<b>Statsguru:</b> Making sense of GDP data
Mining was expected to contract by 1.8% in FY17, but ended up growing by 1.8%
 "graph")
premium
(Click on the image to enlarge)
Last Updated : Jun 05 2017 | 1:10 AM IST
Robust data form the bedrock of policy making. But when the data itself are misleading, how should policymakers respond?
Take the gross domestic product (GDP) estimates. Since January, the Central Statistics Office (CSO) has released three estimates of FY17 GDP, consistently pegging headline growth at 7.1 per cent across all three forecasts as shown in Chart 1. But GVA has been marked down from 7 per cent in the first advance estimates to 6.6 per cent in the latest provisional estimates, as shown in Chart 2. And despite concerns, the investment rate for FY17 has been marginally revised upward, as shown in Chart 3. Presumably this is because of a downward revision in FY16 estimates.
Look deeper and the revisions in estimates are bewildering.
Mining was expected to contract by 1.8 per cent in FY17, but ended up growing by 1.8 per cent, as shown in Chart 4. Similarly, construction was projected to grow at 3.1 per cent according to the second estimates, but after the preliminary data for demonetisation were factored in, it was scaled down to 1.7 per cent later. Such sharp revisions are also seen in the quarterly estimates. As Chart 5 shows, Q1FY17 growth was revised downward to 6.9 per cent in the second estimates from 7.3 per cent earlier, only to be revised upwards to 7.6 per cent in the provisional estimates.
Similarly, an investment contraction in the first half of FY17 has now turned into an expansion, with investments actually contracting in Q4FY17 as seen in Chart 6. So, how should monetary policy have responded? GVA excluding farm and government sector was expected to grow by 6 per cent in Q4. But as Chart 7 shows, it grew only by 3.8 per cent. While GVA by farm and government sector have been revised upward as seen in Chart 8, this too might be revised, as the MOSPI has clarified to Business Standard that only the data for April to February have been factored in for estimating the government spending.