Most forecasts of GDP growth in 2008-09 are down to 8% or thereabouts. RBI has a range of 8-8.5% and some forecasters have brought down the number to 7.5% and lower. However, CMIE?s forecast continues to be 9.5% and this is ascribed to two main reasons: the lower-than-expected tightening of monetary policy and an industrial recovery, with the index of industrial production (IIP) growing by 8.6% in February 2008, compared with low growth between November and January. Implicitly, this raises questions about the quality of IIP data, and this is more than the issue of changing bases and weights frequently. First, data collection systems have been driven by the factory-based Industrial Revolution model of the West, and this has dubious validity in a country where large chunks of economic activity are unorganised and self-employment-based. Incentives for formalisation aren?t adequate. Consequently, even the government admits that the IIP doesn?t cover all manufacturing and the small-scale sector is under-represented, since there is a threshold turnover for cut-off. To take one example, the IIP includes mining and electricity, but not gas, water supply or construction. Second, value figures are deflated by the wholesale price index (WPI), which is itself an index open to question.
Third, in deference to the IMF?s data dissemination standards, IIP figures are released within six weeks, when information isn?t available, leading to differences between advanced and revised figures (there are two revisions). Fourth, the number of items has increased to 538, without ensuring improvements in data delivery undertaken through the government system. There are, thus, reasons for disquiet both on coverage and administrative systems of data collection, documented in the 2001 report of the National Statistical Commission. The website of the ministry of statistics proclaims, ?The industrial growth based on the revised IIP does not, therefore, seem to reflect the perceived ground realities.? It is also the case that post-1991, there has been more deterioration in the quality of industrial statistics, since easing of registration requirements has removed some automatic submission of data. The commission recommended improvements in CSO?s functioning, but fell short of drastic suggestions. Since important policy decisions are taken on the basis of indices like IIP and WPI, the country can?t afford cosmetic improvements. Private and outsourced collection of data, as opposed to processing by CSO, is distinctly possible and will be more efficient.
