Statosphere: Where The Figures Don't Meet

There are many jokes about statistics and the tricks that numbers play. Behind any jumble of numbers, what you see is not always what you expect to get. Any arrangement of numbers can be interpreted differently by different people. So, it was not entirely surprising when the telecom ministry last week ordered an audit of the country’s top mobile operators. The beef? The ministry feels that mobile operators, to avoid paying the government a higher licence fee, are misrepresenting their revenues. Telecom operators are expected to pay a licence fee to the government, which is calculated as a percentage of their revenues. Some services attract a lower percentage compared to some others – for example, while revenues from long distance services attract a licence fee of only 6%, some other mobile services attract 10%, leaving ample room for arbitrage. But, what’s surprising is the discovery made by the ministry – some mobile companies were reporting one set of numbers to the Telecom Regulatory Authority of India, and a completely different set to stock exchanges.
Today’s column is not concerned with the telecom regulatory infractions but is suitably spurred by the interesting example of two sets of balance sheets, representing the same core numbers. Statistical discrepancies have become a way of life in this country. At the platinum jubilee celebrations of the Indian Statistical Institute in Kolkata, in December 2006, prime minister Manmohan Singh observed: “Another key issue has been the discrepancy in the estimates of the same variable from two different sources. The classic example is the difference between the National Accounts and National Sample Survey estimates on consumption expenditure. The present system has also not been able to provide adequate information on basic socio-economic indicators that are crucial for micro level planning.”
This is a concern expressed even by the high level committee on estimation of savings and investment, chaired by C Rangarajan (former chairman of the prime minister’s economic advisory council), which submitted its report recently. At the macro level, the report traces how committee after committee has pointed out the statistical discrepancies between savings and investments estimates, with savings far exceeding investment numbers. For example, the report notes how the Chelliah expert group had noted that the order of discrepancy in 1994-95 was almost 2% of GDP. Many other similar examples abound in the macro-economy.
At the micro level, there is the absurd situation of the same issue being reported differently by different ministries. Take the example of the cotton crop. Curiously, both the textile ministry and the agriculture ministry provide their own estimates on the annual cotton crop and, predictably, both are widely divergent. This has been happening over many years and arises primarily because the two ministries use different methodologies. For example, take the crop estimate of 2004-05: the textile ministry reported it as 243 lakh bales of 170 kg each, while the agriculture ministry number was 16.43 million bales (of the same size, or 164.3 lakh bales). The advance estimates for 2007-08 were similarly at variance: 315 lakh bales versus 258.1 lakh bales, respectively.
The same kind of statistical discrepancy has existed even in the trade deficit figures presented by Reserve Bank of India and the numbers generated by the commerce ministry, with the central bank’s trade deficit estimate typically being higher than the ministry’s. The difference arises largely in imports estimates, with the RBI number based on foreign exchange paid on the day of the import rather than the year’s average rate used by the commerce ministry in its calculations. But, more importantly, the RBI’s numbers include defence imports, which never have to pass through customs scrutiny (and therefore escape getting caught in the commerce ministry’s statistical dragnet) but have to be paid with hard foreign currency.
But, despite these manifest discrepancies (some of which exist even in the developed economies), this column’s heart goes out to the unseen and unheard class of professionals called statisticians who have been given the unthankful job of collecting data from across this vast country, teeming with diversity unknown to any other nation-state, using technology that’s horribly outdated and given budgets that are constantly shrinking. And, despite all this, the fact that they still manage to distil some semblance of sense into that thicket of numbers -- which is then used by policy makers to drive decisions and massaged by politicians for gaining mileage – is creditworthy. Spare a thought for them too when you next boast of how India’s going to grow by over 5%, when the rest of the world economy is shrinking. Without them, you wouldn’t have had a clue.