How CSO plans to avoid issuing wrong GDP data

Having committed a blunder while calculating this fiscal's first quarter GDP from the demand side, the Central Statistics Office on Tuesday said it has commenced a corrective mechanism that would start reflecting in the growth figures from the second quarter.

"The process of rectification began immediately. . . it has already begun. . . the immediate correction was already there. So far what else I can do to improve the quality of the expenditure side figures, I have already started," Chief Statistician TCA Anant told PTI.

Anant said he will take more corrective measures, once an internal department report gives it suggestions.

"I will build in the corrections of whatever incremental changes have been suggested internally immediately, as they come in.. Some of them will become operational from next quarter release," he said.

Announcing the economic growth figures for April-June, FY'11 on August 31, CSO wrongly calculated GDP from demand side (market prices) at 3.65 per cent. The next day, it was corrected to 10 per cent.

Anant said, "Whether there is anything more to be done in the expenditure side. . . Once I have a (internal) report I will decide."

Anant, who is secretary with the ministry of statistics and programme implementation (MOSPI), said the same degree of robustness is still not there in calculating national income figures from the demand side, as there is on the production side.

"It is also the case that these set of numbers have not yet achieved the same degree of robustness as the production side estimates, the factor cost side," he said.

The CSO had correctly calculated the economic growth figures from production side, known as factor cost, at 8.8 per cent for the first quarter.

Anant said, "Initially we released factor cost, the market price estimates and the expenditure side have started to come out only in the last 2-3 years.

"This (wrong calculation) has happened because of some gaps in that process. . . . I am trying to understand what are the assumptions being used in this side of the process and whether better error correction protocol can be put in place to avoid this."

He said some errors crept in, which was not picked up partly because the persons concerned were not calculating the figures.

"Clearly there were errors in the system which avoided this getting picked up," Anant added.

The CSO had attributed the wrong calculation to the use of inappropriate deflators, used for converting figures at current prices to some base year prices (in this case 2004-05).

Anant said, the CSO uses different deflators for computation of GDP from the expenditure side.

To gather data on private final consumption expenditure, the department needs to take into account the government expenditure and exports, among other things. For each one of them they use different deflators.

"Because of that somewhere in the process of compiling those numbers, they made some errors with regard to deflators they were using," Anant added.

He said the figures in the quarterly estimates are part provisional and partly based on real figures -- like revenue collection, government expenditure, Anant said, "When you project on the basis of a few numbers there may be inconsistencies, across different projections. So certain amount of exercise then goes in trying to match them up. If these are not adequately coordinated or matched up, you can get discrepancies in the figure."

"Part of the error collection protocol would try to pick these up. . . there was error in the process at the first place and that error did not get picked up was the weakness of the protocol," he added.

He said the CSO will fix the weakness in protocol immediately, but if the error is occurring at the source of information, then it would take a slightly longer time.