No need for inflation index for govt staff: Sen

India’s Chief Statistician Pronab Sen has said there is no need to construct a separate price index for government employees as suggested by the Sixth Pay Commission.

He said the existing inflation measure for industrial workers was working fine and building a new index would be costly.

The commission has suggested a separate index reflecting the spending of the government employees so that they can be protected from inflation.

“Frankly, it (the CPI for government employees) will not make much of a difference”, Sen said. “The consumption pattern of government employees is similar to industrial workers,” he added.

Sen put the number of government staff (in states and at the Centre) at around 12 million and said their dearness allowance (DA) was revised periodically based on the reference inflation number, which is currently the CPI for industrial workers.

After the commission’s report was submitted to the finance minister in March, the government wrote to the National Statistical Commission on the proposal for a separate CPI for government employees. Sen is also the secretary of the commission, which is yet to reply on the matter.

Nearly 60 per cent government employees are in the Group D category, which is the lower end of the government hierarchy, and their consumption pattern is not different from that of industrial workers, said Sen.

The problem in building a new index is that government employees form a small portion of India’s total workforce, which is estimated at 420 million.

At 12 million, government staff make up less than 3 per cent of total workers. This will make it difficult to find enough samples from the existing surveys. The number of industrial and agriculture workers are estimated at 40 million and 208 million, respectively.

An important rule in random sampling is that each member of the population should have an equal chance of being included. Thus, if we have a sample size of 10,000 workers, it will have only around 300 government employees, which is too small a sample to construct a separate index.