Is it more important that inflation in primary articles has risen 16 per cent in the latest week (compared to the corresponding week a year earlier), or that the index, at 306 in June, was 4.5 per cent higher than what it was in March? Private consumption in the fourth quarter of 2009-10 rose in relation to the fourth quarter of 2008-09, but fell relative to the third quarter of 2009-10; which base period is more relevant for comparison, if one is to understand the latest trends? Questions such as these get repeated each time the official statistical machinery releases any set of data.
At the top of the list of to-do’s for the new Chief Statistician of India, T C A Anant, therefore, should be the job of getting the government to stop putting out year-on-year comparisons, and to do what the rest of the world does: adjust the latest data for seasonality and then annualise it — so that the country gets a realistic picture of what actually happened, for instance, to fourth-quarter private consumption. Once this is done for inflation, the country might find that inflation rates are falling; and if that is indeed the case, it would suggest a totally different course of action for RBI when it frames monetary policy. The logic of a shift to seasonally adjusted, annualised rates applies just as much to industrial production and other data.
Another critical lacuna in the country’s data system is the lack of official numbers on family incomes. Organisations like the National Council for Applied Economic Research do study family income data, but that is not a substitute for official numbers. India is one of the few countries that have no official numbers for this, because the stress is on expenditure — the survey-based poverty numbers look at expenditure levels. Also, the country’s official statistics focus on individual expenditure, whereas the domestic reality (in terms of spending) for most people is the family unit. The country needs to move away from the focus on per capita income numbers, which are statistically derived by dividing GDP by the total population and, therefore, not really useful for understanding what people actually earn and spend. It is much more important to understand personal incomes — which in India are about 60 or 65 per cent of “per capita income”, the rest being accounted for by corporate profits and the like. Along with this shift, the government needs to take the emphasis away from the mean (or arithmetic average, as in the average per capita income) to the median — the median family being one that has half of all families earning more and half earning less. This is what is done in the United States and many other countries, and imparts much greater understanding about who is the “average” nationals of those countries.
It is easy to list many other statistical tasks: the economy needs data on housing starts, inventory and industrial capacity utilisation, to better understand business cycles. Then, the industrial production numbers put out every month are of dubious accuracy, with reporting by companies having become voluntary a long time ago. The data on incomes in parts of the burgeoning services sector are even less reliable. An economy without reliable statistics is like a plane flying with defective meter readings; the pilot is likely to take the wrong decisions. Improvements are possible only if a major constraint (the lack of statisticians) is tackled. As Prof Anant’s predecessor told this newspaper, there are 1,200 vacancies in the official statistics collection/analysis wing of the government.