Abheek Barua: Can we get better economic data?

The way round this problem is to engage a professional agency, perhaps a market research firm.

One of the soft indicators of development or progress in an economy is the quality of economic data that its government agencies put out. On this metric, India scores rather dismally. China is perhaps much worse than India but that should hardly be a consolation. While things like GDP growth and indices of industrial production are bandied about rather freely in the media and on the conference circuit, I am yet to find someone who bases his hard business decisions on these numbers.

If you are not convinced by my claim, try this experiment. Ask any fund manager who invests in Indian stocks whether he uses the monthly industrial production index (IIP) (which the government's Central Statistical Organisation releases) to make his forecasts either for market movements or company profits. I can bet my last paisa that this answer would a categorical "no".

Bond traders often have no option but to fall back on the official wholesale price index in gauging inflation expectations. So does the Reserve Bank when it takes a call on monetary policy. However, there seems to be a broad consensus on the fact that the official indices often "misreport" inflation.

The latest controversy surrounding inflation relates to commodities like metals whose prices have firmed up continuously in the last four months, touching historic highs. Yet the "metals and alloys" sub-index of the Indian wholesale price index has remained virtually flat in this period. While there could be some difference between domestic and international prices, discrepancies of this magnitude are hard to fathom.

There is another set of issues that users raise about price-indices. These relate to their "representative-ness". There is, for instance, a common perception that the basket of goods used to compute the index (that has the year 1993-94 as base) has lost its relevance. I am not sure of the exact composition but the intuition that a bunch of things that represents the mix of production in 1993 ceases to achieve that end in 2006 has something going for it. There is the bigger issue of the choice of index when it comes to measuring inflation. Headline inflation in India refers to wholesale prices and is used to guide monetary policy. In most other economies, change in households' cost of living measured by the consumer price index (CPI) is taken as the appropriate inflation measure. Again, while services constitute the biggest chunk of GDP, there is no direct index of service prices. This is an endless list of problems, and I must, here, acknowledge the fact the government is not entirely indifferent to this issue. The whole-sale price index, for instance, is being entirely recast. The base year is being shifted from 1993-94 to 2000-01 and the new basket of commodities is being enlarged to include 1,100 commodities rather than the 435 used earlier. A working group is looking into this and their report should be out by the middle of this year. Efforts are also on to spruce up the IIP and to create separate rural and urban CPIs.

This is certainly a commendable effort. In fact, I attended a workshop on the new price index, and was quite impressed by the enthusiasm with which ministry officials spoke about this exercise. However, I have a niggling suspicion that unless the entire paradigm of data collection changes, the lack of credibility will remain.

One of the key problems is of course that data reporting is entirely voluntary. While it is possible for government agencies to exert some pressure on PSU units to report their figures, it is impossible to use the same tactic for private sector companies. It is also irrational to expect an employee of a private sector unit (or a PSU for that matter) to put the task of reporting data high on her list of priorities. Thus, if the current system of data collection continues""one that relies on sending questionnaires and hoping that people respond""quality will continue to suffer.

The way round this problem is to engage a professional agency, perhaps a market research firm. Data disclosure will still remain voluntary but the presence of a professional collection intermediary can make all the difference. Let me give an example. Our bank is sponsoring a purchasing manager's index which measures the manufacturing sector's performance on different parameters at a monthly frequency. We have hired a market research firm to collect data from the sample of companies. The response has been remarkable. I am certain that if we had just mailed some questionnaires, we would get half the response that we do now.

Who pays for all this? I would recommend pricing some of the data and also see a ready market. Construction companies, for instance, that use this index to price government contracts would be happy to pay for the sub-index relevant to them if they are convinced of its accuracy. So would banks and bond firms, to take other examples.

As far as commodity prices go, more active collaboration with exchanges like the NCDEX or MCX would certainly in help in getting accurate, timely data. Another issue raised at the workshop was that private companies are often a little squeamish about sharing information with the government, particularly the stuff that relates to their balance sheet. This is perhaps a valid issue and can be addressed by making data agencies of the government semi-autonomous and not just a sarkari department.

Finally, I was surprised not to see a single representative of a major manufacturing firm at the workshop. I would presume that the prices of their wares would be a key input in the index. It is important to give this exercise as much publicity as possible if companies are expected to participate in this data drive. A couple of advertisements in the national dailies might just help.