Friday, March 4, 2022

The confusing case of India's GDP data: Economic Times 5th February 2022

 

The second advance estimates (AE) on GDP were released on February 28, while the first AE on January 12. The revised estimates (RE) for FY2021 came out on January 31. There is excitement each time these numbers are released, quite like when the Monetary Policy Committee (MPC) minutes are released - although for the latter, there are diminishing returns, as most of the views of the members are known in the interim period. In any case, does what each MPC member feels really matter once the outcome is announced in the policy? Since by the time the MPC minutes are out, one gets ready to debate the content of the next policy and, to quote John Maynard Keynes, when facts change, people change their minds.

Coming back to the National Statistical Office (NSO) and GDP, the logic of these advance and revised estimates is baffling. First, the January 7 release of first AE for FY2022 is based on data that is available up to November for most components. This means that a call is taken in four months, with caveats like the impact of omicron not considered. So, what purpose is served? Answer: this is the basis for the edifice of the Union budget brought out on February 1. So far so good.

The next release - the first RE for FY2021 - comes out on January 31. This is for the past, and should ideally pass by like idle wind. But that's not so, since the calculators are out and everyone is busy calculating how FY2021 finally turned out to be. And this time we got to know that growth, instead of falling by 7.3%, fell by only 6.6%. It is supposed to mean that the economy did better. Which means that growth for FY2022 that was supposed to be 9.2% 24 days back is now actually down to 8.8%. Not because the economy slumped in this period, but because FY2021 was better than one thought. This is the infamous 'base effect'.

This base effect is often trotted out as a reason for anything that happens for which we can't find an explanation. Therefore, even while we pay high prices for our vegetables in the market, negative inflation is explained by the base effect.

The Union budget, which is preceded by the Economic Survey, still talks of 9.2% growth in FY2022 based on the 24-25-days-old story, because the January 31 RE comes out after the Economic Survey is published. Hence, there is confidentiality, as the NSO does not let out the numbers to the Department of Economic Affairs (DEA), which brings out the Economic Survey. But the budget uses the 9.2% to project the FY2023 GDP growth rate and accordingly fixes the fiscal deficit ratio.

Next, the RBI comes out with its policy on February 10 and still sticks to the 9.2% growth number for FY2022 to extrapolate the FY2023 number, which is put at 7.8%, The spreadsheets are out again on what the RBI estimate would be in case it used 8.8% as the growth number, which is the real rate after January 31, after all.

And just while we are digesting these numbers, there is a second AE for FY2022, which comes out on February 28 that now says with more certainty that growth will be 8.9% - which is closer to what the economist's calculator showed post 6 pm on January 31 and lower than the 9.2% that the Survey and RBI spoke of.

But wait. Here, too, the data does not take the January performance of industry. Instead, it uses other imputations. This means we are still lagging by three months.

Welcome to the world of Indian real time statistics. It is nothing short of mind-boggling, keeping economists and statisticians awake for two full months. We need to simplify everything and not have a plethora of estimates that confuse. The sequencing of data is important. Why not have the RE for FY2021 first and then the AE for FY22? The latter is needed to ensure that the budget has a chronological basis.

The second AE can be done away with, and we can have the provisional estimates that come out normally on May 31 as the next touchpoint. While such a flow of data keeps economists excited and fills TV airtime and newspaper space, is it really worth it?

 

No comments: