Let’s be pedantic once again and quibble about data. Manufacturing has registered and unregistered categories. For registered manufacturing, annual data are collected through the Annual Survey of Industries (ASI) and this is part sample (survey) and part census. Data on unregistered manufacturing surface once every five years, through a survey. Plus there are SSI (small-scale industry) and IIP sources. In August 2001, the National Statistical Commission submitted a devastating critique of India’s statistical system and all four sources of manufacturing data — ASI, unregistered manufacturing, SSI censuses and IIP were hauled over the coals. There are time-lags and sampling and non-sampling errors. Units that shouldn’t be included (they have closed down) are included. Units that should be included are excluded. Different databases of unregistered (census or survey) manufacturing vary widely. Perhaps understandable, because differing concepts and definitions are used. This is compounded by the lack of adequate data on SSI and unorganised traditional industries. Finally, other than the problem that the IIP represents only 80 per cent of manufacturing, there are problems associated with low response rates, small samples, unsatisfactory weights and non-representation of the unorganised sector. As a generalisation, manufacturing data are therefore somewhat satisfactory for registered manufacturing and extremely unsatisfactory for everything else. But having said this, one doesn’t get the sense that this manufacturing growth is in the realm of lies, damned lies and statistics.

Manufacturing’s (not industry’s) share in GDP is now around 18 per cent and no one denies it should be higher. But how high should it be? It is impossible to answer this question, and comparisons with China aren’t fair. Targets of 30-35 per cent float around, although people don’t always explain whether they mean manufacturing or industry. Shares in GDP are a function not only of what absolute growth in manufacturing is, but also of what happens to services. Unless one conjures up a dismal service sector scenario, and that is unlikely, manufacturing (and industry) picks up what agriculture loses. In the GDP share, it takes manufacturing 20 years to pick up five percentage points. Therefore, even a share like 25 per cent should be great. Instead of shares in GDP, it is better to have absolute manufacturing growth rate targets of something like 12 per cent. Are we there, or is this a flash in the pan?

... contd.