to the estimates made in January 2014 which used data from the Annual Survey of Industries. And just last week, core sector growth for March 2013 was revised from 3.2% to 7%.
Increasing the number of items for which data is being collected is as critical as deciding which items to drop, or which weights to change—with cereals consumption falling in relative terms, for instance, the WPI weights have to change for the index to have any meaning. Doing this, however, may be the easiest part of the revision. The manner in which the data is collected, and validated, is much more critical—prices in the WPI for electricity, for instance, were not updated from October 2011 till August 2012, when they were suddenly raised 13%. In January 2012, a wrong entry for sugar production saw IIP rise 6.8% instead of 1.1%; the July 2011 IIP surge, turns out, was due to an impossible 517% hike in insulated cables and wire—remove these, and the capital goods growth was just 0.3% in that month instead of the 63% reported.
Which is why, once the IIP/WPI is revamped, it is important that the results be validated. In the case of IIP, the comparators could be PMI as well as data from company financials—when government data showed engineering exports rising from $38 billion in FY10 to $68 billion in FY11, company financials showed exports for the top engineering firms had risen just 11%; two months later, the government admitted a data glitch. In even the case of the CPI which is a new index, this newspaper’s complaint has been the data doesn’t reflect the ground reality. Garbage in equals garbage out, never mind whether an index uses new weights or new commodities.